As pressure on IT organizations to become more agile increases, servers and storage systems inside data centers are being turned into appliances that can be deployed in a matter of minutes. Looking to capitalize on that shift, Hitachi Data Systems (HDS) has unveiled 2U server and Flash storage appliances that can be provisioned in a matter of hours.
Bob Madaio, senior director of product marketing for HDS, says the 400 Series edition of the Hitachi Hyper Scale-Out Platform (HSP) is specifically designed to handle Big Data application workloads, including the Pentaho business intelligence software that HDS acquired last year. Based on two Intel Xeon E5-2620 series processors, the 400 series can be configured with 192GB of memory, 8GB of Flash memory and 72GB of magnetic storage.https://o1.qnsr.com/log/p.gif?;n=203;c=204663295;s=11915;x=7936;f=201904081034270;u=j;z=TIMESTAMP;a=20410779;e=iMaking use of REST application programming interfaces (API) and a management framework based on OpenStack, Madaio says the 400 series is the latest example of a software-defined appliance that provides IT organizations with a platform that can easily scale out to handle Big Data analytics workloads as they grow over time.
Meanwhile, a new A Series addition to the Hitachi Flash Storage line-up takes a similar appliance approach to primary storage. Based on a pair of controllers that can be configured with up to 60 solid-state disk drives, the A Series storage appliance provides access to 384TB of addressable capacity at a rate of one million IOPS.
In general, Madaio says, appliances are crucial to bi-modal approaches to IT management that enable IT organizations to more flexibly respond to increased demand for both compute and storage capacity. While most application workloads continue to run on premise, Madaio notes that IT organizations in the age of the cloud are under pressure to make IT resources available whenever and wherever needed. The challenge IT organizations now face is figuring out how to manage various application workloads differently based on characteristics that in some instances lend themselves more to appliances rather than traditional servers.