posted by: Jack Santos
The “Cloud” topic is much in vogue these days, and runs the risk of the hype cycle; I feel “clouds” are at a peak hype stage and ready for a big disillusionment phase. That may be so, and we may be tiring from cloud-this and cloud-that. But I believe that fundamentally, references to the cloud really occur at two levels, both of them significant.
The first level of cloud (what I would call “True Cloud”) is really virtualization gone wild. Now that most datacenters have implemented and gotten comfortable with virtualization, many are experimenting with it in such a way to maximize capacity and expedite failover or load management. Especially for organizations with large (>10,000 sq ft), multiple backup datacenters. By focusing on virtualization workloads, datacenter operators now have the flexibility to truly separate applications from the physical hardware. Amazon and Google have become masters at this and have productized their offerings. Nick Carr (in “The Big Switch”) theorizes that this is the start of the utility computing era, much like Edison power plants sounded the death knell for private company-owned power generation.
The next level of cloud is more intangible. It’s a notion that anything can be built, run, or used over the internet. I’ll call this “Cloud Think”. It may be predicated on the “True Cloud” advances (and it may not), but it certainly is a major advance in business thinking of IT: rent – don’t buy. And just as the holy grail of home ownership has become a mantra for our American generation (and to some extent the root cause of our current economic morass), for many IT folks, locally produced and operated datacenters have been sacrosanct – until now. The model is beginning to look more like the Euro view of buy vs. rent. Yes, some buy, but many rent, and that’s OK. Technological advances with “True Cloud”, and psychological advances with "Cloud Think”, now makes us realize we have that option, and that it is economically feasible.
But what worries me more is the systemic risk associated with clouds – and whether that is on anyone’s radar – in either the “True Cloud” (for raw R&D) or “Cloud Think” (for decision-making purposes) camps. Systemic risk is on everyone’s mind, since we just pulled back from the economic abyss – and still not quite sure why or how. Wikipedia's definition of systemic risk is what I’ll use:
- In finance, systemic risk is the risk of collapse of an entire financial system or entire market, as opposed to risk associated with any one individual entity, group or component of a system. It can be defined as "financial system instability, potentially catastrophic, caused or exacerbated by idiosyncratic events or conditions in financial intermediaries". It refers to the risks imposed by interlinkages and interdependencies in a system or market, where the failure of a single entity or cluster of entities can cause a cascading failure, which could potentially bankrupt or bring down the entire system or market. It is also sometimes erroneously referred to as "systematic risk".
What everyone now knows, when it comes to our complex financial system, is we don’t know what we don’t know. Those that make a living in economic research (like Nobel Prize winner Paul Krugman) anticipated significant upheaval, but are still as shocked as anyone at the speed and breadth of the recent collapse. Lowell Bryan and Richard Rumelt talk about it with McKinsey, in a recent article.
And we don’t have to go too far back in history to discover the effects of systemic risk on the most over engineered and revered system on earth: our electric power grid. The northeast blackouts of 1965 and 2003 didn’t occur in a backwater country with gum, duct tape, and bailing wire distribution systems. It did start with seemingly small innocuous events that everyone thought had been designed for, spreading wildly to affect millions for an extended period.
As a result of the recent financial collapse (or near collapse) two concepts related to systemic risk have come to the forefront: “too big to fail” , and “too interconnected to fail”. Both of these concepts are entirely appropriate for what we in IT are defining as the cloud. Burton Group research is very concerned with this and the topic of risk management. Bob Blakely often writes about it (Risk Management: Concepts and Frameworks as an example).
Bob Metcalf, the inventor of Ethernet, had to eat crow when he predicted the collapse of the internet due to systemic risk issues. It never happened. I certainly don’t want to follow in his shoes on that topic, but now that we are building the cloud on top of the internet, the systemic risk only grows greater; he may have been wrong on timing, but right on the danger of collapse. In the 20 years since he made that prediction, we have only become more dependent on our computing infrastructure. Cloud – “true” and “think” – developments promise to make that dependency even stronger, and the risk factor even greater.
If all goes according to the plan of cloud pundits, cloud systemic risk will converge with financial system systemic risk…and the stakes will be as high as ever.