Dec. 15, 2010 12:45 PM
JNBridge, a provider of interoperability tools that connect the Java and .NET Frameworks, on Wednesday said interoperability will make or break cloud computing success in 2011 and beyond.
"We believe that what is typically discussed under the rubric of cloud interoperability is so insufficient it will destroy the promise of the cloud," said Wayne Citrin, CTO of JNBridge and interoperability expert. "Our vision of cloud interoperability is one where applications and services can be developed in any technology and deployed to any cloud, where cloud services can be accessed by any other cloud service or from any client on the ground. In other words, any platform, anywhere, and until that exists, there will be severe limits on user adoption of cloud computing."
Citrin predicts the top three trends that enterprises should expect in 2011:
1. The realization among cloud users (if not also among cloud providers) that true cloud interoperability really doesn't yet exist – despite cloud vendors' claims to the contrary. There's an "emperor's new clothes" situation with regard to cloud interoperability. Despite what many vendors say, it just doesn't work. Citrin says this will become important in the coming year as users continue to adopt cloud platforms. Initially they will be migrating legacy applications to the cloud, since those applications will be immediately available. But the roadblocks that will come up, particularly with "platform as a service" offerings, may cause users to sour on cloud offerings and cause adoption to be delayed.
2. The introduction of tools and technologies to support interoperability – particularly, the ability to migrate legacy non-cloud applications into the cloud (regardless of the platform for which they were written, and regardless of the cloud platform to which they are being migrated), and the ability to access cloud services from clients running on any platform, regardless of the cloud platform upon which the service resides. However, he cautions that it needs to be easy. "Customers want things to 'just work,'" Citrin said. "For interoperability to 'just work,' it either will need to be built into the cloud platform offerings themselves, or will need to be third-party cloud services that can just be plugged into directly. The more technically sophisticated customers who create their own cloud services from scratch will have no problem adopting conventional tool-based offerings, but other customers who want to wire their offerings together from best-of-breed services (that are available regardless of platform) will just want these offerings to work together – either right out of the box or through the addition of easy-to-use interoperability services. These customers are used to the idea of connecting services together, and they will be comfortable plugging in a service that allows interoperability."
3. Initial steps supporting the notion that interoperability can be a service, too – as well as a conventional tool. Cloud services supporting interoperability will begin to be introduced, and will start gathering early adopters. However, Citrin says customers may be disappointed that the services they create will not be usable by up to half their potential market. If they write Java-based services, they will shut out .NET-based customers, and if they write .NET-based (Azure) services, they will shut out Java-based customers. Creators of cloud services may be disappointed by the adoption of their offerings, which again will slow the further roll-out of cloud services.
"Interoperability in the cloud is currently a real problem. Supporting real interoperability is critical to the success and survival of the cloud computing model, and right now it is a large stumbling block that prevents adoption," said Mark Driver, vice president of research at Gartner. "Initiatives that support cloud interoperability in its various aspects will benefit cloud users, vendors of cloud services, and vendors of the underlying cloud platforms."
Aside from interoperability, Citrin sees one other cloud trend in the coming year. "We will see greater support for the 'housekeeping' capabilities that will need to be part of any offering built on cloud platforms. Particularly, features like user access and feature control (also known as 'licensing') and metering/billing support. I suspect these will arrive in the form of service offerings themselves that can be plugged into the newly developed cloud offerings," he said. "However, without robust support for licensing and billing, software developed for cloud platforms won't be anything more than just a toy."