I saw a fantastic article from Nancy Gohring of InfoWorld yesterday, on how “Amazon said that it would back Eucalyptus’ efforts to support Amazon Web Services’ APIs”. Great article, well worth reading in full.
For me, however, it was the a priori assumption in the first paragraph (and the headline) that really stood out:
Eucalyptus has become far more attractive to enterprises wishing to build private clouds, now that the No. 1 cloud provider — Amazon Web Services — has thrown its weight behind the software company.
I am not buying this at all.
In my experience with many enterprises actively moving to the cloud, most every large organization sees Amazon Web Services (AWS) as an aspiration, but not a preference. They tell me that they want to be like AWS, but typically only use AWS for edge cases and new developments – and typically non-mission critical applications, rather than mainstream production. At least for now. (I cannot comment on attitudes to Eucalyptus – I do not know any enterprise that is considering it.)
What’s more, in a separate (and also excellent) article on GigaOm, comments from a Public Relations Manager for Amazon suggest to me that the world’s largest cloud provider still doesn’t really ‘get’ the reality of enterprise computing:
“Many enterprises today have legacy applications and a good deal of investment in those legacy applications. This type of arrangement provides the added flexibility to more freely move workloads between those existing IT environments and AWS … [O]ver time, most enterprises will not run their own data centers.”
This thinking sounds great as a PR statement, but in the real world it discounts the clear intentions of almost every large enterprise, and contradicts almost all research data. Unless “over time” means on a geological scale, the opposite is actually true.
Amazon simply does not support the multitude of platforms, systems, and vendors that are typical of “legacy applications” in large enterprises. For many, it does not accommodate a lot of mandatory requirements for management, security, compliance, etc.
In a new article published yesterday in CIO.com, the inestimable Bernard Golden notes that for many enterprises, moving ‘legacy applications’ to a cloud environment is not practically possibly:
Cloud computing is not going to solve legacy application challenges and costs. I recently talked with the CIO of a large media company who commissioned a study of his legacy apps to determine how many could operate in a cloud environment. The results: 10 percent.
At very least, Bernard continues, IT would need to completely re-engineer most legacy applications for AWS. However, in my experience very few enterprises see a compelling need to re-engineer millions of lines of legacy code that still does what it needs to. The other realistic option for legacy applications, Bernard points out, is “shifting to on-demand SaaS applications”, not migrating to an IaaS provider.
Which leaves AWS out in the cold when it comes to this supposed movement of legacy applications.
So then, how then does “this type of arrangement” allow a large global bank to “more freely move” its SWIFT processing onto an AWS cloud? How can a mineral exploration company “more freely move” to AWS for geological data processing from a remote oil field in Azerbaijan? How can a research institute “more freely move” to a hybrid AWS cloud to run trillions of calculations a second to model 100 years of global climate change? Or any of the other thousands (millions?) of legitimate, actual, legacy (and current) enterprise use cases for non-commodity environments?
I don’t think it does.
Even for new applications, most enterprises are not showing a preference for actually using AWS; and Eucalyptus is still being roundly trounced in the market for enterprise private cloud.
This is not to say that AWS is not a viable choice for enterprise adoption. It certainly is capable of running large-scale enterprise applications in production, so long as they are engineered for the environment, and supported by capable third-party tools for security, orchestration, assurance, service levels, and so on. At CA Technologies I am working every day with our enterprise customers, encouraging them and helping them with public cloud (including AWS) adoption.
However, I do not believe that Amazon’s new alliance with Eucalyptus clears any significant barriers for enterprise adoption of public or private cloud. Enterprises that were adopting AWS for certain use cases will continue to do so; presumably any enterprise that is building a Eucalyptus private cloud will continue to do so.
But neither makes the other any more ‘enterprise-ready’ than it already was.
*btw, bad Australian joke for you: What does a Koala do at a party? Eats roots and leaves! Yeah, the Aussies Antipodeans will get it. 🙂