Latency to the Provider
CloudSleuth, and other external third party monitoring tools, do a great job of showing different scenarios of latency to the data center. The latency can help in indicating how an application might perform when hosted within this data center.
Just as important and usually even more elusive to determine is the jitter to and from the data center. Jitter is the measurement of one packet’s latency to another, and a large jump in latency can cause an application to underperform just as bad, if not worse, than a consistent set of high latency. It’s important when looking into a provider of your cloud services to evaluate not just the low latency but the consistent obtainment of that low latency.
You want to make sure over long time periods the provider has been able to hit performance levels above your defined expectations. Your customer base or user base might be geographically diverse. Finding a provider that can deliver the service to all your customers with geographic diversity is a must when your service expectations are so spread out.
Think about picking a single provider for this service. In doing so, you can truly realize cloud economics in not only the management of your diverse systems but in cost and scale capturing your overall buying power. This is part of the transformational value that the cloud delivers and in this value you truly enable your company to realize value in this transformation.
PeakColo’s embraced this model from day one and our diverse geographies allow our business partners to build truly unique and diverse services for their own value creation. Our partners truly embrace this transformation delivering their services not just whenever but WHEREVER the end customers want to consume this. All technologies benefit from lower latency - from virtual desktops, VOIP, ERP, website loads, mobile data, IPad Apps, and just about any application requiring user interaction.
Next week, I will dive into the need for latency reductions within the data center.