Softpanorama (slightly skeptical) Open Source Software Educational Society

May the source be with you, but remember the KISS principle ;-) Softpanorama Search

Conclusions

Prev | Contents | Next

Nicholas Carr represent a interesting threat to IT as a profession, the threat which I would call obscurantist threat.  The danger here is nor only his own ignorance of technological aspects of IT masked by definite talent as a writer, but also because people who know little or nothing about IT (and that, unfortunately, includes part of the media) take him seriously because of the stylistic quality of his prose.  When media lends credibility to his ignorance in IT technology, he is able to spread naive or completely absurd hypothesis like is the case with his hypothesis that "cloud computing as the next best thing since sliced bread".

As convincingly demonstrated in this paper the problem with Nicholas Carr is not only that his initial article and subsequent books suffer from the high doze of ignorance of  IT history and lack of understanding both electrical transmission networks and the group of technologies under the umbrella term  "information technology" (or "IT"). 

The problem is also that he falsify evidence: he cherry picks historical facts to fit his needs instead of trying to provide an objective picture (his usage of electrical transmission networks analogy is a good example of his approach to selection of facts and analogies). To be more correct Carr tortures facts and stretch analogies  to  fit his fundamentalist (server is sacred, desktop is dead) Utopian vision. The central idea of the article "IT does not matter" is simply a fallacy and it is not supported by the facts on the ground... 

Proposed remedy of "in the cloud" providers is a plain vanilla Utopia. In no way events with cloud computing will unfold in the direction of moving everything on the remote, cloud-based servers. What will succeed is a hybrid approach with significant client based component because of tremendous computing power of modern laptops, netbook and other clients (it is the computing power of desktop which actually dooms dummy clients).   This presuppose a long search for the optimum, economically justified equilibrium between server and desktop components for each and every important application. At the end of the day the codebase will be refactored into two communicating (and unique for particular application area) parts:

• "Cloud-based" subsystem, which might be much more smaller then Carr predicts or event close to zero for processor intensive applications like Photoshop)
• Desktop subsystem (which might became closer to aggregation of "on demand" installs from central server for many applications (and may be even multiple OS images) as modern laptops are formidable computational devices and to discard this local power is suicidal

Also the location of "the cloud" is open to review. For example private, corporate-based cloud is a pretty attractive and rapidly evolving solution.  They actually are called remote datacenters ;-)

Also open to review is the spectrum of applications which can demonstrate better cost of ownership under pure "in the cloud" mode. Yes, it is undeniable that there are several classes applications that emerged as suitable for 100% on server SaaS approach ( email, CRM, supply chain management, corporate benefits and other HR-related applications). And several that proved to be much less suitable, for example, Office applications, especially spreadsheets, ERM (SAP/R3, etc) most of other CPU intensive applications (it is expensive and rather stupid to provide the remote power equivalent to a dual core 3GHz CPU and several gigabyte of 1.33GHz or faster memory ). 

At best Carr managed to ask several  interesting questions, but provided inferior, simplistic and by-and-large completely misleading answers. The initial HBR paper was intentionally controversial, extremely weak on facts but rich with fuzzy (and faulty like in case of electrical power network)  analogies. The latter  makes Carr a prominent representative of IT obscurantism. Subsequent books added almost nothing to the story but continue Carr's obscurantist tradition.  

This is a typical non-professional approach to a complex problem and attractiveness of Carr's book and articles is generally reverse proportional to the level of understanding of this complex technological area.  And instead of writing new books Carr probably should try to explain why in the five years since publishing of his article (very long time for such fast developing technology as IT)  utility computing has not yet become a mainstream proposition. So what has happened to it in the past five years, and why it still might have chance of getting mainstream. Fundamentally, the deals proposed by "in the cloud" providers proved to be not as exciting as they were hyped up to be.  In comparison with the situation fives years ago now there is an additional factor: availability of more or less mature virtualization solutions but this factor can be played both for and against "utility computing": businesses are more than happy to use virtualization technology within the limit of their private datacenters and get the same (or better as they do not need to feed a middleman) flexibility and efficiency gains.  Carr fails to do the job of identifying the most promising areas were IT can be profitably commoditized and moved to an IT utility service provider.  Such areas definitely exists along with areas when it can be suicidal for the company to outsource to the "in the cloud" provider. 

Carr has more significance as a blog writer then as a book writer.  The irony of Carr's position as a technology forecaster is that for the last five year since the publication of his HBR article local datacenters actually flourished and had shown no signs of impeding demise.  Also his "in the cloud" euphoria does not take into account mainframe-style social problem that are typical for outsourced IT services as well as several technical problem (with the cost efficiency and the bandwidth cost as probably the most important two).

The author uses an old tried and true "snake oil salesmen" formula: shocking title,  bashing existing situation with circumstantial evidence, red herrings, and logical leaps and then proposing something absurd as a solution. The key idea is to bank on the fact that his audience is uninformed and gullible. The author writing  doesn't scare me (especially given the absurdity of this main thesis) as much as all the praise and attention he got. The latter makes me wonder at the level of gullibility of the audience. Authors of miracle diets and "make rich fast" books should be beware of competition as well as a potential new field into which they can profitable extend their business ;-)

Both the initial Carr's HBR article and two subsequent books provides a simplistic and flawed hypothesis of where IT is heading. All three major ideas that Carr put forward in his HBR article (and he never added anything substantial to them later, in two subsequent books) are demonstratably wrong:

• Carr's hypothesis about disappearing competitive advantage from IT is open to review. He does not the possibility of moving this advantage to the higher level of integration and more customized to the need of the enterprise architecture. Hagel and Brown in their book The Only Sustainable Edge suggest that competitive advantage is the relentless building of distinctive capability on the edges (frontiers), both within an organization, and the networks it operates in. That means that organizations have to choose what they are going to excel in, as it is not possible for any single organization to excel in everything. That also means that some doze of movement software services to external providers might be healthy: something needs to be sacrificed to stay in focus. So while some parts of IT can be outsourced or done by service provider, other are not and those parts will always require IT staff organized in some form of  datacenter (datacenter=hardware+software+architecture+IT professionals). 

  There are two issues here: 

  • First issue is that the whole is always bigger then the sum of its parts so availability and the level of standardization of hardware and software does not prevent from gaining significant competitive advantage, which is many cases is long-term.
  • The second issues is that copycat adoption of technologies that Carr advocates is good only as far as there are people who understand what needs to be replicated. Such people simply might not be available without strong local IT organization. Also not all technologies are available in pre-packaged form. For example,  Amazon lures key people from Wal Mart to create its distribution system. Wal Mart sued and the case was later settled. First of all as IT is more like nerve system: like with human nerve system weakened of IT backfires in areas sometimes not directly connected to IT activity.  

  Usually IT troubles are symptom of general stagnation, low employee morale and/or sickness of the organization.  They are simply more visible. Corruption of corporate governance with subservient boards, unchecked greed and rat race for bonuses by executives are much more serious and widespread problems that undermines competitiveness of enterprises then any level of overspending on IT, if such thing exists (it might briefly existed during dot-com bubble).  Since 2000 overspending on IT is history, and the reality is high-profile corporate scandals. Just in 2002 they have wiped out hundreds of billions of dollars in shareholder value, in industries ranging from telecom to energy to healthcare.  The next wave of corporate scandals, now in financial industry is just starting. Corruption flourishes where there is no transparency and accountability and IT tends at least marginally increase the level of transparency.  But what breeds corruption is the asymmetric government policies (profits are private, but losses are socialized), the problem especially visible recently in investment banks.  In addition there are several other important factors: 

  • Individual executive's set of values.
  • Red tape and a lack of transparency
  • Moral hazard issues beyond asymmetric monetary policy and socialization of losses for large institutions ("too big to fail")
  • Widespread nepotism and neo-tribalism. The corrupt leaders tend to protect each other and install their relatives in positions of power. Also old tribalism changes into more subtle forms like graduates networks (some firms refuse to hire HBR graduates not because all of them are bad, but because they afraid of subsequent "infiltration" and creating closed clan of graduates) 

  In a sense Carr's recommendation is a recommendation similar to throwing out passengers from a sinking boat. IT ins this respect it is similar to problem with human nerve system.  I strongly doubt the Bear Sterns or Enron went down because of overspending on IT systems.  

  Even if we presuppose that 70-80% of all IT is undifferentiated (commodization) that means that at least 20-30% of all IT is highly differentiated. Also the differentiation in IT is quite different then just  differentiation of parts (it is more like Lego-style differentiation). For example, two cars may be more or less identical in terms of the components but there is no question that the buyer can distinguish a Ford from a BMW, a Hyundai from a Honda. The actual source of competitive advantage is the relentless building of distinctive capability including distinctive IT architecture within an organization and in networks it operates.

  The entire discussion about IT and its impact on competitive advantage boils down to availability of smart, highly trained professionals who are capable of broad strategic thinking and able to connect seemingly disparate ideas, systems and protocols. Without them there can be no competitive advantage. But you cannot employ just smart people. So some kind of IT organization is necessary although it might be quite different then the current structure.
   

• Benefits of liquidation of local IT (and especially local datacenters) and moving everything "into the cloud" in the name of cost cutting are questionable. We discussed some pitfalls above. There might be some value is shaking out completely "dilbertalised" IT departments via outsourcing in a manner of 'creative destruction" but such period should be brief and clearly intended in recreating IT after a brief period rather then permanently destroying it. People, preferably smart people, are the key. "Ivan the Terrible" strategy of exiling "useless" IT personnel into part of the IT which is subsequently outsourced is one possible approach that might help to preserve existing talent pool. The problem is the that people at the top should be able to recognize such talent and in many cases "fish rots from the head".

  As for the benefits of dismantling IT organization,  such a move unless done in a very limited and controlled fashion entails serious dangers as exemplified by outsourcing experience of many organizations. The cure might be more dangerous then disease: an important side effect of total dismantling of IT organization is that instantly makes a company a donor in the hands of ruthless external suppliers and contractors. Consultants (especially large consultant firms) can help but they also can become part of the problem due to the problem of loyalty. We all know what happened with medicine when doctors were allowed to be bribed by pharmaceutical companies. This situation which is aptly called "Viva Viagra" and in which useless or outright dangerous drags like Vioxx were allowed to became blockbusters was fully replicated in IT: myth about independence of IT consultants is just a myth (and moreover, some commercial IDS/IPS and EMS systems in their destructive potential are not that different from Vioxx ;-).

  The key problem is IT is its dilbertalization and outsourcing everything to service provider just move this problem to service provider level. Moving services to remote "in the cloud" service providers (or even local outsourcers) does not solve this problem as the latter also are not immune from the disease.  It might be even more pronounced in such an environment.
   

• "Everything in the cloud" software services as  the "bright future of IT" is an unrealistic approach.  There are several competing technological trends which might at the end of the day command higher market share that "plain vanilla" remote computing over WAN ("in the cloud" software services providers). Also the power of modern laptops is such that remote datacenters are now in serious economic disadvantage as for their ability to provide comparable level of computing power remotely. As an advocate of utility computing, Carr is far from being convincing as he ignores the problem of "bandwidth communism" as well as social problems in centralized utilities which are essentially identical to the problems in large local data-centers with the same or higher level of bureaucratization, neglect of customer needs and excessive costs. the latter omission looks really glaring as with all his technical incompetence he has some sociological talent as is clearly visible in his blog.  Anyway, the reality is that utility computing is still too new and too immature to be the mechanism using which a sizable enterprise can built world-class IT infrastructure. It also can be more expensive as costs include not only direct costs paid to outsourcer but also additional costs the enterprise suffers from due to necessity of additional control mechanism, additional security mechanism, loss of operational flexibility ("everything becomes contract negotiation"), etc.

  While WAN connectivity dramatically improved due to ubiquity of fiber lines and Internet became more diverse and powerful in recent years in no way that suggest the dramatic switch to "utility computing" is under way and that the best and/or most economical way to implement now standard (mature) levels of IT technology (services as Carr calls them) is to implement them "in the cloud" (bandwidth communism). Carr predictions also looks far from being plausible from several other standpoints.  Utility computing is a viable trend but it is not absolute and has severe limitations;  as such is far from being dominant and defining trend of the last five years. Nothing suggests dramatic acceleration of this trend in the future.

  While "in the cloud" computing has its merits the artificial link of  its success to dismissal of local datacenters is just another Carr's fallacy: they are more like complementary not antagonistic technologies and each has its strong and weak points.  The hypothesis that businesses will "en mass" move mission-critical applications to "the cloud” is far from plausible.  One of the problems at the moment is economics. “On Demand” computing might be economical only for very CPU intensive operations. Although telecom prices for bandwidth have fallen and bandwidth has increased, processing power of a typical server and laptop has increased much more rapidly, which means that it's not that easy to transmit significant stream of data over the cloud. It looks like "In the cloud" service providers face serious headwinds ahead.  Competing technologies include but are not limited to application streaming, virtual appliances and standardization of datacenter infrastructure (right now represented by "cloud in the box").

  There are some clear analogy between old "mainframe-based' datacenter and new service provider approach. And user hostility to IBM in the era of mainframes was a symptom of underling problems. Something similar might happen with remote service providers attractiveness. In addition one of the major problems in utilizing remote WAN-based service providers is bandwidth problem and the cost of solving this problem depends on the type of service. That means that some services like email and Web hosting are inherently more suitable to outsourcing to providers in the cloud then others (application streaming, etc).  For high bandwidth services there is no free lunch -- you need to pay for bandwidth in order to get respectable reliability. 

His recommendations fare even worse then his key ideas which is typical for any Utopia: they are dangerously naive. Companies which follow them can definitely be hurt. Carr's article could actually negatively affect economic growth by giving CEOs justification for withholding the necessary IT investment. I would like to remind the reader Carr's proposals  which  we discussed before in more detail:

• Spend less. That's a good, generic, risk-free advice applicable to any large corporation department and any situation. Essentially, this is another way to say that penny saved is penny earned. The problem with it is that "stupidity is punishable" and "scrooges usually pay twice". You need to understand where you need cut spending and where you need to spend more; elimination of local IT that Carr advocates prevents you from seeing this and as such is a stupid solution to the problem akin to throwing baby with a dirty bathwater.  Weakening IT as akin to weakening of nerve system and can translate into more general trend and lead to "me-too" company with average products, average management, and an average future.
   
• Follow, don't lead. This is very questionable advice, as it is IT IQ that provides strategic value to the business; a large company might be save a few millions by using copycat technologies but as each business is individual enough and that strategy can backfire and at the end lead to losses not gains as the product adopted can interfere with the established businesses process and decision making.
   
• Focus on vulnerabilities, not opportunities.

  This is really a very dangerous advice as Carr completely, fails to understand that value of sound architecture in security and disaster recovery. Right now too many companies are focused on issues of vulnerabilities and compliance, but it is strategy and operational mistakes destroy more shareholder value. As one industry study states [New Report Reveals Causes for Shareholder Value Destruction]):

  "Only 13% of the decrease in shareholder value in these companies resulted from compliance failures. Sixty percent of the value destruction was attributable to strategic mistakes, such as misjudging customer demand or competitive pressure, or management ineffectiveness.

  An additional 27% was due to operational blunders, such as cost overruns or poorly managed integration during mergers and acquisitions."

Contrary to Carr's diagnosis the key problem with current IT is not excessive costs or underutilization of equipment (another Carr's fallacy): it is disconnect from the interest of the organization due to strangulating and corrupting everything bureaucracy -- dilbertalization of IT. It might be a sign of maturity in best "utilities" style but it is a pervert sign much in the of "mature socialism" Brezhnev-style.

What is implicit in all this "IT does not matter" noise is compete absence of understanding of the value of architecture in It both on software level and infrastructure level. While businesses must articulate strategy and align IT with that strategy the results depend on how closely the selected architecture suite the states business goals.  to achieve that organizations need a talented IT personnel including but not limited to IT literate business leaders and business savvy, talented and loyal IT managers.

The truth is, IT departments are more like nerve system. they are not the brains that drive the business forward and they should not be. Much like the nerve system IT  connects different parts of business organization together in such a way that makes possible for them like a players in a good orchestra to function in unison.  That fact also limit the value of external consultants. Consultants don't know your business - they know the toolset, and they can excel at holding hands but not much more. Business processes can be well supported only by a flexible, evolving  architecture; The latter  (often discussed under the label of adaptability) is probably the greatest challenge in IT. Processes which are cast in stone (like in some popular CRM and ERP; also typical for "in the cloud" providers) can be more of a liability then an asset.