A hyper converged infrastructure (HCI) is an architecture type that consolidates and orchestrates virtual resources, networks, various storage types, and computing from individual machines. Integration of these disparate architectural components allow infrastructures to be governed as one system through a set of tools, toolkit, or user interface (UI). Nodes can be added to an HCI system base to extend workload capability, especially for virtualization and workloads within clouds.
With a converged infrastructure (CI), the premise of bringing all the components of the architecture within an infrastructure is the same as HCI, but with HCI, the integration of everything from storage to computing, and server virtualization to the data center are more securely and steadfastly unified. Through the use of software engineered specifically to maintain workloads of all kinds with a specialty in virtual workloads, such as those done across cloud environments, HCI offers a more protective—and in some cases, faster—architecture type for ventures using private or enterprise cloud environments.
While they may be the hottest new trend in the world of IT and development operations (DevOps) right now, HCIs are still relatively new to the end-user, and they do have their shortcomings. CI may be a better option for your enterprise because HCI can cause issues of latency and overall performance that you just won't find an issue when using CI. HCI does deploy quite rapidly, but even with the fastest deployments, if you're up against latency challenges, it will take a lot longer to execute and compute. For this reason, testing the two different infrastructure solutions within your architecture is a good idea—but if you don't have the time or resources to do that, consider that what you're choosing when you opt for HCI is usually greater speed... with greater chances for latency events occurring far more regularly.
While it certainly has its downfalls, HCI solutions and the software that they use to manage the architecture of your enterprise do offer up some advantages. For small businesses who only need a very small enterprise cloud environment and integration between very few aspects of their infrastructure, HCI can be used to do something called building a data center in a box. It can be a great solution for ventures like these, but it isn't just plug and play—a data center in a box run through HCI requires a great deal of upfront and backend testing before it does what admins and other users need for it to do. In cases where a data center in a box might work well, most companies would rather turn to other easier solutions, including legacy storage and traditional local area networks (LANs) until they are large enough to effectively use an enterprise cloud platform integrated with an all-flash array (AFA) data center.
When it comes to virtualization, Tintri has always been the cloud-friendly company, with storage systems engineered specifically to make it easier to handle your data on virtual servers and other machines. While Tintri does offer options in both CI and HCI, the larger deployments it engages in typically demand the performance (minus the latency) that CI offers. The most basic tenet of the Tintri promise to end-users is the simplicity of storage and data management through Tintri Enterprise Cloud, and a simplified user interface for managing data, software, and other information types.
If it’s true Web-scale you want for your enterprise, HCI is not the answer you’re looking for. Take a look at the systems used by the largest online companies in the world, including Amazon, Google and Facebook—none of them use HCI systems, and there’s a reason for that. While HCI does have its place within the market, enterprise cloud and Web-scale services are not it—not yet, anyway.
Unique control with VM-level actions for infrastructure functions including snapshots, replication and QoS make protection and performance certain in production, and accelerate test and development cycles.