Enhancing Excelero’s NVMesh storage with Intel Optane technology
When it comes to migrating any IT environment into the cloud, from simple archives to the most complex ERP environments, one imperative applies. Cost efficiency is critical, and teams must quickly define what it takes to deploy the most cost-efficient method to solve application or workload needs. Although this discussion centers on migrations to the cloud, the scenario rings true for both cloud and on-premises deployments.
Achieving cost-efficiency involves many areas of the infrastructure, most notably compute, network and storage. Different applications depend more heavily on one or the other of these elements for optimal performance. For example, taking archives to the cloud demands superior network bandwidth and storage throughput. In contrast, taking a database deployment to the cloud requires low latency storage access and maximum CPU horsepower.
Storage is often the silent bystander to data center efficiency challenges. Often, we see teams unknowingly increasing compute to overcome storage performance struggles, effectively lowering the overall cost efficiency. In the cloud, where network bandwidth can be very high, both storage throughput and storage access latency are the true bottlenecks, creating inefficiency in nearly all environments.
Excelero’s NVMesh storage software was designed to remove the storage bottleneck so teams can deploy their core, performance-hungry applications in the cloud and do it with efficiency that often exceeds typical on-premises performance. NVMesh for Azure is a software-based storage solution that runs on Azure virtual machines to create a high-speed, shared access storage environment that can meet the performance needs of core Oracle/SQL databases, AI/ML/DL workloads, Kubernetes deployments, and any other applications that has a thirst for speed. Because it runs on the virtual machines themselves, NVMesh pools the lightning fast NVMe drives on the machines and distributes them across the network to deliver 100+ GB/s of throughput at less than 100 microseconds latency. By distributing the storage across multiple machines in the cloud, NVMesh technology also distributes the risk of data loss due to drive or VM failure across all the nodes as well. The result: a fast storage layer, and just as importantly, a highly reliable, enterprise-class storage solution for your business-critical environments.
So, what does all this have to do with efficiency? A lot.
First, when infrastructure bottlenecks go away, every other component can run more freely. It’s like trying to drive a car with your foot on the brake. The engine struggles to achieve performance and guzzles gasoline. Take the foot off the brake, and the engine can run free. The car arrives at the destination in less time and the fuel mileage increases.
The same goes for increasing the performance of storage to let compute run freely. Application run times will be lower, delivering end user results faster. Once in the cloud, IT teams can then shut down the VMs until the next run. Efficient.
Second, NVMesh technology is an ultra-low overhead storage software solution that requires very little CPU overhead to operate. With the high numbers of CPU cores in today’s virtual machines, those machines running NVMesh software can also run applications – meaning the cost overhead for the storage environment is not completely additive. To reuse the car analogy, think of this as driving a large group in a bus to their destination. Taking two or three group coordinators with you becomes far more cost-efficient when they take a seat on the bus, versus driving a separate vehicle to the destination. By comparison, when running an Oracle database that needs 60 cores to operate, IT teams can choose a 64-core machine with NVMe drives and NVMesh technology will just use a couple for the remaining cores to provide high-performance storage. Again, efficient.
Last but not least, a third path to efficiency is with the storage itself. Intel Optane technology is a game-changer that enhances the efficiency of Excelero’s NVMesh storage software even more. Intel’s Optane drives bring high-speed memory technology to the front end of storage drives. The data hits the memory front-end so the read/write doesn’t need to always go to the drive media.
In the car analogy, this is like dropping a supercharger on the motor of the car. The car needs to be able to handle that burst of performance, but when it can, it will scream! Optane technology delivers the speed for IO and NVMesh storage software provides that fastest pipe across the drives.
On top of just performance, the Optane technology provides an additional capability to NVMesh storage: tools to create erasure coding across more cost efficient (QLC) NVMe drives in the cloud. Without diving too deep into details of NVMe technology, there are basically two types of NVMe drives: MLC enterprise drives and QLC cost efficient drives. In order to get a better price for capacity – especially when going to the cloud – the QLC drives are necessary. When architectures combine QLC drives with Intel Optane drives, NVMesh can deliver breakthrough speed to applications, while more efficiently using the drives with erasure coding to boost storage utilization.
Migrating applications to the cloud, especially performance hungry databases, AI/ML and Kubernetes environments, is an opportunity to reexamine existing infrastructure and achieve a tremendous boost in cost-efficiency as well as performance. See what Excelero’s NVMesh storage software can do to help hit your performance targets with both resiliency and efficiency. Together with Intel’s Optane technology, you can take cost-efficiency to the next level.
Learn more at www.excelero.com.