I was recently working with an international service provider who was experiencing ongoing issues with storage IO performance in its VMware private and public cloud infrastructures. The impact of this was very significant for their support organization because they were fielding a large number of support calls on a daily basis. Many of these incidents resulted in three or four of their best engineers from storage, network and virtualization support teams spending two to three hours trying to figure out the exact cause of the problem. Even more troubling for them was that they had lost a number of customers to alternative service providers.
Now, the service provider is looking to put a more comprehensive monitoring solution in place to detect these IO performance issues more quickly, and offer subscribers reports that provide better visibility into the health of their service. As we dug deeper into their requirements, it became very clear that the real problem they needed to solve was not a monitoring or reporting problem. Their ultimate goal was to prevent storage performance incidents from happening in the first place. If they could achieve this, it would enable them to significantly reduce their support costs, free up their best resources to focus on activities that generate value for their business, and keep their customers happy (preventing them from choosing alternative vendors).
The service provider is also launching a new premium service that offers SLAs for storage performance. Having a way to prevent storage performance issues would allow them to deliver this service without significantly over-provisioning IOPs capacity.
The requirements outlined by this service provider align with a common issue: solving the intelligent workload management problem. That is, ensuring workloads get the compute and storage resources they need to operate reliably while maximizing the efficiency of the infrastructure.
At best, the combination of monitoring solutions and the best experts in the world can only hope to get to a point where you have greater visibility into the cause of storage performance issues, but not prevent them from occurring. Furthermore, adding high-performance disk is often an unnecessary and significant expense.
Storage performance issues in virtualization are one of the most complex issues to address. VMTurbo provides a unique way to prevent problems, eliminating the need for costly support and infrastructure investments. Learn more about how VMTurbo is tackling the problem of how to prevent storage performance issues.