There is the data center, there is the cloud, and in between there is a minefield of systems, processes and policies to make it all work together.
Like it or not, that’s the situation the enterprise finds itself in these days. The promise of a new dawn in IT infrastructure is tantalizingly close, but it’s not a done deal yet.
If it was a simple matter of data migration, the hybrid cloud would be well on the way to becoming the dominant data architecture by now. Yet, many organizations are afraid to pull the plug until they are assured that this new paradigm is secure, reliable and resilient enough to entrust with critical data and applications.
According to Gregory Ness, vice president of marketing at CloudVelocity, even the costs of cloud computing are not fully known because things like migration, integration, governance and the like are rarely mentioned up front. Most providers like to tout the flexibility and operational benefits of their platforms after everything is up and running. The cloud is at its best when it can be integrated into legacy architectures, but that requires a much more robust management regime than many organizations are willing to implement – one that can maintain highly dynamic service and application mapping as well as broad resource relationship maintenance, and then take care of the authentication and LDAP responsibilities that arise during the deployment phase.
This is one of the reasons why functions like migration should become integral components of the upgrade process, says tech blogger Rick Blaisdell. Currently, most organizations treat them as items on a “to-do” list – important, yes, but not part and parcel of the entire cloud experience. But just as early operating systems removed memory and management responsibilities from programmers, so too will migration become part of the cloud fabric. Until that time, however, IT will continue to struggle with cloud migration and will need to keep abreast of all the tools and strategies needed to make it happen.
Already, cloud management platforms and migration systems are moving past the initial development cycle and are starting to shed some of the earlier approaches. The practice of snapshotting application images onto the cloud using virtual machines or container-based constructs is already coming under scrutiny, says CliQr’s Tenry Fu. An image on a cloud is only moderately useful because it lacks the ability to seamlessly connect to resources and other applications, particularly if you hope to maintain robust app portability. While it’s true that imaging simplifies pesky deployment tasks like discovery, it’s inadequate when it comes to the kind of end-to-end orchestration needed to convert the cloud from a basic storage platform to a robust IT application environment.
Clearly, none of this will be possible without a healthy dose of automation, and a fair number of small development firms see the cloud as a once-in-a-lifetime opportunity to gain leverage in the enterprise. Cloud Technology Partners, for instance, has issued a beta release of the PaaSLane platform, which gazes right into application source code to ensure compliance with established rules and processes. The idea is identify coding errors and other issues that would inhibit the app’s ability to function across disparate cloud architectures, thus ensuring that applications can not only be ported to the cloud but integrated into working, even customized, PaaS environments. The system utilizes Java and .NET C# deployment rules, as well as specialized Cloud Foundry rules developed in collaboration with Pivotal.
Establishing a presence in the cloud is the easy part. Integrating that presence with other cloud instances and legacy IT infrastructure is a challenge. In fact, it will be the single-most challenging aspect of IT in the new millennium.
There will be tools and platforms in abundance to help make it happen, but the significant legwork will have to come from the IT department, which must establish clear notions of what they want out of the cloud and how it should all work together. Without that, the cloud will be too unwieldy, and too ephemeral, to provide any real productivity gains for the knowledge workforce.