If you're in the cloud space, you've heard the term "noisy neighbor" before. If not, here's a tip: It's not about the house next door with the weird kids who play German death metal at full blast. It's about your cloud provider's support of tenancy.
Cloud workloads that put large demands on the server, storage, database, or network -- hurting the performance of other workloads that share those resources -- have been a problem both inside corporate data centers and, in particular, on SaaS and IaaS cloud platforms. When you own the servers you at least have some control over the situation. If others control the servers -- the case in the cloud -- all you can do is to complain to them.
[ Amazon can no longer take its cloud leadership for granted | Get the no-nonsense explanations and advice you need to take real advantage of cloud computing in InfoWorld editors' Cloud Computing Deep Dive PDF special report. | Stay up on the cloud with InfoWorld's Cloud Computing Report newsletter. ]
This "noisy neighbor" issue is a core reason why many enterprises have not moved their important applications to the cloud. The fear is that someone running an intensive application on another machine instance in the same infrastructure you're sharing may affect your performance. In my experience, however, this fear isn't borne out in practice.
This is not to say that "noisy neighbor" problems don't exist; they certainly do. Nonetheless, the majority of performance issues I've seen relate more to bad cloud providers than someone hogging resources on the platform.
Cloud providers are charged with creating and maintaining a multitenant platform that can isolate workloads from each other. If you're impacted by another workload, then the cloud provider is not doing its job. It's not a cloud problem but a vendor problem.
Most good cloud providers -- IaaS, SaaS, or PaaS -- support architectures that minimize the "noisy neighbor" problem. However, enterprises should be skeptical of anyone who claims it provides consistent performance. Do some long-term benchmarking to be sure.
I suggest you run workloads that simulates your use of the system, then monitor performance over a few weeks. You will see that cloud platforms don't provide entirely consistent performance (neither do your data centers). Regardless, the performance you see should work within your accepted tolerances for your workloads.
If not, you have a bad provider. Don't blame your cloud neighbors.
This article, "Suffering from cloud contention? It's not the cloud's fault," originally appeared at InfoWorld.com. Read more of David Linthicum's Cloud Computing blog and track the latest developments in cloud computing at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.