Hyperconvergence: Convergence in Warp Drive

Hyperconvergence is a relatively new technology paradigm that has been garnering attention. To understand hyperconvergence, consider traditional and converged IT infrastructures. The traditional infrastructure is comprised of server, storage, and networking systems, all of which are siloed and, consequently, often require separate administrative groups to manage. Converged infrastructures advance this approach by coupling compute, storage, and networking with a virtualization software layer. Converged solutions offer virtues like simplified datacenter design and can be offered by a single vendor. Although they may consist of bundled components from multiple vendors, there’s one phone number to call when an issue arises.

Hyperconvergence takes the integration paradigm JetStor Hyperconvergenceeven further. This strategy tightly consolidates compute, storage, networking, and virtualization resources in a commodity hardware box supported by a single vendor. Going well beyond virtualized servers and storage, hyperconvergence also integrates other key functionalities such as backup, replication, data deduplication, WAN optimization, and public cloud gateways, rendering standalone solutions unnecessary. In essence, the IT infrastructure consists of homogeneous, self-sufficient appliances.

This everything-in-a-box approach offers advantages. Again, there’s a single vendor and because all the boxes, or nodes, are aggregated, they’re managed as a single federated system. Everything shares a common interface and is administered via common tools, eliminating the need for multiple teams.

Expansion is simplified. Instead of scaling up by adding more compute or storage resources, you scale out by adding more boxes to the environment. Scalability is granular and can be done appliance by appliance, stabilizing IT budgets. The downside, however, is you can’t add only more computing performance when needed or only more storage. With each node containing all resources, buying more computing power means also buying more storage. But because the appliances are x86 server chassis, their costs are reasonable.

Hyperconvergence is not a panacea for all IT environments. Because the nodes don’t share between them, applications generating large volumes of data that require storage might be better served by storage arrays, for example. Hyperconvergence is a good solution for smaller sites or distributed locations where IT resources are sparse and the costs of enterprise storage arrays can’t always be justified. Here, hyperconvergence can deliver IT services efficiently and cost-effectively.