Most corporate data centers are complex collections of hardware and software that have gradually expanded over a number of years. Need more storage? Add a storage appliance. Need to beef up security? Add a next-generation firewall. At the same time, organizations are under pressure to reduce management costs and speed the rollout of new applications and services. Unfortunately, an IT infrastructure with dozens of add-ons is inherently complex, which isn’t exactly conducive to lower management costs or agility.
The need to reduce this complexity has led to the rapid adoption of hyper-converged infrastructure solutions. Hyper-converged infrastructure is essentially a data center in a box. Compute, storage, networking and virtualization are tightly integrated in a pretested, preconfigured solution that is supported by a single vendor. In addition, hyper-converged infrastructure solutions typically include natively installed management software, data reduction services, WAN optimization and other features.
Hyper-convergence emerged as an alternative to the best-of-breed approach of building a data center from the ground up. It accelerates deployments and simplifies management by minimizing complexity and the risk of compatibility issues. A modular design makes it easy to add resources as needed without increasing complexity.
While the simplicity of hyper-convergence is its most appealing characteristic, first-generation systems have serious limitations. First, they typically run on traditional servers with latency-inducing software wrapped around legacy, open-source files systems. Second, they use network overlays and lack cluster network integration. This design not only creates performance issues but makes these systems difficult to deploy, troubleshoot and maintain. As a result, hyper-convergence has often been limited to light workloads, with many organizations hesitant to risk deploying latency-sensitive and mission-critical applications on first-generation hyper-convergence solutions.
Last year, Cisco introduced Cisco HyperFlex Systems, which apply a new architectural model to hyper-converged infrastructure. Based upon Cisco’s Unified Computing System (UCS) technology, HyperFlex integrates the Cisco HX Data Platform, a distributed file system built specifically for hyper-convergence, with low-latency UCS fabric and deep network policy engineering. As a result, HyperFlex delivers 40 percent greater performance than competitive products and applies policy-based automation across the entire system.
HyperFlex also includes powerful data management services that reduce the data footprint by up to 80 percent, and enables flexible, adaptive and independent scaling of compute, network and storage capacity. Together these features help to overcome the limitations with performance, flexibility and management that has plagued first-generation solutions.
Last month, Cisco announced new capabilities for its HyperFlex Systems products. The latest versions of HyperFlex provide support for 40Gbps UCS fabric networking and feature high-capacity, all-flash nodes that enable greater IO consistency and dramatic performance gains. Other enhancements include the intuitive HyperFlex Connect management platform, advanced analytics and orchestration capabilities, enterprise-grade data protection and management, and hybrid cloud mobility.
While first-generation hyper-convergence had significant drawbacks, the new generation is faster, simpler and capable of supporting mission-critical applications. Cisco HyperFlex Systems are leading this new generation. Let us show you how HyperFlex overcomes the limitations of legacy solutions and delivers on the promise of hyper-convergence.
By: Michael Hritz, Sr. Cisco Business Development Manager, Sigma Solutions