Dan's Take

More Efficient Virtual Machines Through Intelligent Caching

Infinio promises greater vSphere performance.

Arun Agarwal, co-founder and CEO, and Scott Davis, longtime friend and CTO, both of Infinio dropped by the other day to both introduce the company and discuss the problem they're solving. Although I know many of the company's executives from their other appointments, I hadn't heard of Infinio.

Infinio brought together a number of really interesting and innovative people to address the growing challenge of getting data in and out of virtual machines (VMs) to accelerate virtual workloads. The company understood that many virtual workloads executing on the same physical host present a different storage profile than the same workloads executing directly on a physical host.

Having followed Davis through different companies, I know first-hand that he really knows his stuff; he understands how data flows through virtualized environments; and, last, has great ideas about accelerating those data flows for improved performance.

Infinio believes that intelligent caching algorithms combined with a tight integration with the VMware hypervisor makes it possible for needed application components and data elements to be in the cache when workloads demand them. If done correctly, this improvement will be seen regardless of whether the actual data resides on rotating media, solid-state media or some combination of the two.

It's the company's belief that this type of intelligence will offer significant performance improvements over similar computing environments without that caching.

Dan's Take: Marrying Virtual Processing and Virtual Storage Is Critical
If we step back for a moment to examine the architecture of most industry-standard, x86-based systems, it becomes clear that the storage subsystem is usually designed to address the needs of a small number of workloads and the load they impose on the storage devices. Only so many requests for data reads or writes are in the queue at any moment. Furthermore, where the data items actually live on the storage medium can be predicted ahead of time so that the right data will be found in the system's cache a high percentage of the time.

If we add virtual processing software to the mix, the number of requests for data can skyrocket, and where the data actually is living is often much less predictable. If we also add the different performance characteristics of today's different classes and types of rotating media; the emerging use of solid-state, flash storage; and the storage interconnects being used, placing the data in the proper location can mean the difference between a system configuration that performs well or poor under the imposed workload.

Infinio's take on storage virtualization for VMware-based virtualized environments appears both innovative and quite useful in helping enterprises make the most of their computing environments. Its caching mechanism appears to be intelligent enough to make the best use of the available storage devices and storage interconnects to accelerate virtual applications. The company's tight integration with the VMware hypervisor means that management of the environment will be easy and productive for administrators. Although the company is new on the scene, it has decades of experience. I expect to see good things coming from it.

About the Author

Daniel Kusnetzky, a reformed software engineer and product manager, founded Kusnetzky Group LLC in 2006. He's literally written the book on virtualization and often comments on cloud computing, mobility and systems software. He has been a business unit manager at a hardware company and head of corporate marketing and strategy at a software company.


Subscribe on YouTube