The traditional data center infrastructure is comprised of discrete realms (compute, storage, network, etc.) each with discrete appliances (servers, SANs, switches). As a result of the inherent dissimilarities found in traditional infrastructure builds, you end up with equally dissimilar management interfaces even within a particular realm. For example, EMC and NetApp SANs, or Dell and HP servers, or Cisco and Juniper switches all require utilizing unique management interfaces.
Converged infrastructure first came onto the scene in the form of built-to-order plug-and-play rack-sized appliances that are comprised of compute, storage, and network appliances. These racks-to-go enabled IT departments to skip the whole rack, stack, and configure dance and just plop these beasts into their datacenter.
Hyperconverged integrated systems (HCIS) have entered the market as the next step in furthering convergence. Hyperconverged infrastructure is the integration of those discrete realms in a single, relatively small, appliance. These appliances act like nodes that can then stack, which creates a modular scale-out infrastructure (think Legos).
The three types of infrastructure we’ve discussed (traditional, converged, hyperconverged) have also been referred to as blocks, racks, and appliances, respectively.
Here are a few distinct benefits of hyperconverged infrastructure (other than sounding exciting and awesome):
- Administrative simplicity: With many HCIS appliances, you also benefit from a single pane of glass rather than the multitude of management interfaces, which can include the ability to provision compute resources.
- Physical simplicity: Adding additional nodes is a plug and play installation and minimizes your data center’s hardware disparity.
- Fine grained scaling: HCIS takes advantage of the scale-out approach (many tiny machines working together) rather than the scale-up approach (a few massive machines). This can heavily reduce CapEx spending when it is time to scale since you’re only adding ‘little’ nodes in place of ‘large’ traditional appliances. However, this is not to say that these appliances are cheap.
The linear network, compute, and storage scaling of HCIS make its most touted use case virtual desktop infrastructure (VDI). Other popular use cases are for branch micro-data centers and for general purpose servers. Where HCIS begins to make less sense is when you have higher requirements for either computing or storage, for example, heavy database crunching. There may also be security concerns that arise from the flatness of an HCIS environment and security compliance concerns that arise from the tight integration inherent in HCIS.
Hyperconverged stands on the shoulders of converged, but with the rate of technology development and the weight of IT driving business decisions increasing there is already discussion about what will stand on the shoulders of hyperconverged. How can your data center benefit from convergence, or a blend of infrastructure solutions? Reach out to some of our infrastructure experts Brian Alletto and Jerin May.
Join the conversation! Whether about the current maturity level, adoption rate, or future of convergence, we would love to hear your thoughts.