Reducing Latency in the Cloud

Now that more and more enterprises are starting to get some cloud experience under their belts, attention is starting to shift from getting the technology to work to building an optimal cloud experience.

The primary fear of cloud providers and their platform suppliers is that failure to reasonably duplicate traditional data center performance will eventually hamper deployment. After all, a cheaper architecture is of limited value if it ends up diminishing productivity.

So it’s a safe bet that demand for the new generation of cloud optimization products hitting the channel will be high as public cloud providers and enterprises building private clouds seek to leverage the technology to its full potential.

Certeon made its move into this arena this week with a new version of its aCelera optimization system tailored to the cloud. The aCelera Cloud Optimization solution reduces many of the rough patches involved in connecting private clouds to public ones. With built-in multi-hypervisor support and central management, the system accelerates application performance across cloud platforms and maintains the ability to shift loads among various cloud providers.

And because cloud services generally rely on wide-area networks, it’s no surprise that most of the leading WAN optimization solutions are quickly migrating to cloud operation. Riverbed already had a wedge into the public cloud realm through its support for Citrix-based services, such as Amazon’s EC2. Now, it’s targeting the larger enterprise market by supporting VMware ESX environments. The announcement came on the heels of a partnership with automation specialist CloudSwitch aimed at reducing latency for legacy applications that extend into cloud environments.

Meanwhile, IBM and Akamai have pooled their respective DataPower and transport services to foster a more dynamic network environment for cloud services. The package consists of two versions, one for the enterprise and one for SaaS providers, utilizing the Akamai platform to chart optimal pathways for Web traffic, even bypassing the Internet if necessary, while IBM handles the final link from Akamai to the customer. The companies say they can cut application latency 30 to 50 percent.

Since the cloud is largely an extension of the virtual infrastructure that has arisen at most enterprises, it shouldn’t come as a surprise that it alleviates one concern — the need to provision maximum physical resources — but replaces it with another: ensuring optimal network performance. Virtualization, after all, greatly reduced physical server requirements even though it increased storage demands.

Ultimately, the trade-off for the cloud should be a net gain for the enterprise. Not only will you get a larger resource pool, but you should see a more vibrant, dynamic data environment all around.