SGI uses Ivy Bridge to build Hadoop cluster and NoSQL equipment

As early as 1980, Silicon Graphics has become a well-known brand in the field of High-performance graphics workstations, after a series of 2009 years of restructuring and acquisition of assets from Rackable Bae, Silicon Analysys (SGI) has shifted the focus of its business to High-performance computing servers, clusters and devices. As Intel releases the Ivy Bridge Xeon processor, SGI is ready to build a server product based on this new CPU family.

SGI has been providing dedicated hardware for the Hadoop project for a long time, including preloaded Cloudera Apache Hadoop releases (CDH) and devices dedicated to MarkLogic NoSQL databases. These products are now ready to meet a new wave of performance upgrades from the Ivy Bridge Xeon CPU, Gigabit Ethernet, or InfiniBand high-performance node network.

A rigorous approach to Hadoop

SGI has been treating hadoop with a rigorous attitude. In an interview previously launched by ZDNet to the top net, Bill Mannel, vice president of SGI product marketing, explained this. SGI not only build hardware and software, but also implement the following scenarios for the SGI Hadoop cluster:

After the CDH is installed in the cluster, SGI tests each component from the system level, not the rack level;

The company employs Cloudera certified Technical personnel responsible for providing customers with the ideal cluster installation services;

SGI also has a professional service team, working with Cloudera certified developers to deal with MapReduce and other large data coding and development.

SGI claimed that its previous generation of Sandy bridge Hadoop clusters maintained a performance record of the Hadoop Terasort benchmark. It is clear that the company is looking for a performance upgrade through Ivy Bridge to keep its products in the lead in the future.

NoSQL.

SGI a rigorous approach to Hadoop and NoSQL. SGI and MarkLogic Company established an exclusive partnership, so we can not see in its products Datastax/cassandra or MongoDB figure. The NoSQL devices offer One-fourth, One-second, full rack and even multiple rack configurations, and are rigorously tested at the system level like the Hadoop cluster. SGI will also offer customers a development model that includes one, two, or three nodes to help them gain a nosql experimental environment.

SGI currently has a range of U.S. government agency clients who have built large Hadoop clusters. SGI also provides a large Hadoop cluster for the E-commerce giant ebay's PayPal division, which is designed to help PayPal conduct a fraud detection analysis.

SGI's commitment to the Big data field shows the company's high level of confidence in the mainstream market of NoSQL and Hadoop. In addition, SGI will also be in the increasingly prosperous Hadoop turnkey solutions in the field of a fist. In this case, SGI no longer as an enterprise software vendor, but as a high-end hardware vendor to deliver turnkey solutions to customers.