There’s nothing worse — well, few things worse — than picking a technology, switching your company over to it at considerable effort and expense, and then discovering you’re locked into a dead end platform.
It’s like betting your company on Beta right before VHS went big. Or betting on VHS right before everyone switched to DVDs. Or betting on… anyway, you get the idea.
In a perfect world, we’d be able to accurately predict which technology will dominate, and make our investments accordingly. But, in practice, people just wait. They sit and watch other guys place their bets, and they continue to sit and watch until a clear winner emerges.
But, what if there is no clear winner? What if you have a technology with multiple — and incompatible — approaches? That’s the situation with clouds right now. There’s the Amazon approach to cloud computing. The Microsoft approach. The IBM approach. The Google approach. And, the ever-popular VMware approach.
Picking a vendor too early can cost the company significant amounts of money down the line when it is time to switch. So after the early adopters have all adopted, if there’s still no clear winner, someone has to take a step to break the logjam. This is usually when the vendors — all except the market leader — get together and decide on a standard.
That’s what seems to be happening now in the cloud computing space. Amazon is the market leader. But its cloud platform is proprietary, which means its customers aren’t going anywhere. So, Amazon’s name isn’t popping up in any of the standards discussions, except when people discuss whether Amazon will actually be the standard. It’s like Windows was the standard for office computing for decades.
So Amazon’s competitors, including IBM, Cisco, EMC, CA, SAP, and Red Hat, have just gotten together and agreed on a cloud standard called Topology and Orchestration Specification for Cloud Applications (TOSCA). In November, VMware and partner SAP also said they would support TOSCA.
For companies still sitting on the cloud sidelines, this is good news in three ways. First, if they pick a vendor that supports TOSCA, and later want to move their cloud to another vendor, they will be able to do it at a lower cost than if they were moving between proprietary platforms.
Second, if they have a private cloud running on their own servers, they can move processes from it to a public cloud, and back again, to handle sudden spikes in demand, to use as backup systems, or to roll out new applications faster. This is the “hybrid cloud” approach, and many companies are looking at doing clouds this way so they have the control and security they need, when they need it, while still being able to benefit from public cloud infrastructure.
Third, companies will be able to buy pre-packaged cloud-friendly software systems that can be deployed with any of a number of cloud vendors. That would reduce the dependence on a cloud vendor’s own solutions, allowing more choices, and, in the end, lowering costs for everyone.
It remains to be seen how many other cloud vendors will support TOSCA, and to what degree they will implement it. If they don’t support it, enterprises will be stuck paying extra for cloud middlemen, vendors that package up cloud applications for customers in a way that enables them to be deployed to different clouds. Sure, those vendors add more value, as well. For example, they might monitor cloud performance and automatically redistribute applications as needed. But if the middleware vendor itself uses a proprietary system, that’s just stepping out of one quagmire right into another.
