Which cloud you go to, depends on what you are moving. Move machines, code, processes or content. This is the clearest way to determine what type of cloud an operator is offering, ask them “what do I move to you – machines, code, processes or content”.
Alistair is qualifying “move”, in that some code might need to tweaked for the features/functions/services of the cloud, or perhaps re-written in the case of tightly coupled legacy code.
Clouds as a utility
Nick Carr, Big Switch. Common need set, not a differentiator, economies of scale. When these things happen, this pattern exists, you get a utility model.
Alistair says the real question is, “is computing like electricity?” Now referencing Chris Anderson’s to be released book, Free. These points are consistent with Alistair’s Interop talk, which I blogged about previously.
One aspect of utility is decreasing cost. Another is degree of elasticity.
Computing is not like electricity, because (according to Forrester’s James Staten) no interoperability standards, and only fits loosely coupled applications (SOA).
Other reasons from Alistair, can’t have electricity breach, electricity doesn’t have to be monitored for HIPAA compliance, electricity doesn’t lock you in (plugs, interfaces). Electricity outages have easy back-up plan, batteries and generators. Electricity doesn’t suffer from utility sprawl.
Issues with cloud computing
Alistair says “cloud bursting is nonsense”. (IBM won’t like that). It’s all about the data. The biggest cost is moving the data. It’s also a time issue. Can the data be moved quick enough to support unplanned spike? If not, data needs to be kept synched with cloud (burst destination). This is expensive. Might require all data to be moved to hosted environment connected to the cloud.
Who owns the metadata? Typically, it’s the cloud provider. This is issue when crowd sourcing results in metadata that provides value to your business.
Too much choice right now. A lot of people are sampling, but not a lot are buying. Alistair implies there will be more buying when there is less choice, because industry will be settling out.
Cloud can dump you. Cloud provider has freedom in relationship, can discontinue service even if consumers hasn’t broken terms-of-service.
Taxation issues. State of NY wants to tax transactions done in environments hosted in NY, even if business (cloud consumer) isn’t subject to NY tax. As a former retailer, I’d call this Nexus.
What is you are a Cloud? Has the ship already sailed?
If you aren’t thinking in terms of shipping containers, instead of racks, you shouldn’t be considering cloud computing to compete purely on economy of scale.
Beyond shipping containers, can you build by dam, or buy nuclear power plant.
Answer, specialize. One potential market, legal drivers that force local (geographic) data storage. Canadian government won’t store data, use applications, in US based clouds. (Patriot Act implications)
Others: offer different contract terms, user interface, support – more help to onboarding. Niche markets.
Cloud Futures – Why Clouds are Cool
Clouds are cool not for themselves, but for what flows over it.
Enterprises today think of clouds as peripherals. Something on the edge, that you plug in.
Enterprises need to think about disruptive aspects – perform tasks couldn’t before, speed up the organization, etc.
What will matter more as computing becomes less expensive:
- data synch
- portability
- always connected
- transcoding in the cloud
- situational awareness
Alistair gives example of how cloud computing, and the ability to have split-second auctions, will be the downfall of NYC taxi medallion program in favor of “just-in-time” town car arrangements.
He says this is one example of how cloud computing will change our everyday lives.
[I’m not convinced that it’s cloud computing that changes everything. I think the internet, cellular, smart handhelds, constant connectedness changed everything. The cloud is an extension of prior works. But, not all internet applications need to be cloud resident.]
Not sure I agree with his comments about the cloud not being like electricity. Electricity has to be monitored, has to be reliable, scalable, etc. The electricity service has been perfected and commoditized over the years. The cloud or computing, will get there some day but it will take a while (just as it did for electricity).
Posted by: Mike Kavis | June 22, 2009 at 02:56 PM