Every business application is ultimately measured by its performance, especially over the local area network and...
Internet. As workloads move to private clouds, and even take aim at hybrid cloud deployments, the question of a workload's network performance is more important than ever. It's not a question of just adding bandwidth -- that's simply throwing money at a problem and hoping things improve. Instead, business planners and IT professionals need to take a proactive stance on cloud network performance through improved monitoring, careful utilization, accommodating architecture, and a variety of redundancy tactics.
It's important to remember that cloud network performance is not a one-time effort. It takes constant monitoring and frequent evaluation to determine whether performance meets business needs, or take proactive steps to prevent potential problems from becoming critical impairments. Regular testing is another important element of cloud network performance, ensuring that any network resiliency technologies and configurations work as expected -- if not, adjustments and corrective action can be implemented before problems arise that impact workloads and users.
1. Monitoring network and workload performance. The first step in ensuring private cloud performance is up to par is to always be watching. Luckily, there are network and application performance monitoring (APM) tools that can help you with this time-consuming job.
But what about when control of your workloads has been offloaded to a public cloud provider? Though your company may lose some control, it doesn't mean they lose the opportunity to monitor performance, and these same tools should be deployed.
2. Control workload sprawl and resource use. Cloud sprawl, or the overprovisioning of cloud services, eats up resources and can cause huge problems with performance -- especially in the long term. Once you find issues with network performance using APM tools, set up policies and procedures to mitigate the sprawl that's causing these dips in performance, and pull back resource and network use when necessary.
3. Design the cloud network for resiliency, prepare backup. Like most things, a solid foundation and planning can ensure future success. When designing your private cloud network, make sure it's ready to face inevitable failure or disasters. Redundancy is key to this design. Likewise, ensure cloud backup is in place to take over in the event of performance problems.
4. Qualify alternative public cloud providers. A cloud admin's job in researching potential providers is never done. Companies should be consistently evaluating new and existing cloud providers for compatibility with their workloads in applications, just in case something happens and a vendor switch needs to be made.
About the author:
Stephen J. Bigelow is the senior technology editor of the Data Center and Virtualization Media Group. He can be reached at firstname.lastname@example.org.
Amazon virtual private cloud takes back network control
Cloud concerns rise following an Azure outage
AWS fighting off cloud performance question