The cloud is intended to provide unlimited scalability. But what happens when your cloud provider's infrastructure is tapped out or suffers an outage?
Ideally, your provider would simply offload resources to another provider, and then another, and another, ad infinitum, so that no matter what happens, the enterprise has the means to support data requirements without spending millions on new capital projects. To do that while retaining control and visibility into wildly disparate infrastructure is easier said than done, however, primarily because it requires a fair amount of coordination between the enterprise, the cloud provider and the provider’s provider.
The working term for this state of affairs is the "cloud of clouds." That is, one cloud provider integrates his platform with multiple other providers, enabling an overarching cloud topography that looks and acts like a single entity for enterprise data and infrastructure management facilities. It's a worthy goal, but as any system admin will tell you, it will require coordination across a number of fronts, including IT, application development, database administration, possibly even carrier networks. And with clouds providing backup for other clouds, expect the cost to trickle down to the enterprise eventually, even if such massive distribution of resources spreads the cost over large numbers of users.
One way that large organizations are accomplishing this is through brokering. Federal agencies are developing their own cloud formats, which are then made available to various service providers to enable widespread and redundant cloud environments. The Department of Energy recently launched its own brokerage service, dubbed YOURcloud, which provides a self-service portal for IaaS environments across multiple providers. Coupled with the ONEvoice collaboration platform and the OneNNSA network overlay, agency officials have access to a virtually unlimited set of resources while maintaining strict security and accessibility requirements.
Clearly, interoperability among cloud platforms will require a fairly sophisticated set of standards. Management developer OASIS is looking to establish its Cloud Application Management for Platforms (CAMP) system as a portable PaaS solution designed to provide a cohesive cloud environment across multiple providers. The company has already drawn the likes of Oracle, Red Hat, Rackspace and Huawei into the fold, bringing full application lifecycle management to multi-cloud deployments. The platform provides a common development vocabulary and APIs for both PaaS-aware and PaaS-unaware development systems, as well as an interoperable protocol that cloud providers would use to enable functions like self-service provisioning, monitoring and control.
As well, a number of cloud interoperability solutions aimed select vertical markets have hit the channel. Iveda Solutions recently teamed up with Power-All Networks to enable multi-cloud support for video surveillance services. Using PAN's Inter-Cloud Computing platform, Iveda users gain broader scalability through multiple points of presence spread out across 10 nations around the globe while retaining a single user interface for multiple storage and network video recorder (NVR) software formats.
The drive toward the cloud of clouds is essentially a continuation of data systems development going all the way back to the first silicon chip. One processor is fine, but two or more on a motherboard is even better. The PC was a step up, but then came networked PCs, followed by the Internet, and now the multi-cloud.
If nurtured properly, this kind of data federation could ultimately lead to a single, open computing environment in which all applications, platforms and data environments work seamlessly with little or no configuration requirements on the user end. It probably won't get to that point, but it’s nice to have a dream.