{"id":428994,"date":"2010-03-15T09:00:03","date_gmt":"2010-03-15T13:00:03","guid":{"rendered":"http:\/\/gigaom.com\/?p=105653"},"modified":"2010-03-15T09:00:03","modified_gmt":"2010-03-15T13:00:03","slug":"appistry-joins-cloudscale-storage-fray-and-brings-hadoop-with-it","status":"publish","type":"post","link":"https:\/\/mereja.media\/index\/428994","title":{"rendered":"Appistry Joins Cloudscale Storage Fray, and Brings Hadoop With It"},"content":{"rendered":"

\"\"<\/a>Appistry today added another element to its cloud-computing application platform<\/a>, announcing the April availability of CloudIQ Storage. With the release, the St. Louis-based company joins the growing ranks of companies seizing on demand cloud storage solutions that can maintain performance in the face of rapidly growing data volumes. Appistry hopes to distinguish its scale-out storage offering from the competition, however, with two key innovations: (1) an ace-in-the-hole that it calls \u201ccomputational storage,\u201d and (2) CloudIQ Storage Hadoop Edition.<\/p>\n

Customers can achieve the performance benefits of computational storage by launching a commodity-server-based cloud with Appistry CloudIQ Engine, and installing CloudIQ Storage on the same pool of servers. This is a big change from the standard model of having separate islands of processing and storage<\/a> connected by what Appistry VP of Product Management and Marketing Sam Charrington calls \u201ca straw.\u201d In a computational-storage model, processing tasks automatically route themselves to the relevant data, wherever it\u2019s located across the file system. Charrington says this capability lets applications access data at bus speed (i.e., within the same box), thus eliminating the network bottleneck.<\/p>\n

With the Hadoop Edition, Appistry hopes to \u201cupgrade\u201d the performance and availability of Hadoop-based applications by replacing the Hadoop Distributed File System (HDFS) with CloudIQ Storage. While Hadoop is wildly popular<\/a> right now, one issue is its use of a \u201cnamenode\u201d \u2013 a centralized metadata repository that can constrain performance and creates a single point of failure<\/a>. Appistry\u2019s approach retains Hadoop\u2019s MapReduce engine to assign parallel-processing tasks, but attempts to resolve namenode problems with CloudIQ Storage\u2019s wholly distributed architecture.<\/p>\n