Hyper-converged solutions use software-defined technologies to integrate compute, network and storage resources into modular blocks that can be combined to form data center infrastructure. Each module provides predictable performance which makes planning upgrades and managing ongoing operations easier. For example: Let’s say you need a data center solution that has 10TB of space and 25 cores of computing power to handle 50 workloads that are running data analysis. With traditional architecture, you would select the appropriate storage solution for the workload, then the appropriate compute/servers to host the virtual workload, and finally the network switches to create the solution. In a Hyper-converged world you would look for the module that most matches a fraction of the requirement and then implement multiples of them to reach the needed resource levels. The modules combine together to form a single data center and aggregate their resources across the total. This aggregation also provides redundancy for the whole solution.
Hyper-Converged Solutions for Mid-Market Businesses
Hyper-converged solutions provide a much needed break for overworked IT personnel, for whom managing data center operations is a secondary duty. The need to manage and monitor external storage systems and networking separate from the virtualization stack is eliminated. Instead the solution is managed at the virtualization management console, which is where administrators spend most of their time anyway. When new projects and workloads need to be addressed, it becomes a simple matter of purchasing additional modules to increase the aggregate resource pool. Is performance the issue? Hyper-converged products include mechanisms to cache active data on each module to provide Flash or better storage performance. Keeping the data near the CPU removes delays associated with traditional architectures and leads to faster workloads. Hyper-converged solutions also work well for remote offices where resources may not be available to diagnose and resolve technical issues.
Hyper-Converged Solutions for Enterprise Businesses
Enterprise businesses deal with both Mode 1 (scale-up) and Mode 2 (scale-out) applications. For scale-up applications, Hyper-convergence provides a predictable performance platform to ensure that mission critical applications operate at peak capability. For scale-out applications, the simplicity of adding modules to expand available resources provides cloud burst like functionality but without the complexities that usually come along with it.
Mavenspire’s Hyper-Convergent Solutions
We work with two hyper-converged partners – Pivot3 and Nutanix.
Pivot3 was the first hyper-converged solution on the market. Initially focused on the video surveillance industry, this solution is highlighted by its use of a global cache, which provides high-performance storage for even the most demanding application, erasure coding, which in turn supplies much higher storage efficiency than competing products, as well as a fully integrated VMware focused management. Recently, Pivot3 acquired another software-defined storage company which, once integrated, will provide workload-based quality of service. This will be particularly valuable for service providers and enterprises with multi-tenant environments needing to eliminate noisy neighbor problems.
Nutanix is the most visible of the hyper-converged products. Originally focused on VDI and VMware relationships, Nutanix has blossomed to include hardware platform support from SuperMicro, Dell and Lenovo, as well as support for VMware, Microsoft and Linux Hypervisors.
Recently Nutanix has tightened their focus on delivering an experience that optimizes cost and provides functionality similar to OpenStack through their Prism management and Acropolis platforms.