Tuesday, October 18, 2011

Cloudera and SGI: Hadoop for Geeks AND Suits

Some of you may be too young to remember it, but there was a time when business technology decision makers looked askance at open source solutions such as Linux and the Apache Web server. These and others are now common components of many business technology infrastructures.

What happened? Some pioneering companies such as Red Hat and Ubuntu began wrapping packages of supporting technologies and services around Linux and related offerings. And those packages began to overcome the fears and objections of the "suits" who approved the budgets of the "geeks" already sold on the technological advantages of Linux, et al.

Fast-forward to today. Apache Hadoop is a combination of a distributed file system and nifty behind-the-scenes software. Hadoop makes it easier and faster to build and run applications that require lots of data and the shared power of multiple computers. And so far, Hadoop has followed a trajectory similar to that of Linux in its early days as a viable business computing platform. The geeks already love it or are champing at the bit to take it out for a spin, but the suits are largely unfamiliar and wary.

Enter the alliance of Cloudera and SGI, the venerable Silicon Graphics International. Cloudera is doing for Hadoop what Red Hat, Ubuntu and others did for Linux and related technologies. Its Cloudera Enterprise offering combines a suite of management software with support services to make Hadoop adoption and realization of its benefits faster and easier. SGI, meanwhile, combines the strengths of its server hardware with its strengths in markets that have proven strong for high-performance computing (HPC) and are likely to be equally enthusiastic users of Hadoop. These markets have included government, research and telecommunications segments, according to SGI.

Providers of cloud-based business applications and services will become one of these key markets very soon. Hadoop-powered computing clusters as a service, and the kinds of applications those clusters can support, should prove irresistible to users with advanced requirements, and to Cloudera, SGI and their partners.

Meanwhile, the Cloudera-SGI alliance is presenting itself as equally appealing to technical and business decision makers. For the geeks, SGI announced yesterday that it had set new Hadoop performance records by running Cloudera software on SGI servers. For the suits, the companies announced yesterday the availability of SGI Hadoop Clusters with the Cloudera Enterprise Management Suite already installed. "The relationship will also enable the two companies to jointly build, sell and deploy integrated, high performance Apache Hadoop-based commercial solutions," the announcement said. I can't imagine that the two companies will take very long to turn their collective expertise and experience into packaged solutions and services designed for the suits and the geeks in specific markets and/or facing specific challenges.

If you work at an enterprise that you think can benefit from business-enabled, purpose-built Hadoop deployments, keep an eye on Cloudera and SGI, especially as they ramp up competition with Oracle's software and its Sun hardware. And whether your company needs or wants Hadoop or not, look at the Cloudera-SGI alliance as a harbinger of more such relationships to come.

Vendors understand that users prefer solutions that let those users focus on running their businesses, not any particular technologies. Responsive vendors will increasingly deliver combinations of purpose-built, pre-configured physical, cloud-based and cloud-enhanced "appliances" that are easy to deploy, require minimal user management and deliver benefits rapidly. And for the solutions that meet those criteria, the inner workings matter far less to buyers and users than how well those workings work. Whether those buyers and users are geeks, suits or geeks wearing suits, a rapidly growing demographic segment...

1 comment:

  1. Thank you for sharing this article about SaaS and cloud computing, I learned a lot from it.
    us vpn

    ReplyDelete

Comments welcome!