The old saw that says you can’t manage what you can’t measure applies just as much in the virtualized world as it does in traditional environments. But practice and tools are still in catch up mode.
Indeed, for a long time, IT pros could fly by the seat of their pants – looking at a range of indicators to get a sense of what was happening inside the server. But virtualization makes that a practical impossibility –there’s simply too much to going on. As a result, monitoring practices – especially server monitoring and the tools that support it – have been evolving.As workloads grow in complexity and become virtualized or cloud-based, IT monitoring tools have become a necessity, making it easier to shift workloads based on performance data and resource use. In a best case situation they can even anticipate changes in demand.
Cloud adoption continues to grow and accelerate as organizations discover more and more workloads that can do just fine off-premises. However, cloud computing is not without performance and management issues. And while there’s lots of experience around for running traditional data centers, much of that expertise and tooling doesn’t apply, or at least not so clearly, in the cloud.
That has raised the importance of products like DataCore’s SANsymphony which can manage things like storage and servers virtualized with VMware – even across different locations.
Although cloud and virtualization are closely related, some of their management challenges are different. And that has prompted the development of new capabilities in familiar management tools such as management stacks from BMC and others, as well as entirely new products.
To some extent, it is a matter of thinking of the cloud as a series of lifecycle management issues, as virtual loads come and go – orchestration, in a sense. And monitoring needs to be able to respond to that dynamism.
To date, many of the products that can help with cloud management are not as complete as traditional data center tools, though some of those traditional tools can provide capabilities across virtual and cloud assets.
But given the dramatic growth in cloud, these products are not likely to be the end of the story. For instance, a Computer Weekly article highlighted research from University of Modena and Reggio Emilia in Italy.There, researchers suggest that treating each virtual server as a single “black box” – the approach tools generally take – may be too cumbersome and slow. Instead, they advocate an algorithm based process by which similar virtual servers are lumped together by their behavior or workload and analyzed, thereby, more quickly and with potentially more useful insights.
In addition, a useful summary of some cloud monitoring tools can be found here. In general, it’s probably a good idea to look for vendors you trust that have some staying power, because they will likely continue to evolve with the nature and focus of IT workloads in the cloud or across environments.