Is your head in the Cloud (computing)?
I’ve been having a discussion with Millo Avissar from iSkoot regarding cloud computing, and it’s major technical problems.
It is my opinion that Cloud Computing is/was sold on its technical merit, rather than business merit. Selling to technologists may work for enabling technologies — but virtualizing datacenters begins to creep into the business.
I know, from working with businesses in multiple market segments, that companies seem to have similar concerns about their data and services (particularly those businesses whose revenue is driven in part by online commerce.)
In most cases, the issues are the same as the buy vs. build considerations for datacenters vs. colocation.
There are always trade offs for either solution (cloud vs. no-cloud.) Larger cloud-providers definitely have the buying power, ability to scale, and market share; These can lead to greater creation and acceptance of standards across the board, along with stabilization of a cost model, economic and business stability, and realistic service levels on par with the requirements for business critical and business essential operations (AKA mission critical and mission essential.)
To wit, it would be reasonable to say that larger cloud-providers will become a technical platform for smaller cloud-providers; much as the datacenter providers (rackspace, verio, abovenet, L3, etc.) became the primary hosting-providers for smaller companies to launch their own hosting platforms.
By “outsourcing” their non-core capabilities to another larger provider, the smaller hosting-providers were able to grow their technical offering and provide services that were not only competitive, but in some cases frameworks that were essential for the success of other more mainstream companies
If indeed the large cloud-providers can develop, agree upon, and ratify a standard framework for cloud-computing, the smaller cloud-computing resellers and co-developers will not need to worry about the macro issues of capacity/scalability on demand, reliability, accessibility, serviceability, and inter-connectivity; The smaller companies will be able to focus on the micro issues, which may include standard development frameworks, monitoring, reporting, and overall performance that will make their specific products attractive to buyers.
Much like a web application framework, or even Java became an abstraction from the “lower level” computing technology, having a mutually agreed upon interface to “the cloud” could prove beneficial for both the cloud companies and the software developers. Companies are now afraid of “lock in” and single vendor reliance. A single development model or framework with the proper extensions could provide the necessary similarity between service providers that could ease the transition from non-cloud based operations into the cloud, and back out again, integration between multiple cloud providers, and of course intercommunication — all seamless to the end developer.
Think about it, application packages that may be purchased or downloaded, able to be deployed to any cloud provider with a single click… just like installing software to a local server.
That being said, I do not believe that any single software company can produce this necessary framework without the support of an alliance between non-cloud companies, such as Oracle, Microsoft, and IBM. [EDIT: Although even these companies are getting “into the Cloud”; See Azure as an example.]
How different is Google from Amazon? How completely alien is Joyent from HP AiaaS? The control panels are different, the access methods differ, the means of data delivery differ, service levels differ. This is very scary to any business that needs higher service levels.
On the other hand, clouds are very attractive to young companies, who don’t mind being tied to a single platform, can control cash flow by paying for what is actually utilized (pay as you go), may require infinite scalability (all you can eat), and love to be on the bleeding edge of technology (or haven’t been burned by undelivered promises… yet.) These younger companies can not only benefit from the cloud computing model, but are contributing to its growth and maturation.
We haven’t even touched upon the larger issues (which may or may not be pertinent to your business) such as data access, logging, “auditability”, integrity validation, service prioritization, service levels, and business continuance upon service disruption.
Fears realized, and further reading:
Cloud Vendors A to Z (Revised)
Don’t buy cloud computing hype: Business model will evaporate
Google Users Live By the Cloud, Die By the Cloud.
Amazon EC2 Outage Wipes Out Data
Do you feel like you’re on vacation, hearing the kids chanting: “Are we there yet?” No. “Are we there yet?” No. — while passing signs that read “If you hosted your applications in the cloud, you’d already be there!”
We ain’t there yet, bubba. Keep drivin’, the future holds promise.
