Converged infrastructure is moving “from infant to adolescent” writes Wikibon CEO David Vellante in his latest alert, which examines the overall trends in the converged hardware segment. Kicked off by Oarcle’s 2008 “Database Machine” announcement, the converged infrastructure marketplace now includes multiple entries from every major player from IBM to EMC, plus some startups such as Nutanix.
At the same time, the technology is moving forward. First-generation systems were more repackaging of exiting products than real innovation. IBM’s entry into the market in April 2012 was significant, Vellante says, in that it raised the level of convergence sophistication, essentially combining the strengths of the original, highly integrated, single-chassis systems, such as Oracle’s “red stack”, with much of the flexibility of the reference architecture systems, without requiring all the architectural, testing, and tuning work involved in implementing the latter. IBM’s PureSystems are totally integrated and tested single-chassis/single-SKU systems built to order with a menu of choices (e.g., different hypervisors, Power of x86 processors, a variety of network switches from several vendors, storage viurtualization that can act as the front end for additional external storage). This sounds analogous to ordering a new automobile and may signal the start of a new trend.
Simultaneously, hyper-convergence and flash storage are entering the market, Vellante writes. The former is championed by small, entrepreneurial vendors such as SimpliVity, Nutanix, and Scale Computing. They champion the Software-Defined Data Center with software-led, scale-out architectures that emphasize granularity and independent scaling of compute and storage.
SimpliVity and Nutanix also are pioneering flash storage as part of their architecture. Cisco has added options for Fusion-io and EMC VFCache flash storage modules to UCS. Sales volumes are small so far, but flash is beginning to have an impact.
Converged systems are basically IaaS in a box, and the cloud service providers (CSPs) have been the main customers so far. CSPs often do not have the ingrained ITO silo structure of most large corporate organizations, removing the political resistance from converged-system adoption.
The vendors tout their systems as silo-busters that promise a solution to one of the main budgeting problems in IT, the fact that more than 60% of infrastructure spend is on labor. While converged systems will allow CIOs to break down traditional silos and vendor lock-ins and decrease personnel costs somewhat, Vellante warns that they may find they have traded the traditional server, storage, and network organization for a set of highly converged silos. “As often the case when children become adults,” he warns, “the exact outcome can be unpredictable.”
As with all Wikibon research, David’s full report is available without charge on the Wikibon site. IT professionals are invited to register for free membership in Wikibon, which allows them to post comments and questions on research and publish their own Alerts and white papers for the Wikibon community. It also gets them invitations to Wikibon Peer Incite meetings, at which their peers discuss their experiences with leading edge products designed to solve pressing IT problems.
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