VMware’s release of the first beta of its VSAN virtualization software for server-based flash storage and disks is creating a surge of interest in hyperconverged infrastructure, writes Consultant and Wikibon Analyst Scott Lowe in “VMware VSAN vs the Simplicity of Hyperconvergence”. Although not itself a hyperconverged system, VSAN allows users to create a software-led environment from server-based storage across a data center that, he says, can replace a physical SAN. It is also “a major validation” for companies in the hyperconverged space, such as Scale Computing, Nutanix and SimpliVity.
VSAN, Lowe writes, does provide the ease of use of hyperconvergence, allowing management of server-based flash and disk storage on multiple devices to be managed as a unit in vCenter. In the beta version scalability is limited to eight nodes, although Lowe expects the production version will handle 32 nodes, the maximum of a vSphere cluster. Inside the cluster, flash is used as a distributed read/write back mechanism, with 70% of capacity dedicated to the read cache and the remainder available as a write cache. Writes can be mirrored across solid state disks on different nodes. The flash storage tier is used to cache writes and convert them into IO patterns that are more sequential, improving efficiency when they are then written to the hard disk tier. It requires that both flash and hard disks be used in nodes, with each storage node containing up to eight disks.
By comparison, most hyperconverged vendors sell their software pre-installed on hardware, which greatly simplifies setup and maintenance, one of the selling points of these systems. Scale Computing’s HC3 and HC3X platforms, for instance, can be configured in an hour. Scale uses the open source KVM as its only hypervisor, further reducing costs, and adds tools that provide enterprise-class features to a system designed for SMBs.
Nutanix has the widest production selection in the space, with appliances containing one-to-four independent server nodes. Its wide range of solutions allows customers to match the product they buy more closely to their needs. It has released a Hadoop Reference Architecture demonstrating the appliance’s ability to support big data analysis.
SimpliVity differentiates itself from the purely software driven vendors by building its performance and capacity optimization into its OmniCube Accelerator Card. Although SimpliVity can be run without the card, an option it provides via Amazon Web Services, performance suffers. Its product, the OmniCube, is powered by the SimpliVity Data Virtualization Engine (DVE). It also loads a virtual machine known as the OmniCube Virtual Controller that contains all the intelligence in the system and talks directly to the Accelerator Card.
The full report, available without charge on the Wikibon Web site, provides a great deal of information on all four systems, of which this curation is only a brief synopsis. IT professionals are invited to register for free membership in the Wikibon community, which allows them to post comments, questions, and their own tips, alerts and white papers.
Latest posts by Bert Latamore (see all)
- Wikibon view: IBM legitimizing Spark to compete with Hadoop - June 19, 2015
- Red Hat flying high as big customer event opens next week - June 19, 2015
- Wikibon: Key ingredient of future storage architecture is leadership, not tiering - June 11, 2015