One of the most interesting things we encounter when my consulting company works with clients is their reaction to the infrastructure architectures of cloud providers. When we explain that they achieve robustness by keeping multiple copies of data on commodity hardware, rather than the traditional model of investing in expensive hardware to improve device robustness, we observe a visceral shudder in people.
When we go on to explain that designing applications for cloud environments requires architectures with built-in redundancy because one assumes the underlying resources may fail, we can see a second shudder. What, keep computing resources running at low load just to ensure application uptime? That's ... wasteful.
The notion that one should treat computing resources as cheap commodities assumed to be unreliable goes against the time-honored axioms of IT - hardware is expensive, therefore computing is expensive, therefore its use should be rationed as much as possible.
The thing is, computing isn't expensive. Servers are dirt-cheap. Network bandwidth pricing keeps dropping. And storage ... well, when a 2 TB drive costs $59, it's clear that computer vendors give us a gift every year in the form of lower prices.
Costly IT Friction
Nevertheless, we encounter resistance whenever we encourage companies to embrace the new cloud computing architectures. What they don't recognize is that the expensive factor in computing today isn't kit, it's people. Operators, administrators, finance, help desk.
The people involved in the organisational processes put in place to manage precious, expensive hardware are a huge cost and, by virtue of the fact that they're involved in the various processes of funding, provisioning, and managing that expensive hardware, are a cause of IT friction, which translates into operational languor.
I was put in mind of this while reading Andy Kessler's new book, Eat People. His book offers 13 rules for "game changing entrepreneurs." Every book of Kessler's is stimulating, and this one is no different.
I was struck by his second rule: "Waste What's Abundant to Make Up for What's Scarce." Kessler describes a conversation he had with George Gilder, in which Gilder says "What's abundant is cheap - the price signal tells you to waste it. What's scarce is expensive. Instead of using economics to allocate what's scarce, just waste something else until what you want is no longer scarce."
In later chapters, Kessler continues to discuss the challenge entrepreneurs face when creating solutions that solve scarcity issues: entrenched incumbents not unwilling to use governmental forces to skew markets, markets focused on efficiency rather than effectiveness, and user assumptions about the possibility of changing current modes of operations - all of which the entrepreneur must address to achieve success.
The Kindle Paradigm
The accuracy of his insight was brought home to me by the mode in which I was consuming it: I was reading his book on my new Kindle. The breakout success of e-readers in general, and the Kindle 3 in particular (Amazon claims the Kindle 3 is its all-time best selling product), illustrate Kessler's thesis perfectly.
Much discussion about e-readers focuses on comparing them to the incumbent solution: books. Are they as good as, can you read them as clearly, etc., etc. From my experience, they're not the same, but they don't get in the way of content consumption, i.e., reading the author's words. However, while most of the discussion about e-readers regards efficiency (do they allow reading as well as a book), that perspective ignores their major advantages.
I'm currently in Seoul, and I decided to read Kessler's book the day it came out. The odds of me finding a copy in Seoul on release day are vanishingly small. And, even if it were available, I'd have to find my way to one of the few bookstores carrying English language books. And I'd pay a significant premium.
Instead, I bought it and downloaded it in less than a minute. Amazon has cut deals with wireless providers throughout the world so that it can sell books everywhere. And, if you don't feel like spending money, there are thousands of books, both current and those in the public domain, available for free. All of these books, whether free or carrying a price, can be accessed throughout the world in less time than it takes to read this paragraph.
Moreover, the Kindle offers an advantage to the traveler. Rather than having to lug multiple heavy physical books around to ensure that sufficient reading material is available for entertainment, a single 8 oz device can carry enough books to carry one through the longest trip. And, in fact, I've noted at least one physical advantage to the Kindle: I was reading a long hardcover book recently and found it heavy and inconvenient - while the Kindle is far lighter and more easier to manage, no matter how long the book being read may be.
In essence, Amazon has replaced much of the current infrastructure of books by "wasting" computing and network power to replace the entire supply chain of physical books.
And, just as Kessler posits, the incumbent market players are resisting this trend mightily. Despite the fact that creating and distributing a digital book is vastly cheaper than printing and trucking the paper-based counterpart, publishers have resisted reducing the prices of e-books significantly. In effect, they are attempting to shield the pricing of hardcovers by insisting that the e-book equivalent be priced much closer to the commonly available hardcover price.
This leads to the absurdity that some books are cheaper on Amazon in hardcover format than in the e-book one. It's as though publishers can't quite get their heads around the thought that lower prices made possible by the cheaper form factor might so increase sales that they would end up making more money in total. In other words, publishers are resistant to exploiting price elasticity, preferring to protect their current modus operandi, despite its clear obsolescence in the face of e-books. No doubt some publisher will eventually decide to throw caution to the winds and reduce prices for e-book versions of its releases, and then we'll see the entire industry's pricing scheme transform overnight.
There's another radical threat that e-books poses, though, beyond the blunt issue of prices. The low cost of e-books reduces the barriers to publishing. One role of publishers has been to support the cost structure of hard copy printing and distribution; because it's been very expensive to get a book to market, the number of books published is rationed. The second role of publishers, therefore, has been to serve as market taste arbiters, selecting the limited number of books that could economically be justified. The reduced cost of e-books calls that second role into question. In this compelling analysis by a Kindle blog I follow, the author describes how quality books, heretofore ignored by publishers, will now be able to get to market via bypassing the rationing mechanism of publishers.
The relevance of the e-book revolution to IT organisations is this: operational processes built on assumption of scarcity have to be rethought in light of vastly changed economics. The visceral repulsion to "wasting" computing resources, and the high organisational overhead put in place to ration access to expensive computing resources must be re-evaluated in light of cheap computing and an environment in which much of the cost of computing is attributable to the rationing processes themselves.
Too many cloud computing strategies I've seen assume that agile infrastructure can be put in place without re-evaluating other assumptions and processes. This is analogous to publishers loving the potential of e-books (not least because it offers the opportunity to reduce the margin now offered to physical bookstores) while insisting on maintaining the prices that were appropriate for a physical book ecosystem. Publishers will soon wake up to the realisation that all of their operating assumptions and practices need to be rethought for a world in which much of their traditional function is called into question. Likewise, cloud computing strategies need to incorporate much more than quick provisioning grafted onto existing organisational structure and processes - not least because the potential for authors to bypass publishers and self-publish has its counterpart in "shadow IT."
Cloud computing will, ultimately, cause a revolution in infrastructure, application architecture, and IT organisation structure and processes. The cost reduction made possible by wasting computing resources will force the biggest change in IT ever. Having long trod the path of efficiency (wringing more processing out of limited computing resources), CIOs will now need to set off on a path of increasing their parent organisation's effectiveness - in essence, the entire company's productivity.
Bernard Golden is CEO of consulting firm HyperStratus and the author of Virtualization for Dummies.
To comment on this article, please email the editor.
Follow CIO on
Sign up to receive CIO newsletters.
Click here to subscribe to CIO.
Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.