When servers sit in the private cloud, it forces us to reconsider how the customer experience would be affected. For private cloud solutions not all applications are web browser based so the considerations need to be a lot more stringent. Will the service provider have a big enough pipe to service all those requests ? How much contention is there on those links ? How does it vary during the course of the working day ? If you don’t have a dedicated setup for your virtual data centre, then what could be the impact ? The table below shows a point-in-time assessment of the bandwidth and latency that we can expect from the three providers that E-Bit ran some of these tests against.

GoGrid, GoDaddy and a number of US-based providers have latency of between 160ms to 250ms, depending on their location. If your consumers are in Australia and you really care about the end user experience, then hosting in the US may not make sense unless you are servicing the global market or you are on of the few (but growing) groups with applications that are not sensitive to latency eg web, email, telnet type of access. The table below gives you a sense of the performance expectations from a few of the providers.

As you can see, Web24 seems to have the largest pipes available amongst the various virtual server providers at the time. Since then a number of new providers have cropped up all over, the reference table could be used against any you may be considering. Web24’s pipe size and latency made them my first choice of service provider for some applications. I’ve also found them to be flexible in helping set up my virtual Datacenter, rather than using a one size fits all approach, making Web24 a good choice when you need more than just a virtual server in the cloud. However, such providers are few in number and would not suit all businesses.