The culture (specifically, the emphasis on collectivism or individualism) plays a crucial role in business decisions and the results affect both the individual and the group.  Looking at the cultural roots, the recent rise and fall of some technology companies, and the decisions made by their management, perhaps we can learn some lessons to address the current issues facing both Japanese and US companies in improving the efficiencies of current IT infrastructure management.   Current trend toward next generation computing clouds offers an opportunity to reexamine the evolution of POTS, PANS and SANs and learn from the lessons of the past.

While there is a lot of debate about pure individualism and collectivism, free markets and planned economies, history and evolution of POTS, PANS and SANs argue for a middle path.

Kyudõ and the Living Bow

The master moves with graceful elegance.  You can feel the intensity of purpose behind his each move.  You see the master’s face radiating cherubic serenity of a stress-free baby looking into the comfort of mother’s eyes.  Blindfolded, the master focuses his attention on the kiai (the spiritual energy – the vibration) to emulate the four seasons, the spring, summer, autumn and winter.  The mind, the posture and the breathing are all orchestrated in harmony to create that moment of perfection when the master, the bow, the arrow, and the target become one.  All I see is the arrow hitting the bull’s eye perhaps 60 meters away.

According to Jackson Morisawa [1], “The ‘aim’ of kyudõ at Chojen-Ji is never an external focus on the external target.  It is, rather, an internal ‘aim’ focused on the center of the perfect cross.  When one reaches the ‘point zero,’ the body, mind, and energy are in perfect harmony and serenity prevails.  This internal harmony works in relation to the external effect, and the arrow that pierces the mark is penetrating and alive.”  Yamada Shõji [2] discusses the influence of Zen on Japanese archery.  “Dead blade and living sword are Buddhist concepts taught in tantric (Shingon) lineages. We take this principle and merely rename it the dead bow and living bow.  It is the same principle as expressed by the saying “Rejoice in death and live (kõshi sokusei); [Try to] insure life and die (hissei sokushi).” [In other words,] when one’s mind is troubled by fear, one’s bow is dead. When one is willing to sacrifice oneself and regards lightly the loss of one’s own life, then one’s bow comes alive.”

A Zen master in Japan, once, told me that the interesting moment of enlightenment is not when you hit the target.  It is when you miss the target, that the self is revealed, and the understanding of your own shortcoming makes you a better person ready for the next aim.

Watching the unusual feat of the living bow and learning about Zen were part of my studies as a student at Japan American Institute of Management Science (JAIMS) in the paradise island of Oahu, Hawaii.  I became an “accidental student” when my boss, Bud Wonsiewicz, at U S WEST, a telephone company, came to me and suggested that I spend a few months in Hawaii and a few months in Tokyo studying Japanese language, culture, and business.  The purpose was to return to US WEST and establish closer partnership in technology and business, development & transfer between the Japanese Telecommunications company Fujitsu and US West.  Fujitsu has a school in Honolulu that brings students from many countries including Japan to learn business management in a cross-cultural setting.  Top management of Fujitsu and US West decided to choose someone from US WEST management to attend this program and spend a few months in Fujitsu Japan working with various groups learning the Japanese Way.  According to Jinjiro Dodo, the then President of Fujitsu America, Japanese understood the role of personal relationships in creating value in business partnerships and Fujitsu created the school with a long-term view to develop business leaders with inter-cultural understanding and close inter-personal relationships.  He also emphasized that both language and culture play a very important role in business conduct and JAIMS encouraged Japanese students to experience the culture through internships in the US and non-Japanese students to experience Japanese culture through internships in Japan.

US West, one of the baby Bells created by the divestiture of the then regulated monopoly, AT&T, was a quintessential example of the rugged western culture of individualism, set out to create dust so as not to fall behind and eat the dust others created (as the US WEST commercial in this blog says).   Bud Wonsiewicz was a renaissance master with a lineage from MIT and Bell Labs who came to US WEST to bring innovation and transform the plain old telephone company to an information services powerhouse.  He had worked with other masters who made UNIX at Bell Labs in Murray Hill, New Jersey.  He was starting experiments at US WEST Advanced Technologies using video walls and information appliances in the late 80s.  He had started collaboration with various universities working on different projects including the transmission of holograms over telephone wires at the MIT Media Labs.  He was impressed with the long-term focus of Japanese companies and took the opportunity to collaborate with them.  Later US WEST was to start a project with Time Warner, Toshiba and Itochu to bring information services to home on cable.  More recently, an Itochu executive told me that the Time Warner deal turned out to be a very profitable venture for Itochu.  The decade that followed was a very exciting time at US WEST for innovators, until a few executives were carried away with the Internet bubble and merged with Qwest.  It did not matter to them what havoc the dust they created wreaked on the universe!  The birth and death of US West should be a good case study for the IT industry to learn how the bubble mergers with inflated valuations led by individual egos can go awry and affect thousands of employees, their dreams and the entropy of the universe.  On the other hand, the case study of Steve Jobs would show the positive impact of a single individual.  Even when everybody else had given up on Apple, he resurrected the software innovation and hardware design, leveraging the institutional knowledge.

The next few years, beginning with my first encounter with Dodo San, turned out to be an adventure in discovering new insights into life, technology, quality management, Japanese business, personal relationships and myself.  Learning about Zen, and watching kyudõ with awe in Honolulu, were just the first steps of a journey that opened my eyes to the power of innovation in America, human resource potential in India, the long-term focus and process savvy of management in Japan.

Zen, derived from Buddhism, which originated in India, connects the oneness of individual’s body and mind to the oneness of the universe with the principle of connectedness of all sentient beings.  The resulting concept of “Sangha” has played a role in the evolution of collectivism.  As Lama Surya Das summarizes [3], Sangha implies true community.  “It is taking refuge in true community itself; communion with others – collaboration, connectedness, engagement, responsibility, working together in an organic way, and not in an addicted manner.”   The concern for the group often played a prominent role in Japanese decision-making process, which often resulted in the individual willingly or unwillingly compromising self-interest in the interest of the group.

Later, when I met Dodo San in Tokyo, he told me over a bottle of sake in a small sushi bar, that globalization is creating a new connected universe where collectivism will transcend national boundaries and he saw building global interpersonal relationships as critical to Fujitsu’s future.  The Internet had not yet become the force that it is today.  Broadband was just beginning to take form and Fujitsu aspired to lead the broadband revolution with emerging Asynchronous Transmission Mode (ATM) Switching and SONET transmission technologies.  Fujitsu wanted to play a major role in computers (acquiring ICL in Great Britain) and the PC revolution.  Cisco was still a Silicon Valley startup struggling to understand what it takes to build telecommunication grade systems for the Internet, which originated from a “send and pray” protocol.

The connectivity Dodo San was talking about was very different from the global connectivity today at the speed of light, to which we are accustomed.  Twenty years later, Cisco has established itself as the leader in bringing Telecom grade trust to IP networks.  ATM switching and all the associated companies have disappeared and Fujitsu is struggling domestically and has toned down its global ambitions.  New companies such as Microsoft, Google and Amazon have emerged, with that unique American innovation, as companies that are the new global communication, collaboration and commerce enablers.   They are offering new ways to access computing services through clouds challenging the conventional Information Technology infrastructure solutions.  Their introduction of computing clouds using the emerging virtualization technology is reshaping the IT infrastructure and its management.  This new revolution will claim new casualties in the conventional IT infrastructure market.  A host of storage and server companies will cease to exist as the next generation servers and storage devices become commoditized.  Some storage companies are already for sale while others are migrating to server and network infrastructure domains to find new ways to survive.

Space -Time Services Continuum & Information Services Changing the Landscape from Banking to Healthcare

The canonical nature of space and time manifests itself in many domains namely, spirituality, theory of relativity, quantum nature of the universe and in just day-to-day living.  In day-to-day life, people have learnt to trade space and time to better organize their lives (manage entropy) and improve productivity.  People are willing to pay to save time as they do when they use faster transportation to move about.  People are also willing to pay for services that allow them to avoid travel by bringing objects and information to where they are and when they want.

The services that people are willing to pay usually have the following attributes [4]:

  1. Choice based on individual requirements
  2. Interactivity
  3. Mobility
  4. On-demand
  5. Cost based on fulfillment-latency tolerance (faster fulfillment at higher price)
  6. Pay per-use  and
  7. The ability to bundle and unbundle the received services

Current generation of information services coupled with high-speed transportation networks have the ability to fulfill these attributes and improve productivity of the individual or group.  Many of the new service offerings in e-commerce, electronic banking, mobile banking, and mobile wallet [5] fall in this category.  Similar attempts are being pursued to improve productivity of health care services.  Information access at speed of light has raised the bar for consumer information services.  The ability to correlate information from many quarters at the speed of light offers the possibility for a new class of services that were not possible before.  In addition, the explosion of social networking, multimedia interactive services and collaboration of groups transcending time and space have placed new demands on service creation, delivery and assurance infrastructure.

In contrast to business services that have higher profit margins, the consumer services have far lower profit margins and infrastructure that the businesses can afford becomes unaffordable for consumer service providers.  While the affordable price points start falling, the volume of services delivered (scaling) increases drastically along with unpredictable fluctuations in demand.  This was true with Plain Old Telephone Service (POTS) and it was true with Pretty Amazing Network Services (PANS) that the Internet made possible.  We are again seeing the same phenomenon with emerging mobile internet services and social networking services.

Services Management and the Information Technology Black Hole

On one hand, any service that is not managed cannot scale.  On the other hand, services always out run the pace of their management infrastructure.  There are several reasons for services to grow wildly before they are properly managed to globally scale and interoperate with other services:

  1. There is a cost involved both in time and in resources to include the management aspects in the design of services.
  2. It is not certain whether the services will be successful or not
  3. The profit margins and management affordability are not clear in the beginning
  4. Time to market pressures force short cuts that become burdensome when the service becomes wildly successful
  5. Evolutionary product and technology changes add layers of services and management infrastructures that become burdensome over time

When the services are distributed, their management becomes an important factor for service assurance to meet the latency tolerance of the service consumers.  Globalization at the speed of light makes automation of distributed services management critical to reduce the human latency and improve the consistency of user experience.

If history is any guide, one lesson that I clearly take away from studying the evolution of POTS, PANS and SANs is that the history repeats itself.  In Bell Labs, we always debated about services and services management and it always turned out that the management platforms were never on time for services to meet the market needs.  Management services were always added as an afterthought, when scaling requirements forced the issue.  Local Switching Generic Requirements came first.  Operation Support Systems Generic Requirements came next.  Finally, the Total Network Operation Processes were designed.

When Cisco first introduced the routers, I remember the lack of basic element management, let alone network management.  When SANs were introduced, I asked if Hitachi storage devices supported dynamic FCAPS (Fault, Configuration, Accounting, Performance and Security) management and people looked at me as if I was an alien.  In all these cases, a huge services industry was created to fill the management gap and it disappeared when the network element providers eventually addressed management.  Recently, when I mentioned dynamic FCAPS management to a couple of Silicon Valley Venture Capitalists (VCs) with respect to virtualization and cloud computing, they gave me a quizzical look and said that they don’t think 90% of the VCs know what FCAPS mean and were skeptical about what it had to do with clouds.  What the current generation of young VCs, the high priests who broker other people’s money for investment in innovation, miss is the fact that telephony is not about voice.  It is about networking abstractions.  It is about end-to-end service creation, delivery and assurance with telecom grade trust, universal access, massive scaling and global interoperability at an affordable price.  Interestingly, SUN got the slogan “network is the computer” right but unfortunately missed the opportunity to create a computing cloud platform before it was acquired by Oracle, thus wasting a generation of innovation and institutional knowledge.  With a broader vision, Bell Labs was able to invent the Transistor and the Laser that transformed the simple voice connectivity to what it is today.  Bell Labs is another example of wasted institutional knowledge dismantled by the short-sighted appetite of free market forces coupled with the poor decisions of a few AT&T executives, regulators, and Judge Green.

Current IT data centers have evolved to meet the business services needs in an evolutionary fashion from server-centric application design to client server networking to Storage Area Networking without an end-to-end optimized architectural transformation along the way.  The server, network and storage vendors optimized management in their own local domains often duplicating functions from other domains to compete.  The application developers also started to introduce server, storage and network management within their applications.  For example, Oracle is not just a database application.  It also is a storage manager, and a network manager besides being an application manager.  It tries to optimize all its resources for performance tuning.  No wonder it takes an army of experts to keep it going.

The result is an over-provisioned data center with multiple functions duplicated many times by the server, storage and networking vendors.  Large enterprises with big profit margins throw human bodies, tons of hardware and a host of software and shelf-ware to address their needs.  Managers in some of the data centers do not even know what assets they have and of course that is yet another opportunity for vendors to sell an Asset Management System that discovers what is available and services to provide asset management using the asset manager.  Another system is de-duplication software that finds out multiple copies of the same files and removes duplication.  This is another example of how expensive it is to clean up after the fact.

Heterogeneous technologies from multiple vendors that are supposed to reduce IT costs actually increase the complexity and management costs.  Today, many CFOs consider IT as a black hole that sucks-in, expensive human consultants, and continually demands capital and operational expenses to add hardware, software and shelf-ware.  Even for mission critical business services, enterprise CFOs are starting to question the productivity and effectiveness of current IT infrastructure.  It becomes even more difficult to justify the costs and complexity to support massive scalability and wild fluctuations in workloads that consumer services demand.  The price point is set low for the mass market but the demand is high for massive scalability.  I understand that a simple service like Facebook alone uses about 40,000 servers.

Unless the cost structure of IT management infrastructure is addressed, the mass-market needs cannot be met profitably.  The cloud service providers are justifiably looking to alternatives just as AT&T looked for alternatives to replace operator services in Telephony.  However, will history repeat itself?  In the rush to provide cloud services and lock-in their solutions, are the cloud service providers making the same mistakes that POTS, PANS and SAN infrastructure vendors made initially and added layers of service management as an after-thought?

Computing Clouds and the Last Best Hope at Crossroads

The evolution of POTS, PANS and SANs influenced the IT infrastructure to change from being server centric in 80s to being network centric in the 90’s and to being storage centric during the past decade.  I remember in the 80s, server vendors roaming the IT data centers to influence the sale.  While the differences between the servers from different vendors were minor, the sale was always made through wining, and dining of a few decision makers.  Sometimes, the new decision makers even reversed the decisions of their predecessors.  In the 90s, the network vendors and during the last decade, storage vendors dictated what the enterprise CTOs and CIOs bought.  Today, storage and its management have become the highly visible costs in the current data center.

With the introduction of server virtualization, and multi-core, multi-CPU green server initiatives, the tables have turned away from storage vendors and we are already seeing the consolidation in the industry where most of the storage companies are on the block for sale, going the way of ATM switch companies.

Amazon, Microsoft and Google in that order seem to have grasped the need for massive scalability and price point sensitivity.  They have successfully exploited server virtualization (some more than the others have) to create cloud environments at the right price points.  They have just started to address availability, performance and security issues.

Today, Cloud Computing (public or private or hybrid) offers the last best hope for saving the enterprises from the IT black hole.  However, unless cloud providers address telecom grade trust and provide end-to-end application visibility and control including business customer mission critical applications, clouds will be relegated to non-mission critical archiving, e-mail (mission critical enterprise e-mail will still be on Exchange) and social networking applications where customers are price sensitive and tolerate occasional downtime. In POTS, end-to-end visibility and control was provided through the automation of operation and business support systems.  The Internet through IETF (Internet Engineering Task Force) request for comment process incorporated end-to-end network operation and management automation.  Unfortunately, in the IT data center, end-to-end application (CPU to Spindle) operation and management is provided by an evolutionary patchwork of systems sold by multiple, server, network, storage, and management software vendors focusing on promoting their own products, which address only piece parts of the solution.   While all vendors pay a lip service to the standards by promptly including them in their roadmaps, they have not helped in reducing the complexity in inter-operability management.  In POTS, PANS and SAN implementations, we repeatedly learnt that the end-to-end Total Cost of Ownership (TCO) is not optimized by optimizing the Cost of Ownership of piece parts independently. This is what Total Quality Management was trying to teach us.

Japanese management should know this better because of their preserved institutional memory of the lessons learnt from the evolution of POTS, PANS and SANs.   Server, network, storage, telecommunications and software expertise is preserved in the Japanese institutions.  On the other hand, in the US, the dismantling of large R&D labs such as the Bell Labs, reduced government R&D, the demise of companies like US WEST and SUN, and the emphasis on startups to create innovation (with the disadvantage of starting from scratch with limited individual knowledge that the new teams bring) have all contributed the duplication, complexity and increased cost of current IT data centers.  Youth, inexperience, quest for quick profits from investors, loss of long-lasting institutional knowledge dismantled by unbridled free-market forces and a culture of making dust without concern on its impact on the universe, limit the current American Management, who depend solely on the culture of individualism, to benefit from the lessons of the collective past.  As the master said, those who ignore the lessons of history are condemned to repeat it [7].  US WEST later was acquired by a long-distance service provider Qwest whose CEO ended up in Jail.  All the institutional knowledge and the investment in innovation went to waste.  It is unfortunate that the rest of the world has to live with the consequences based on the decisions of a few that control money and decision-making. It is unfortunate that the innovation and hard work of many that bring out the best of American individualism is squandered by the decisions of a few paying attention to mostly self-serving short-term interests.

As Professor Ballon [8], pointed out in my class on current issues in Japanese management twenty years ago, waste, duplication and occasional mismanagement are of no visible consequence in a culture that promotes individualism when the resources are abundant.  When the resources start becoming less and less abundant, the group as a whole cannot afford increased entropy.  Similarly, in a collective culture, as resources start becoming more and more abundant, the group can afford a little waste, redundancy and individualism to foster innovation.  Groups usually decide where they stand in this spectrum of possibilities from unbridled individualism to stifling collectivity and adjust to optimize their chances of success both as an individual and as a group.  Success seems to lie in the middle.

The Fujitsu management, I knew, twenty years ago had a long-term vision to build global relationships and develop co-prosperity.  Dodo San twenty years ago, on a winter night, over another bottle of warm sake from Akita, told me that business is like sailing.  “Current wind speed is not your main concern.  You carefully watch for the rate of change of wind speed and direction.  You anticipate the future based on your observations at the present moment with intense focus to make your decisions and act.”  I do not know sailing, but what he said made a lasting impression just as the Zen master hitting the bull’s eye did.  Is the current Fujitsu management watching the rate of change and making its goals such as selling 500,000 servers anticipating the future?  Has the Japanese management learnt from their misses of the past?  This is the subject of my current study and I would like any one with appropriate insights to participate.

References Used

[1] Jackson S Morisawa, “The Secret of the Target”, Rutledge & Kegan Paul Inc., 1988, p35

[2] Yamada Shõji, “The Myth of Zen in the Art of Archery”, Japanese Journal of Religious Studies 2001 28/1–2

[3] Lama Surya Das, “Mind is Mightier than the Sword: Enlightening the mind, and Opening the Heart, Doubleday Religion, 2009, p 168

[4] Rao Mikkilineni, Greg Pugmire, “The Connection Between Profit and Services in the Next Generation Network”, Annual Review of Communications, volume 54, 2001, p 521

[5] http://www.mfino.com/

[6]  https://scalr.net/login.php

[7]  http://en.wikiquote.org/wiki/George_Santayana

[8]  Robert J. Ballon, “Human Resource Management in Japan”, Issue 23 (Vol. 12, No. 1), June 2002, pp. 5-20.  Robert Ballon is a professor Emeritus in Sophia University, Tokyo.  He has written many articles and books on Japanese management, role of individualism and collectivism in business management.  His lectures inspired me to write my thesis under his guidance in 1990 titled “Current Issues in Japanese Management – “Is Japanese Software Thrust As Powerful As Their Hardware Thrust?”

[9] Kenneth Kushner, “One Arrow, One Life: Zen, Archery, Enlightenment”. Tuttle Publishing, 2000

[10] E. Herrigel, “Zen in the Art of Archery” New York, Pantheon, 1971”, Routledge & Paul Keagan Inc., New York, 1988.


 The moon glows the same:

It is the drifting cloud forms

Make it seem to change.

Attributed to a Japanese Priest [1]




Hot Springs and Hadaka no Tsukiai

Most memorable part of doing business in Japan during early 1990 was engaging in a heated discussion on the future of Japanese software industry while moon gazing in a very hot rotenburo (an outdoor hot spring)  in Kusatsu, with some Japanese colleagues,  and, drinking warm sake on a cold winter night.  The discussion was about a thesis that the Japanese manufacturing prowess will catapult Japanese companies into leadership in software.  While process innovation and human network management in a factory approach has helped develop, with high efficiency, a host of software such as compilers, databases and telecommunication software where standards are well developed, Japanese management had difficulty in rapidly adopting new innovations such as object oriented programming and graphical user interfaces.  Some of my Japanese colleagues wanted to find ways to transfer software innovations faster as they did with hardware innovations. Later this would translate into Japanese foray into object databases, video service networks and service management systems with 3D graphical user interfaces and so on.

Tsuki mi (moon gazing), Onsen (hot springs) and collectivism in Japan are closely woven into the fabric of Japanese culture. According to Scott Clark [2] “Japanese have transformed what was once considered a mundane though important and pleasurable practice into an expression of Japaneseness and tradition.  The bath and bathing are today both a familiar daily practice and an act of exoticism, the transformation of the ordinary into the extraordinary in a reflexive discourse on being Japanese.”  Japanese use sake and hot springs to build individual and collective relationships.  As Kuniko Miyanaga describes [3] “In Japan, two strangers, when introduced, go and drink together as a means of investigating one another.  Using alcohol as an excuse, they become outspoken about themselves, ask questions of each other, and, in general, deviate, from normal etiquette, by “exposing their bare selves.”  This is called hadaka no tsukiai, or having a naked relationship without formality or social ornamentation.  In doing this, the two have entered into an extremely informal situation; an important social rule is that such drinking occasion remains separate from all other occasions.  The two people involved remember what has transpired during the drinking occasions, but they do not reveal this when they have returned to more formal settings.  The mutual exploration will continue on the next informal occasion, again by drinking together.  In this way, the process of building intimacy is accelerated.”

Hot springs have played an important role in Japanese business.   A hot spring hotel in Hakone called Hotel de Yama, in a picture post card setting with Mount Fuji and Lake Ashi in view from its beautiful gardens, became very famous in the year 1950 by being associated with Dr. Deming’s lecture given to almost 80% of Japanese business leaders during that period [4].  That historical event changed the course of management in Japan and the rest of the world through the Total Quality Management (TQM) movement that followed.  In the mid 1990s, when we were implementing TQM at US WEST to improve telecommunication business processes, I appreciated its far-reaching impact in shaping the culture and behavior of management itself.

De Yama Hotel

L’Hôtel donne sur le lac Ashi


The Hakone hot spring in Hotel de Yama hosted the most powerful hadaka no tsukiai of the century that changed the course of history.  According to Jerry Bowles, [5] “A more likely reason for Ichiro Ishikawa’s Deming dinner is that he wanted Japan’s new industrial leaders to hear, from this tall, loud, terrifying gaijin–a slightly derogatory word used to connote anyone who isn’t Japanese–what has been Deming’s central message for the past 60 years;  that they–management–were the problem, and that nothing would get better until they took personal responsibility for change. And on that score Deming delivered.”  Ichiro Ishikawa not only founded JUSE but also chaired Keidanren, the country’s largest trade association, which exerted great influence in post-war Japan.  Deming’s message to substitute educated human power to compensate for lack of access to raw materials was music to the Samurai business leaders attending Deming’s lecture.

An interesting email, in 2007, from Tedd Snyder from Wisconsin, [6] points out the irony of cultural changes in Japan since 1950.  He had hoped to see the actual location where Dr. Deming met with the Japanese leadership. Unfortunately, the hotel staff that he spoke with was unaware of this aspect of the hotel’s history.  Mr. Snyder was curious whether any pictures exist from Dr. Deming’s sessions at the Hotel de Yama.  I am sure the Union of Japanese Scientists and Engineers (JUSE), which organized the conference in Hakone has preserved the historical memorabilia but the significance seems to have been lost at the hot spring site.  I wonder if Deming and Ishikawa discussed the importance and the significance of that event during a hot bath. What were the discussions that followed the meeting during dinner conversations?  Whatever they were, the event had a significant impact not only on the Japanese economy but also on the rest of the world.

Current Issues in Management

The world is a different place today.  There is a debate on the role of management in both Japan and the US.  The deep and painful “Lost Decade” Japan endured and the recent Wall Street “Melt-down”, both have in common, the busted real-estate and stock bubbles resulting from management failure.  Collusion between government policy makers and big business interests, loose credit policies, lack of regulatory oversight from government bureaucrats and short term easy profit pressures from stock market gamblers have all fueled a move toward deep scrutiny of the role of free markets, managed economies, individualism, collectivism, private greed and public good.  With their cultural biases toward individualism and collectivism respectively, USA and Japan are reacting in different ways.

Japanese people reacting against the government policies, bureaucrats and status quo chose to throw the ruling government out.  The new government has promised to focus on people’s interest as opposed to the business interest. A recent article by DPJ leader Yukio Hatoyama sharply criticizes the U.S. business model for growth that Japan had emulated during the postwar period, and promises a Japan free of what he calls the “unrestrained market fundamentalism and financial capitalism, that are void of morals or moderation” to better protect the finances and livelihoods of the Japanese people.  As Japanese population is graying and the employee pool is decreasing, Japan has started to expand its borders through outsourcing. Japanese businesses are embracing outsourcing to India, China and Vietnam in very large scale.    As exports start decreasing, business focus is turning inward as opposed to global ambitions.  R&D efforts have drastically reduced which is in total contrast to the days when Japan believed in increasing R&D budgets during recession.  The mood in R&D labs, where institutional knowledge exists to meet the challenges of the new world, is very depressing unlike that which prevailed in the early 1990s.  Participation of Japanese engineers in international conferences is noticeably less compared to the 1990’s.

USA has reacted to the recent crisis by the government taking a more active role than ever before. There is also growing momentum toward retracting from large outsourcing forays and focusing on new green, healthcare and IT initiatives inside the US.  Businesses are taking minimal risk. Even the Venture Community, whose raison d’être is to focus on high potential, long-term ventures, are focusing more on short-term sure-shot revenue generating ideas. One VC in Silicon Valley pointed out that the VC community is frozen with fear focusing more on incremental less-risky solutions rather than long-term solutions with potential 10X opportunity even though they think that the idea will happen in the long term.

While the first and second largest economies are struggling with tighter resources, and figuring out better management strategies, the evolution of networking technologies from POTS (Plain Old Telephone System), Internet based PANS (Pretty Amazing New Services) [7], and SANs (Storage Area Networks) to “Computing Clouds” is presenting a unique opportunity to radically improve the next generation IT infrastructure.  It is a new technology opportunity that requires convergence of telecommunication Next Generation Network (NGN) and IT infrastructure. Such an infrastructure will fundamentally alter the way executable business workflows are globally designed, developed and distributed, and radically improve the service economies worldwide.  The scope of such transformation exceeds the TQM impact that Deming and Ishikawa facilitated at Hotel de Yama.

The purpose of my blogs, is to suggest that the Japanese management today is at crossroads, very similar to the situation when they faced Deming in 1950 in Hakone.  Today, they have the accumulated institutional knowledge that they have preserved from their forays into POTS, PANS and SANs. They are equipped with the human resources and experience to create convergent computing cloud infrastructure, which requires both NGN expertise and current IT infrastructure hardware and operating systems expertise.  For example, large companies such as Hitachi, Fujitsu and NEC have both telecommunications and IT R&D groups that they have preserved while such institutional knowledge has been dismantled in the US by free market forces.  Another advantage these companies have is their hardware, systems and  human network management expertise.  This is in contrast to US companies that are making headway in cloud computing today such as Amazon, Google and Microsoft, which are purely software companies.  History shows that networking technologies evolve with a combination of hardware and software technologies making tradeoffs to meet the requirements of massive scaling and global interoperability.  My analysis suggests that a paradigm shift with 10X improvement can be achieved with slight modifications to server and storage devices to replace the current system administration model with real-time dynamic device management.  With these modifications along with NGN switching and mediation concepts, the computing clouds will be able to provide massive scaling, global interoperability and telecom grade “trust”.

However, the cultural, political, and economic changes that have occurred since the days of Deming and Ishikawa are profound and I would like to pose the question “Is current Japanese management up to the challenge?”  Will they be able to break the organizational barriers between telecommunications, server, network and storage silos within the same company and develop the convergent architecture?  They failed to do so in the past in bringing their telecommunications and computer divisions to work together and successfully compete with Cisco’s foray into the telecommunication-grade router market.  Will Fujitsu, Hitachi and NEC be able to collaborate with each other and service providers such as NTT to create the required mediation and management interface standards for the convergent Computing Clouds?  Will the Kasumi ga Seki cloud project provide a good opportunity to create and test the ideas of convergence.  This is the theme of my study, as a sequel to my thesis published in 1990 on current issues in Japanese management. I am looking for input, comments, and collaboration. I invite anyone with appropriate insight to participate.

In order to understand the requirements to make a profound change beyond the current incremental approaches of the Silicon Valley VCs and established server, networking and storage vendors with their vested stakes in the status quo, we need to examine the history and the evolution of networking technologies.  We also must acknowledge the economic and political realities that drive management behavior.  As Deming and Ishikawa, demonstrated, true leadership that causes paradigm shifts and enables global transformation by decreasing global entropy, usually transcends the political and economic realities and identifies true value for all the participants they rally to their cause.

From POTS, PANS and SANs to Computing Clouds – Current Issues in Information Technologies Management

POTS, PANS and SANs provide a fascinating study of the power of networking, scales of economies and the impact of global interoperability and massive scaling.  The evolution of switching, transmission and access technologies associated with POTS, PANS and SANs, the politics of their penetration and global impact on the economies of nations have taken different routes but with equally profound economic consequences.  POTS altered the communication landscape by connecting billions of humans anywhere any time at a reasonable cost.  It provided the necessary managed infrastructure to create the voice service, deliver it on demand and assure the connection to meet varying workloads and individual preferences with high availability, optimal performance and end-to-end connection security.  The service assurance set a standard known as “telecom grade trust”.

The Internet enabled the connection of billions of computing devices using the Internet Protocol (IP) network and enabled PANS that support today not only businesses but also a large consumer segment through the World Wide Web.  The growth of e-commerce and social networks are a direct consequence of this evolution with global impact.

The resulting explosion of data put a strain on the Information Technology (IT) infrastructure and the management of petabytes of data required a networked shared storage strategy that was addressed by the storage networking introducing custom hardware and software technologies.

With the advent of multi-CPU, multi-Core servers and server virtualization technologies, a new networking architecture called computing cloud is evolving that allows shared CPU, memory, IO, bandwidth and storage via the Internet or the Intranet to create and deliver services on a massive scale.  Virtual servers allow computing and storage resources made available to service developers for developing and delivering business and consumer applications to their customers using the computing clouds.  The computing cloud technologies are just evolving with several benefits already demonstrated but also many enhancements that are still lacking to make them massively scalable and globally interoperable with telecom grade trust [8].

Telecom grade trust today is the result of a long history of evolution of technologies and processes starting with AT&T.

According to AT&T [9], during 1894 to 1904, over six thousand independent telephone companies went into business in the United States, and the number of telephones boomed from 285,000 to 3,317,000. Many previously unwired areas got their first telephone service, and many others got competing companies. However, the multiplicity of telephone companies produced a new set of problems — there was no interconnection, subscribers to different telephone companies could not call each other. This situation only began to be resolved after 1913.  Theodore Vail in 1907 the then President of AT&T believed “that the telephone by the nature of its technology would operate most efficiently as a monopoly providing universal service. Vail wrote in that year’s AT&T Annual Report that government regulation, “provided it is independent, intelligent, considerate, thorough, and just,” was an appropriate and acceptable substitute for the competitive marketplace.”

From those monopolistic beginnings of AT&T to its remaking into today’s at&t through cycles of regulation and deregulation (accelerated by down-right fraud from its competitors such as MCI Worldcom), much has changed, but two things that have remained constant, are the universal service (access on a global scale) and the telecom grade “trust” (providing reliable, secure and high performance connection at a reasonable cost) that are taken for granted.

The Internet on the other hand evolved to connect billions of computers together anywhere, anytime from the prophetic statement made by J.C.R. Licklider [10] “A network of such [computers], connected to one another by wide-band communication lines [which provided] the functions of present-day libraries together with anticipated advances in information storage and retrieval and [other] symbiotic functions.”.  Starting with three computers connected in 1969, the network grew to 213 by 1981.  The IETF was formed in 1985 and used the Request for Comments (RFC) process to develop and promote the standards that drove the growth of the Internet transforming the “send and pray” network to become reliable, secure and high performance network.  However, government support from DARPA and the telecommunication Act of 1996 played key roles again in nurturing and fostering innovation.

Storage networking and resulting NAS and SAN technologies have changed the dynamics of the enterprise IT infrastructure in a significant way to meet business application needs but were not able to meet the cost constraints dictated by the mass services market.  Wild fluctuations in the workloads make it impossible to provision for meeting peak-load requirements cost effectively.  This created a new class of service providers such as Amazon, Google and Microsoft to abandon the expensive and management intensive SAN strategy to develop an alternative storage strategy using commercially off the shelf (COTS) hardware and distributed web based service oriented software architectures.  The resulting divergence between IT infrastructure supporting Internet based consumer services such as social networking, email and video streaming applications and mission critical business applications requiring high performance and low latency tolerance is currently providing a new opportunity to reexamine the economics of IT management.  The divergence between the mass-market service infrastructure and traditional business application infrastructure took a wider turn with the introduction of server virtualization that allowed dynamic server provisioning, application migration, dialup, and dial down of application performance to meet wildly fluctuating workloads.

cloud evolution

Figure 1 shows the two divergent paths.  While the business application infrastructure (shown in the left) provides visibility into application availability, performance and security, the control is still labor and knowledge intensive.  The human latency in diagnosing and addressing application specific workload variations and varying business priorities often cost hundreds of thousands of dollars in software and service costs [11].  The silo approach the vendors provide is often called the Viagra approach with point solutions propping up old rickety systems that are not meeting the application centric management requirements demanded for mass scale deployment. Many of the software management tools end up as shelf-ware and service consultants swarm customer data centers to provide measurement, correlation, diagnosis and implementation services to optimize resource allocation to applications and address any contention issues with shared resources.  Often, best practices provided by server, network and storage vendors only optimize resources within their Silos.  These solutions often conflict with end-to-end optimization goals and business priorities that are different for different applications.  Diagnosis becomes even more complex (thus increasing the human latency) with virtualization layer in the server and virtualization in the storage to mask heterogeneous management systems and vendor devices.  While the Viagra approach provides recurring service revenues for equipment, software and service vendors and allows them to sell point solution software, it also increases the cost of visibility and control for the data center operators.  The Return on Investment (ROI) and Total Cost of Ownership (TCO) calculations the vendors demonstrate using their calculators often, drive local optimization in their silos while resulting in end-to-end optimization cost increases.

In addition, the server virtualization introduced to take advantage of multi CPU and multi Core servers to reap the benefits of energy conservation and better utilization with consolidation, adds another layer of I/O virtualization that neutralizes all the optimization strategies that expensive SAN storage vendors provide.

For these reasons, the new cloud providers have opted for cheaper Commercially Off-The-Shelf (COTS) hardware coupled with software approaches for simpler databases that are distributed, and service oriented software development environments for creating web based service applications.  However, while this approach reduces the management complexity, it does not address the visibility and control required to manage shared resource environments with telecom grade trust.  Application specific availability, performance and security management are wanting and piece meal solutions are being attempted which will again lead to similar issues with increase in complexity.

Recent announcement of virtual private clouds by Amazon is a first step in the right direction by allowing secured access to their virtual resources and infrastructure.  However, it addresses only access issue to hosted virtual servers.  For a computing cloud to go beyond being a pool of hosted virtual servers, and provide telecom grade trust at application or service level, it must address end-to-end connection management of computing resources from the CPU to the spindle.  While virtual servers have provided consolidation, flexibility, mobility, performance management and Disaster Recovery (by instantly replicating virtual servers) taking advantage of the new multi-CPU and multi-core physical servers, it has:

  1. Reduced visibility within a virtual server, through layers of virtualization increasing indirection between logical and physical resources,
  2. No control of application to spindle resources (dial-up or dial-down resources on demand) within each virtual server at run time to address changing workloads and business priorities, and,
  3. Increased cost and latency of diagnosing and resolving shared resource conflicts between multiple virtual servers

One question is which vendor cloud approach and lock-in is preferred?   As pointed out by Tim Bray [12],

“The small problem is that we haven’t quite figured out the architectural sweet spot for cloud platforms. Is it Amazon’s EC2/S3 ―Naked virtual white box‖ model? Is it a Platform-as-a-service flavor like Google App Engine? We just don’t know yet; stay tuned.”

Virtual server sprawl and their management have become more complex than physical server farm management. No cloud vendor has addressed virtual server sprawl management by providing visibility and control to assure telecom grade trust to both service development and service delivery.  Three factors raise two important questions:

  1. The inability, today, to easily dial-in or dial-out a virtual server from one cloud to another with telecom grade trust,
  2. The inability to dial-up or dial-down resources used by a single virtual server at run-time based on workloads, and
  3. The lack of visibility and control to resolve contention between shared resources dynamically, among many virtual servers, based on business priorities.

When management issues are accounted for, is the cloud (whether private or public) using current virtualization technology that depends on a whole virtual server replication and or migration, cost effective?  Are there better alternatives?

Paradigm Shift to Eliminate Waste in Information Technologies Management

While current state of the art has not yet assured confidence with telecom grade trust, if approached correctly, I believe that computing clouds present the last frontier in information technology revolution that will reduce complexity, provide global interoperability and telecom grade trust to networked computing resources.  Based on POTS, PANS and SAN experience, it is easy to see that reducing waste and providing telecom grade trust for accessing computing resources globally will have profound economic consequences.

The key to a paradigm shift is to recognize that the current administration and management paradigm that originated with server architecture is static and assumes that the resources (CPU, memory, bandwidth, storage capacity, throughput and IOPs) are allocated to an application at install time.  Changes to workloads and business priorities are assumed to occur at longer time scales, and administration can be performed off line.  This assumed that there were maintenance and administration times that are scheduled and the services the application provides can be interrupted during this period.

With high-speed networks and global connectivity, this assumption broke down and many management systems were added on to improve, the availability, performance and security as shown in the left side of figure 1.  However, the cost and flexibility required for mass scale deployment of services with wildly fluctuating workloads and business priorities has created the alternative computing cloud architecture on the right side of figure 1.

However, even the cloud architecture shown in figure 1 (on the right hand side), still assumes a static administration paradigm and still requires virtual servers to be provisioned at install time.  Configuration cannot be dynamically changed based on changing application workload profiles or business priorities at run time.  This makes application specific performance management, and storage administration very cumbersome, labor and knowledge intensive.  For example, when there is resource contention, one cannot stop one low priority application and divert its resources to another application with a higher business priority without disruption.  The complexity compounds when you have clustered servers (virtual or physical) with redundant paths to clustered storage devices.

Opportunity for New Approach to Develop Computing Clouds

Deming and Ishikawa taught us that removing waste and implementing far-reaching changes in an organization requires management collaboration that transcends organizational and geographical boundaries.

Return on investment must address not only the impact on share holders but also on impact on employment, customer intimacy and loyalty, impact on the ecosystem of suppliers, generation of tax revenues that facilitate better social fabric with national and global harmony and reduce entropy  through waste elimination.  The Japanese management understood this message very well when they started implementing TQM on a national level and global level.  Their emphasis on harmony (Wa),  Kyozon Kyoei (growing together) and their long term focus helped them become the second largest economy.

It is also interesting to note the contrast between 1950 and 2010.

In 1950, educated human resources were suggested to compensate for lack of access to raw materials at that time.  In 2010, the human resources are at a premium (especially in Japan caused by aging population and low birth rate) and it makes economic sense to eliminate IT management latency through automation that goes beyond incremental approach.  By creating the next generation server, network and storage infrastructure with dynamic resource control, the need for massive outsourcing is eliminated while creating a new paradigm for next generation computing clouds that will radically transform the services economy worldwide.

This will be analogous to Strowger’s switch eliminating many operators sitting in long rows plugging countless jacks  into countless plugs and reducing the cost of adding new subscribers that had risen in a geometric proportion. According to the Bell System chronicles, one large city general manager of a telephone company at that time wrote that he could see the day coming soon when he would go broke merely by adding a few more subscribers [13].  The only difference between today’s IT data center and central office before Strowger’s switch is that “fewer, but very expensive consultants, countless hardware appliances, and countless software systems that manage them” replace “many operators, countless plugs and countless jacks”.  In addition, we have to account for the shelf-ware and the human latency involved.

More recently, in a speech at the NASA Ames Research Center in California, federal CIO Vivek Kundra said that the government cannot continue to invest in traditional data centers to support its IT needs, citing a doubling in the energy cost at federal data centers between 2000 and 2006.  Of the $76 billion the government spends annually on IT, $19 billion he said goes toward infrastructure maintenance.

Both US and Japanese governments recognize the problem and are sincerely looking for solutions.

The question is  whether today’s Japanese corporate management and Silicon Valley VCs are ready to mobilize and remove waste and replace over-provisioning of resources with on-demand provisioning in current IT infrastructure with innovative solutions or continue their incremental approaches? Can they create next generation servers & storage to implement telecom grade trust through computing clouds with 10X improvement?  Who is the next Ishikawa with a vision to transform Japanese business in the time of need?  Will the changing political clouds in Japan nurture the seeds of innovation or will it take the planned economy such as China to drive the next wave of innovation?  Or will the Silicon Valley VCs beat them?  Is this an opportunity for a collaboration between individualism and collectivism to establish a balance  between corporate greed and public good, vendor-lock-in and choice fatigue with too many options, and unbridled free markets and fully regulated monopolies?  Did the Japanese management have a better working model with their collective approach before they followed American capitalism “that is void of morals or moderation” as Hatoyama put it?  Would the computing cloud also by the nature of its technology, operate most efficiently as a monopoly providing universal service with telecom grade trust, under government regulation, as Theodore Vail put it, “provided it is independent, intelligent, considerate, thorough, and just?”.  Who will lead the effort to tame the last frontier of computing clouds in Information Technologies by taking the risk and fighting current monopolies?  Is it the Silicon Valley VCs, Japanese Samurai business leaders or emerging technocrats in China?

Having participated in the POTS, PANS and SAN revolutions both in US and in Japan, I am keenly interested in watching how the current cloud revolution will play out.  One sure thing is that it takes a revolution – not an evolution to implement the productivity improvements that are possible and are absolutely essential.  If history is any guide, AT&T resisted digital switching and fiber optics revolutions and lost its technical leadership.  Japanese companies resisted IP networks preferring ATM (Asynchronous Transfer Mode) switches and missed leadership in Voice over IP revolution.  COBOL programmers resisted Object Oriented Languages which eventually led them to their extinction.  SAN franchises will resist commoditization of storage.   Server vendors will resist modifications to current virtualization strategy claiming that unified computing has arrived in bundled server, network and storage blades supporting pools of virtual servers with virtual I/O without addressing the management complexity issue.  Browser vendors will lead you to believe that the browser is the true distributed cloud operating system, Silicon Valley VCs will fund its research & development and Wall Street pundits will confirm it on Cable shows.

It will take another Deming and Ishikawa to convince management to go beyond their current franchises to harness the synergy of POTS, PANS and NGN and make real progress to reduce the true cost of IT management and reduce the entropy of the universe.  This means replacing current form of management intensive SANs and creating next generation servers and storage devices with dynamic FCAPS management and real time mediation.  This means creating real network centric distributed computing architectures and operating systems.  This means simplifying current management stacks by eliminating them.  This means creating new self discovering, self configuring, self monitoring, self healing and self optimizing distributed software systems replacing current high maintenance architectures that assume static resource administration patched with a string of evolutionary management systems.

I have spoken to data center managers who will discard their current SAN infrastructure if they had a better alternative in a wink.  In fact some have already started their transition plans:

“I don’t care who provides my infrastructure …even for my mission critical applications as long as I have visibility into the cloud and have control of my application response time, I/O, throughput, availability, latency and security…and I have the ability to adjust it based on my business priority and changing workloads”  – This quote is from an IT Business Alignment  Manager at a Large Energy Company who is actively re-architecting their datacenter to move away from their current storage strategy.  This is a strong invitation to create next generation computing clouds with telecom grade trust with visibility and control by next generation service providers.  This is a strong invitation to next generation infrastructure providers to commoditize servers and storage with network intelligence and real-time FCAPS management.  This requires expertise in not only Information Technologies but also POTS, PANS and NGN.

Will No 1 and No 2 leading economies recognize the strategic imperative and act as they did when Deming pointed out that the management was the problem or will they freeze with fear and watch, and wait to outsource their IT infrastructure as new players emerge after the current recession to take leadership, and run with it?  Only time will tell.  In the meanwhile, the moon will glow the same while giving the illusion of change through drifting cloud forms to the hot spring lovers practicing tsuki mi and hadaka no tsukiai.

Plus Ça Change, Plus C’est la Même Chose!

 References Used:

[1] James David Andrews, “Full Moon Is Rising”, p32

[2] Scott Clark. “The Japanese Bath: Extraordinary Ordinary”, Re-Made in Japan – Everyday Life and Consumer Taste in a Changing Society, Edited and with an introduction by Joseph J. Tobin, Yale University Press, 1994

[3] Kuniko Miyanaga, “The creative edge: emerging individualism in Japan”,

[4] http://deming-network.org/deming_1950.htm

[5] Jerry Bowles, “obituary/tribute to W. Edwards Deming”, The Quality Executive, a monthly newsletter. January 1994

[6]  http://www.ces.clemson.edu/pipermail/den.list/2007-September/000019.html

[7] Originally coined by Negroponte from MIT Media Lab “Pretty Amazing New Services are discussed in a book by Jean-Jacques Laffont, Jean Tirole, “Competition in Telecommunications”, MIT Press, 2001

[8] http://blog.kawaobjects.com

[9]  http://www.corp.att.com/history/

[10]  J. C. R. Licklider (1960). Man-Computer Symbiosis.

[11] One company is offering a Health Check Service that will cost $25,000 to $50,000. It also plans to offer a product next quarter that will cost around $400 a switch port, plus the number of links being monitored.  http://www.byteandswitch.com/storage/infrastructure/virtual-instruments-tracks-storage-io-in-virtualized-systems.php

[12] http://www.tbray.org/ongoing/When/200x/2008/10/14/Cloudy-Times

[13] http://www.telephonetribute.com/switches.html#Mr. Almon B. Strowger and His Electric Telephone Switch

[a]Fault, Configuration, Accounting, Performance and Security management known as FCAPS has played a big role in reducing the Total Cost of Ownership and build Telecom grade “trust” in telephony.
wordpress stats

古郷や四五年ぶりの煤払 (furusato ya shi go nenburi no susu harai)

My home village–
four or five years of soot
needs sweeping

—-  A Haiku from 1824 [1]


Famous Rock Garden in Kyoto and a Zen Monk in a Learning Center of a Trading Company

A haiku (also known as Haikai) is a brief structured stanza that evokes emotional images with an element of surprise or contrast.  The intolerance against decay, the recognition of the inevitability of accumulating soot over time, the need to intervene and the acceptance of sweeping as an essential part of intervention; all these emotions expressed in this Haiku, are evident in various aspects of Japanese landscape such as rock gardens and tea houses.   They are also evident in Japanese daily culture emphasizing process and continuous improvement (kaizen).

The first time, twenty years ago, when I sat there at the Rock Garden at Ryoanji with a Japanese colleague counting the stones and watching the tracks of the rake, I kept wondering how they make  those lines everywhere and circles around the rocks.  Are there 15 stones and will you really see the 15th stone only when you attain enlightenment?  Do they rake every day?  How many monks does it take to rake the Rock Garden?

Zen influence on Japanese culture is not only visible in the Japanese gardens [2] but also both in business and every day life through:

  1. Emphasis on process in business
  2. High reliance on measurements, data and analysis
  3. Hobbies such as practicing tea ceremony (Chadou), flower arrangements (Ike Bana) and the martial arts

While Zen teaches a way to approach the actions of every day life with awareness and focus on the present moment, it is not really a religion in the western sense of the word.  In fact, Japanese approach to religion itself is unique.  It is said that the Japanese are born shinto, marry like a christian and accept Buddha when facing death and funerals.  Zen practice unlike western religions does not seek to enlist others.  Instead, the Zen masters actually discourage others from entering the monasteries.  I have heard of stories where the “would be disciple” is discouraged to join and doors closed.  Often one would wait for days at the entrance even in rain  and snow showing one’s will to pursue enlightenment when the monks finally take pity and open the door.

The Zen masters recognized the role of entropy both within (affecting human thought processes) and outside (in the relationship with surroundings.)  The concept of Wa (harmony) strives to fight the tendency of entropy, (the amount of uncertainty which remains about a system after all observable states are taken into account), to increase in a closed system.  Constant awareness (or mindfulness) to reduce random thoughts in the mind, struggle to improve harmony between self and the surroundings and the acknowledgement of the need for intervention to decrease entropy (sweeping the soot) have all seeped into both Japanese daily life and business discipline through “measure, manage and optimize” philosophy.  Kaizen or continuous improvement has transformed the way the processes are optimized in manufacturing, IT data centers and even in daily life.

While Kaizen is an equilibrium phenomenon, which attempts to reduce the entropy by increasing the knowledge through measurements analysis and intervention, innovation is a non equilibrium phenomenon striving to escape from the local equilibrium to find a lower or an absolute minimum.

A Zen Master’s quest was two fold:

  1. Find an equilibrium where the self and the surrounding are in harmony and
  2. Strive to continually maintain the balance between the self and the surrounding through awarenesss and intervention.

The amount of intervention is just enough to keep the balance at equilibrium.  As my english teacher used to say – a perfect essay is one where if you add one more sentence, it is one too many and if you take away one sentence it is one too short.

A long time ago, when I was attempting to unravel  the mysteries of classical and quantum liquids using Metroplolis and Green’s Function Monte Carlo methods, I learned that there are multiple local minima in the thermodynaic state of the system and some times you have to create chaos (in this case increase the temperature) to jump to a next local minimum (see figure).



Kaizen is very effective in reaching a minimum when the system is in an equilibrium state (local minimum).  However, the system has to go through unstable equilirium (jump the peak) in order to reach other lower states of entropy.  Innovation in process or product technologies facilitates this transition but at a price in the form of associated inconvenience caused by change.  Sometimes, the system may have to transition through a higher state of entropy before finding a lower minimum.

The raison d’etre for the Information Technologies (IT) in a data center is to provide the computing, network and storage resources required to create and execute business workflows that deliver specific services to the external world.  The external world is influenced by changing business priorities, varying workloads, environmental conditions and latency tolerence of service consumers.  The success of the data center depends on optimizing the harmony between the consumers and the supply of the required resources.  In a data center, the system entropy comprises of three components:

  1. The human latency in IT management that is involved in monitoring the changes in the surrounding and reacting to them with appropriate visibilty and control [3] (e.g., assuring application availability, performance and security)
  2. The amount of shelfware that has accumulated over time (the soot accumulated over the years) that is not only not helping to reduce the latency of response but actually comes in the way through the cost and effort required to maintain them (e.g., unused licences, annual maintenance costs, labor and knowledge intensive tasks of correlating information from multiple systems) and
  3. Lack of visibility into the impact of latency of response on business (e.g., customer turnover, loss of revenue etc.)

The IT management’s mission is to seek an optimal minimum in entropy through innovation and then to establish perfect harmony between the IT systems and its surroundings that they serve using Kaizen.

This is at odds with social and economic reality in Japan where a large number of humans depend on responding to the changes in the surroundings in the form of IT server, network and storage systems management.  In the old days, when the velocity of information was low,  global connectivity was not electronic, and Japan was an island unto itself, the human latency contribution to entropy was not a significant factor because the latency tolerance of the consumers was larger than the human latency involved in responding to the changes in the surroundings.  Today, globalization, the Internet, massive scale of social networking, broadband communications, wildly varying demand and rapidly changing business priorities are straining the current IT infrastructure, and human latency has become a significant part of the entropy contribution.  The soot, that is accumulated over time in the form of a patchwork of management systems, (evolving from a server centric IT to a fully network centric IT), designed to provide local optimization in server, network, and storage domains adds an additional overlay of people, processes, and technologies attempting to integrate the silos and contributes more entropy.    To meet the changing latency tolerance requirements of IT resource consumers, the infrastructure has to transform itself and serve the new reality.  Agility through dynamic reconfiguration, and customization have to become the norm and not an exception.  Connecting the right resources to right consumers based on their latency tolerance and business priority becomes a differentiator.  Changes in consumer demand at electronic speed mandates visibility and control at electronic speeds and no less.

The advent of switching technology in the first decade of the 20th century, eliminated the human latency in operator based manual switching in telephony introduced in the late 19th century. Will a “Cloud Switching technology” eliminate human latency in the 21st century data center?  If history is any guide, the disruption is inevitable.  One can only minimize the impact with careful design and planning.  Japan has witnessed such transformations before and has led some of them.  Japanese management has successfully managed human network connections with great organizational ability, adopted Total Quality Management,  and established leadership in telecommunications infrastructure in the past.  How will the new generation of managers establish harmony between the IT systems and their global surroundings? Will they move from Kaizen, with which they feel so comfortable, and cause planned chaos required for innovation? Or will Japan become an island unto itself again or reclaim its past position in global technology play.  Will their ambition go beyond selling 500,000 servers?  This is the theme of my study. I am looking for input, comments, and collaboration. I invite anyone with appropriate insight to participate.

[1] http://haikuguy.com/issa/search.php?keywords=sweep&year=

[2] http://learn.bowdoin.edu/japanesegardens/gardens/intro/index.html

[3] http://www.soacenter.com/?p=189
wordpress hit counter

“First snow
on the half-finished bridge.”

– Haiku by Matsuo Basho

Photo:  Path to an open air Hot Spring bath (Rotenburo)


Snow plays a very important role in Japanese life.  People romanticize snow and雪国, yukiguni or snow country in enka singing. “Yuki mi” or snow watching, gives an excuse to drink sake and ponder about the impermanence of everything and nothing. The hexagonal shape of the body and roof of yukimi lantern is designed to capture the snow on its roof, and instill a sense of purity and serenity to the viewer. The Sapporo Snow Festival, one of Japan’s largest winter events, attracts about two million people to Sapporo to see the hundreds of beautiful snow statues and ice sculptures which line Odori Park, the grounds at Satoland, and the main street in Susukino.  For seven days in February, these statues and sculptures (both large and small) turn Sapporo into a winter dreamland of crystal-like ice and white snow. Yasunari Kawabata’s “Snow Country”, captures the Japanese “kokoro” in a novel that won the Nobel Prize.

Japanese snow also features in a trade controversy as the following excerpt states:“The oldest forms of barriers are tariffs (taxes or “duties” on imports) and non-tariff barriers, such as quotas (quantitative restrictions), both of which are imposed at the border.  Other less visible trade barriers include more indirect, inside-the-border practices such as official acceptance of collusion among domestic businesses, distribution systems that discriminate against foreigners, and restrictive government procurement practices.  One of the most notorious instances of unduly restrictive safety and health standards was the onerous set of technical specifications that Japan imposed on imported skis in the 1980s on the grounds that Japanese snow was different from that of other countries.” [1].“Japanese snow” since has become a euphemism for “collusion politics” and “crony-capitalism” that have been associated with Japanese business especially during the peak of Japanese success globally.  Often, western vendors complained that their products had to be validated to meet unreasonable demands that stem more from low risk tolerance of Japanese management or high reluctance to embrace anything that they are not used to.

Collectivism and the need to bring a whole group of people together who have disparate vested interests to agree upon any action, creates a natural barrier to innovative approaches, technologies and ideas from outside. Even within a company, “groups” exist and “soto and uchi” play a very important role on how decisions are made.  The telecommunication division views computer division as “soto”.  Domestic market division looks at the global market division as “soto”.  In reality, the collectivism is practiced as a hierarchy of groups who cooperate within the group but fiercely compete with outside groups.   At any level of the hierarchy, the subgroups are loyal to the main group but fiercely compete with outside groups even within the same company.  This is one of the reasons why the institutional knowledge dispersed in different subgroups of a company becomes ineffective in delivering synergy.  This also explains why every Japanese technology company that has communications, chip and computer expertise, has not been able to leverage the synergy to successfully develop convergent technologies inside Japan.  Ironically, in the US Silicon Valley, where individualism reigns supreme, companies collaborate with each other to harness complimentary expertise to create synergy in the products they develop.

History shows the uniqueness claim has been a double edged sword.  While it may have protected some interests in the short run, in the long run, it has isolated Japanese technology from global markets and also prevented efficiencies that stem from innovation to permeate in the Japanese fabric faster.  Example of the effect of “uniqueness” is the recent turmoil in Japanese mobile manufacturing sector.  By embracing their own unique way of implementing the mobile communications, they have missed out on global expansion and are now trying desperately to catch up on Iphone revolution. Is recent withdrawal of Japanese manufacturers from mobile phone business when Iphone is creating a record in growth a direct result of Japanese uniqueness play?

In my update of the “current issues in Japanese Management”, I am investigating how the current generation of Japanese leaders embraces innovation. I am especially interested in how the current generation is approaching IT management and what the west claims to be the “next big thing” called Cloud Computing.  Japan’s Ministry of Internal Affairs and Communications, plans to build a massive cloud computing infrastructure to support all of the government’s IT systems. The Kasumi ga seki (literally means the fort of fog) Cloud, named after a Tokyo district where the IT infrastructure will be housed, will be built in stages from now until 2015. The goal of the project is to consolidate all government IT systems into a single cloud infrastructure to improve operation efficiency and reduce cost.  The major components of IT operation costs are:

  1. Computing equipment,
  2. Communications equipment,
  3. IT operations management software and
  4. Processes and People who support IT

While the equipment costs have gone down dramatically over the last two decades, the software and people costs have not.  Ironically, while their hardware revenues draw fewer margins, the vendors who sell computing and storage devices have looked to increase their revenues and profits through selling software management systems and service offerings often at exorbitant rates in the form of systems engineers, process consultants and health check experts. Unless, the escalating software (often called shelf-ware because it is not very useful and gets stored on a shelf) and services costs are brought under control through new innovation similar to what happened with telecommunications network and service management automation, consolidation of IT in Kasumi ga seki only moves the problem from many places where it exists today to one place by 2015.

True cloud computing productivity improvements will only come from innovation in hardware and operating systems that allows application software to dynamically adjust its computing, network and storage resources to meet changing workloads and business priorities, software that is self configuring, self healing, self managing and self optimizing and processes that totally eliminate human latency involved in troubleshooting and diagnosis.  More self-serving vendor driven management systems (suggesting to put other vendor products behind their’s) and processes without a fundamental architectural transformation will not lead to drastic productivity improvement. If the Japanese management is not ready for such innovation, the kasumi ga seki cloud will be only a half-finished bridge before the first fall of snow.

I am looking for input, comments and collaboration and invite anyone with appropriate insight to participate in my research.

[1] Stephen D. Cohen, Robert A. Blecker, Peter D. Whitney, “Fundamentals of U.S. foreign trade policy: Economics, Politics, Laws and Issues”, Westview Press, 2003 Edition 2, p96

According to Geert Hostede, “Individualism is on the one side versus its opposite, collectivism, that is the degree to which individuals are integrated into groups. On the individualist side we find societies in which the ties between individuals are loose: everyone is expected to look after him/herself and his/her immediate family. On the collectivist side, we find societies in which people from birth onwards are integrated into strong, cohesive in-groups, often extended families (with uncles, aunts and grandparents) which continue protecting them in exchange for unquestioning loyalty. The word ‘collectivism’ in this sense has no political meaning: it refers to the group, not to the state. Again, the issue addressed by this dimension is an extremely fundamental one, regarding all societies in the world.”

Figure 1 shows the Individualism Index for Japan, Germany and USA [1].

Individualism Index

Individualism Index

In my study of Japanese management twenty years ago, I found many arguments about how individualism and collectivism play key roles in business success.  One argument from Professor Robert Ballon of Sophia University struck me as very interesting.  His thesis is that the social characteristics of a group are related to the resource availability for the group.  If the resources are abundant, system allows individualism and also tolerates resulting duplication and resource wastage.  However, as the resources become scarcer, the group members tend to collaborate with each other to optimize available resources to meet the common goals for survival.  Even if the group starts with competition resulting in violence, eventually they learn that collaboration is a better way to assure the survival of the group as a whole.  This is especially true if the group coexists over a long period of time.  Business success depends on a balance of individualism and collectivism.  100% collectivism tends to foster tribalism and kills innovation thus reducing opportunities for business success.  100% individualism fosters cream skimming, greed, looting and eventual wastage of resources.

The quest for total resource optimization often leads to tradeoff of time against space and collectivism against individualism.  For example, in the days when global connectivity was limited, Japanese management evolved to optimize space that was a scarce resource and organized themselves to put the “group” first.  However, as the global connectivity broadened, people started to become more individualistic.  For example, playing golf in Japan is a very expensive proposition, but today, a Japanese business person can fly to Hawaii and play golf at a cheaper price.  Just-in-time inventory optimizes shelf-space.  However, it requires broadband transport to meet latency tolerance limits of consumers.

There are many theories on whether more dominant collective nature of Japanese management is conducive for success in global business in general and software business in particular which I discussed in my thesis.  However, more interesting aspect of collectivism is its ability to form human networks (humans networking as groups collaborating to achieve common objectives).  The human networks are considered intelligent because they accomplish their goals in multiple ways using information collected from the external world and using it to control it.  The human network consists of a group of individuals operating as a system [2]:

  1. Every system has a purpose within a larger system
  2. All of a system’s parts must be present for the system to carry out its purpose optimally
  3. A system’s parts must be arranged in a specific way for the system to carry out its purpose (separation of concerns)
  4. Systems change in response to feedback (collect information, analyze information and control environment using specialized resources)
  5. Systems maintain their stability (in accomplishing their purpose) by making adjustments based on feedback

According to Vancho Cirovski, [3], the effectiveness of the human network depends on the connections, communication and mastery (or specialization) of the individual human object.  Better the quality of mastery of the individual node, the quality of connection and communication, higher the effectiveness.  Human networks provide a perfect working model for distributed computing.  Malone [4] describes the distributed computing model:

  1. Organization consists of connected “agents” accomplishing results that are better than if they were not connected.
  2. An organization establishes goals, segments the goals into separate activities to be performed by different agents, and
  3. Connect different agents and activities to accomplish the overall goals.

Scalability is accomplished through hierarchical segmentation of activities and specialization. There is always a balance between the cost of coordination of the agents and economies of scale obtained from increasing the network size which defines the nature of the connected network.  Efficiency of the organization is achieved through specialization and segmentation.  On the other hand agility of an organization depends on how fast the organization can respond to changes required to accomplish the goals by reconfiguring the network.

Both efficiency and agility are achieved through a management framework that addresses Fault, Configuration, Accounting (utilization), Performance and Security (FCAPS) of all network elements (in this case the agents).  Project management is a specific example where Fault, configuration, accounting, performance and security are individually managed to provide an optimal network configuration with a coordinated work-flow.  Functional organizations, and hierarchical and matrix organizational structures are all designed to improve the efficiency and agility of an organization to accomplish the goals using both FCAPS management and signaling based arbitration of resources depending on system priorities and workload variations.

What does this have to do with Software, Computing Clouds, and Japanese Management?  Software is the electronically encapsulated human knowledge captured as workflows in executable form.  Therefore, the ability to rapidly develop software or translate domain knowledge into machine executable form is crucial to developing a competitive edge. Current efforts in developing manufacturing systems, banking systems, insurance systems, healthcare systems, and financial systems are just such efforts to translate the specific domain knowledge into executable form, establish interfaces to real world to synchronize the information and execute automated control.  To cope with the rapid increase of the velocity of information, we are depending more and more on complex software systems to manage information in many of these domains.  There are many implementations of the distributed computing model discussed above in developing workflow automation:

  1. Telecommunications network service creation, delivery and assurance
  2. The IP network and infrastructure management
  3. Transportation networks
  4. Power grid management etc.

All these implementations make use of distributed computing model to manage distributed shared resources to optimize the overall system availability, performance, security and utilization.

Computing clouds are emerging as vehicles to share distributed virtualized computing, network and storage resources to implement business workflows. Computing clouds in order to be massively scalable and globally interoperable must support dynamism and end-to-end connection FCAPS management.  Such management demands implementing distributed computing models that are as good as the human networking model.  Will the Japanese management that has successfully implemented distributed computing model in human networks, telecommunication system, be able to build reliable, scalable, secure and dynamically reconfigurable Computing clouds? This is the topic of my research and I invite anyone with appropriate insight to participate in an update to my thesis of almost twenty years ago.

[1] http://www.geert-hofstede.com/hofstede_dimensions.php

[2] http://www.thesystemsthinker.com/systemsthinkinglearn.html

[3] http://www.orgnet.com/MCO.html “Managing the Connected Organization” by Valdis E. Krebs