How hyperconvergence has evolved and making the most of the differences.
The IT infrastructure market is undergoing unprecedented transformation. The most significant transformation is reflected by two major trends: convergence and software- defined data centers (SDDCs). Both trends are responses to the IT realities of infrastructure clutter, complexity, and high cost; they represent attempts to simplify IT and reduce the overall cost of infrastructure ownership. Today’s infrastructure environments are typically comprised of 8 to 12 hardware and software products from as many vendors, with each product offering a different management interface and requiring different training. Each product in this type of legacy stack, however, is grossly overprovisioned, using its own resources (CPU, DRAM, storage, and so on) to address the resident applications’ intermittent peak workloads. The value of a single shared resource pool, offered by server virtualization, is still limited to the server layer. All other products are islands of overprovisioned resources that aren’t shared. Therefore, low utilization of the overall stack results in the ripple effects of high acquisition, space, and power costs. Simply put, too many resources are wasted in today’s legacy environments. This article explores a leading solution: convergence, which ultimately leads to hyperconvergence.
The Evolution of Convergence
The following sections describe the evolution of convergence over the past few years.
The earliest infrastructure convergence solutions have complete network, compute, storage, and virtualization capabilities, but they’re simply conglomerations of existing hardware and software, with little to no actual innovation in product features to be leveraged. Many people look at the solutions offered by major original equipment manufacturers as being ways to lock customers into their technology stacks. Integrated systems do offer a few benefits. Most notably, customers gain a single point of contact for their infrastructure, from purchase to end of life. These systems are always tested and almost always arrive at the customer site fully racked and cabled, so they’re ready to go. On the downside, these systems often have a big step size. When you need more power, you have to buy a big chunk of infrastructure. Also, these products usually don’t solve the serious challenges that so many organizations face.
The converged infrastructure products combine the server and storage components in a single appliance, effectively eliminating the need for dedicated storage area network (SAN)-based storage. These systems provide a localized single resource pool solution, offering simplified management and faster time to deployment. They have effectively virtualized the storage layer and allow it to run in the virtualization platform. Overall acquisition cost is lower, and management (at least, for the server and storage resources) is simplified. With these systems, overall resource utilization is higher than with a legacy island-based infrastructure. Converged infrastructure has some limitations, however:
The systems include just the server and storage resource components.
The fundamental data management challenges have not been solved. It is the functionality of a traditional storage array migrated into the virtualization platform.
Resource ratios (such as CPU:storage:network) are fixed, making them less flexible than some organizations require.
The products can’t always be used by existing infrastructure. In other words, you can’t use a converged infrastructure appliance’s storage from existing legacy systems. In essence, you’re forced to create a separate resource island.
For these reasons, converged infrastructure systems don’t efficiently address performance problems in the legacy infrastructure. Likewise, on the data front, the systems don’t address all data problems, because not all data efficiency appliances are converged. Management may be improved, but it’s not unified global management.
Enter hyperconvergence, also known as hyperconverged infrastructure, which represents the logical next step in the evolution of infrastructure convergence. Hyperconvergence delivers simplification and savings by consolidating all required functionality into a single infrastructure stack running on an efficient, elastic pool of x86 resources. The underlying data architecture has been completely reinvented, allowing data management to be radically simplified. As a result, hyperconverged infrastructure delivers on the promise of the SDDC at the technological level. It also carries forward the benefits of convergence, including a single shared resource pool. Hyperconvergence goes far beyond servers and storage, bringing into the convergence fold many services that make some legacy services obsolete, including the following:
Data protection products (backup, replication)
Wide-area network (WAN) optimization appliances
Solid-state drive (SSD) arrays
SSD cache arrays
Public cloud gateways
Replication appliances or software
Previously, we’ve delved into how hyperconvergence takes convergence to the next level in the data center and provides a plethora of benefits to both IT and to the business.
The preceding convergence options build on one another, with each having key differences. Figure 6-1 describes the high-level characteristics that define each convergence type.
Figure 6-1: Compares improvements as convergence has evolved. Each characteristic is critical to realizing all the traits that business demands of IT in the modern era: lean, mean, and green.