Dan's Take

More Efficient I/Os Through Read/Write Optimization

Condusiv touts its V-Locity, using lab tests to support the product.

Brian Morin, Condusiv's VP of marketing and product management, reached out to chat about his company's product, V-Locity, and a lab test presented by the industry research firm Enterprise Strategy Group (ESG.) While interesting, I found reason to question how useful this information would be to many potential customers.

Morin first took a few moments to speak about V-Locity, what the company describes as "I/O Reduction Software." The product appears to aggregate I/O from Windows and virtual machine  (VM) software from VMware and Microsoft and then optimize storage reads and writes. The product does this by gathering up small I/O requests and making the system read and write larger amounts of data in a single I/O. It also more intelligently places data, rather than using the first available space.

The ESG Lab Test
Morin proudly pointed to research published by ESG in which it audited operational data collected at 100 sites, that included data from 3,450 virtual servers. The report indicated significant performance improvements for workloads when V-Locity was engaged.

The key findings taken from ESG's lab test:

  • Reduced Read I/O to Storage. ESG Lab calculated that 55 percent of systems saw a reduction of 50 percent in the number of read I/Os serviced by the underlying storage. Even more, 27 percent of systems saw a 90 percent or more reduction in read I/Os.
  • Reduced Write I/O to Storage. By optimizing writes to be written in a more contiguous fashion, the size of an I/O consistently increases. In other words, instead of writing four 4Kb blocks of a 16Kb file, V-locity enables the system to write a single 16Kb write, requiring a single I/O operation. As a result of the I/O density increase, ESG Lab witnessed a 33 percent reduction in write I/Os across 27 percent of the systems. Fourteen percent of systems experienced a 50 percent or greater reduction in write I/O from VM to storage.
  • Decreased I/O Response Time. With an average available DRAM size of 3GB across all 3,450 systems, ESG Lab calculated the total time required to process all requests for each system and concluded that on average, systems achieved a 40 percent reduction in response time.
  • Increased Throughput. ESG Lab witnessed throughput performance improvements of 50 percent or more for 43 percent of systems. Further, 29 percent of systems experienced a 100 percent increase in throughput and as much as a 300 percent increased level of throughput for 8 percent of audited systems.
  • Increased IOPS from DRAM. Though the overall goal of the solution is to lower IOPS and improve throughput to the underlying storage, in some cases, because the working set was consolidated and serviced primarily out of DRAM, the number of measured IOPS dramatically increased. This means that the application was able to service requests faster. In fact, 25 percent of systems saw IOPS increase by 50 percent, and a small group of 25 systems achieved a 1,000 percent IOPS improvement. The same can be said for throughput. ESG Lab witnessed throughput performance improvements of 50 percent for 43 percent of systems, and as much as 300 percent increased levels of throughput for 8 percent of systems.
Dan's Take: Beware Non-Industry-Standard Benchmarks
As is obvious from my previous writings on surveys, benchmarks and other tools vendors use to get their messages out, I have strong opinions about my conversation with Condusiv.

First, ESG's comments, after examining the output of Condusiv's V-Locity management console, are clearly interesting. Will they be useful to all enterprises that have fielded virtual environments? No. They're only relevant to those that have fielded virtual environments supporting virtual Windows Servers executing under either a VMware or Microsoft hypervisor. Those using other operating systems or hypervisors will not find the results all that useful.

Second, the data is not the result of executing a recognized industry benchmark; it's the result of examining the output of Condusiv's management software that was observing a Condusiv-based environment.

Condusiv is competing with technology from a number of other suppliers. Citrix/Sanbolic, DataCore and Infinio, to name a few, offer what appears to be similar I/O aggregating and intelligent reading and writing technology. I'm reading reports that indicate that DataCore has gone a bit beyond this to actually making it possible for more than one I/O stream to be executing, as well as optimizing a single I/O stream.

All that being said, if your organization is using primarily Windows-based workloads executing in either a VMware or Microsoft virtual environment, it would be useful to download the software and try it out. If your computing environment isn't based upon that technology, it might be best to look to tools coming from others.

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