VMTurbo is the only vendor offering automated service assurance in the virtualization ecosystem today. Automated service assurance means that you identify the applications that are the most important to you (and the ones that are not), you assign them budgets of virtual resources, and VMTurbo ensures that the service level of the most important applications is not negatively impacted by the resource requirements of less important applications or workloads.
Now there are two very important tricks to getting this right. The first is to be able to measure importance in the eyes of the owners of the workloads or the applications in question. This is where legacy tools in the physical world completely fall down. This is in fact where DRS completely falls down. It is not about ensuring that workload A gets resource X at the expense of workload B. It is about ensuring that workload A’s service level in the eyes of the users of that workload, meets the expectations of those users. Users could care less as to whether or not a workload is getting 90% of the CPU (or less), 100% of the memory (or less) and 90% of the IOPS available in the system. User care about only two things. Those two things are whether the quantity of the transactions required are processed in time, and whether the response time for any interactive items is acceptable. So this all boils down to throughput and response time.
The VMTurbo 3.1 Announcement
With the 3.1 release of their product, VMTurbo has announced two significant enhancements to the product:
- Integration with Application Deliver Controllers (Load Balancers) – Citrix Netscaler being the first supported one. This is significant because in the eyes of some users of some applications the relevant metric is the number of transactions that are processed each minute or second. VMTurbo can find out from the ADC how much workload is going to each web server in a multi-tier application, and tune the infrastructure to meet workload goals. So if you want each web server to do 100 transactions per minute and they are not keeping up, VMTurbo can clone a web server image and spin up another instance, and tell the load balancer to put that new instance in the pool.
- An Open API for Integration with APM Solutions. So what if what the users care about is response time? What if response times need to be less than 500MS 99% of the time, and if this is not occurring then that is a service level violation. Well there is not generalized way to measure end-to-end response time for every combination of purchased and custom developed applications, so this requires integration with the products that measure response time for the applications that matter. Here VMTurbo has opened up their product to be able to receive response time (and transaction throughput) data from any APM solution that the customer has deployed.
The Future of Automated Service Assurance
What this means is that for organizations running business critical custom developed applications supported by vendors like dynaTrace, AppDynamics, and New Relic, the real performance data about those applications (response time and transactions per second) can now be used by VMTurbo to make either recommended or automatic adjustments to how resources in the virtual environment are relatively allocated between workloads. It means that for Applications Operations teams who need to support the performance (response time) of every application in production, as measured by products like AppFirst, BlueStripe, Correlsense, and Extrahop Networks that there finally exists a solution to assure the performance of everything that matters (at the expense of everything that does not matter).
This means that VMTurbo in conjunction with the APM vendors mentioned above are in the position to solve the most important problem facing IT organizations getting pushback regarding the virtualization of business critical and performance critical applications. This just might be the best thing that could have happened to VMware and its customers in the last two years.