Making sure a data center’s compute, network and storage resources are used efficiently is important. But making sure those resources are efficiently used in the cloud is critical. In the data center, finding a better way to balance the compute or storage resources may prevent the purchase of additional servers or storage, but these resources that it has have already been bought and paid for. In the cloud, the organization can effectively give them back, reducing their cost immediately.
What is Cloud Waste?
One of the attractions of the cloud is it enables an organization only to buy resources as it needs them and for as long as it needs them, now down to the second. The reality is that once an organization purchases an AWS EC2 instance, the organization will rarely go back to see if it is using it efficiently. Most organizations will only look at EC2 resource utilization if users are complaining about a lack of performance. Nobody ever complains about too much performance. The problem is the organization is paying for that excess performance even if it goes unused or isn’t needed.
Organizations needs to analyze their cloud resource consumption to make sure it is using only what it needs at the time it needs them. The analysis should also be done in real-time. For example, if an application needs eight cores during working hours but only one core at night, only a continually running process will identify that opportunity for savings. Ideally, this process will determine resource need on the fly, allocating resources the moment they are needed and then reducing resource consumption the moment the need passes.
Another aspect of cloud resource management is identifying what workload demands will be moving to the cloud. Ideally, the same process used to maintain efficient management of resources in the cloud will initially run on-premises, capturing information about the various on-premises workloads, so when the migration occurs, the organization knows exactly what the configuration of cloud resources should look like.
Part of cloud prediction is also knowing when the cost of those cloud resources is more expensive than hosting them on-premises. Again a process that can constantly manage and monitor enables that. It can not only look at current resource utilization but the combined long-term costs of continual cloud payments versus the one time payment of an on-premises alternative. It makes the rent vs. buy decision much easier.
Introducing Turbonomic for AWS
Turbonomic is a management automation platform that uses a supply and demand economic model to manage how resources are allocated in the data center. The company has extended this real-time supply and demand model into the AWS Cloud. The solution can analyze resource consumption, both compute and storage, to identify oversized workloads and make specific upsizing and downsizing decisions. Assuming that the product is also running on-premises, in the data center, it can accurately size workloads before the move to the cloud and leverage the combine analysis of both on-premises and cloud to determine the best location for each workload.
Turbonomic’s ability to continuously analyze to upsize and downsize workloads solving one of the biggest complaints IT has about its AWS experience – understanding the Amazon bill. Turbonomic can track current spending, make sure workloads have exactly the resources they need, and predict future monthly bills based on the company’s cloud plan. Most organizations should see a sizable decrease in their Amazon AWS spend almost immediately, and then also be able to slow the curve of the growth on that bill.
Resource management is important wherever the location of the data center. Resource management in the cloud has a more immediate payoff since unused resources can be “returned,” instantly lowering the monthly spend. IT also needs to understand and predict its monthly spend, something that most customers seem to struggle with.
Turbonomic provides both the ability to make sure workloads have just the right amount of resources and provides IT with the ability to project what its Amazon AWS costs will look like in the coming months. There are very few IT products that are a no-brainer, having a platform that can manage an organization’s Amazon spend and correlate to the on-premises cost is invaluable.