The cloud’s role in processing big semantic data sets was recently highlighted in early April when DERI and Fujitsu Laboratories announced a new data storage technology for storing and querying Linked Open Data that resides on a cloud-based platform (see our story here).
The cloud conversation, with storage as one key discussion point, will continue to be an active one in Big Data circles, whether users are working with massive, connected Linked Data sets or trying to run NLP across the Twitter firehose. CloudSigma, for example, recently publicly disclosed that it is using an all solid-state drive (SSD) solution for its public cloud offering that lets users purchase CPU, RAM, storage and bandwidth independently. The use of SSD, says CEO Robert Jenkins, avoids the problem that spinning disks have with the randomized, multi-tenant access of a public cloud that leads to storage bottlenecks and curbs performance.
That, combined with the company’s approach of letting customers size virtual machine resources as they like, as well as leverage exposed advanced hypervisor settings to optimize for their particular applications, he says, brings the use of the public cloud infrastructure closer to what companies can get out of private cloud environments, and at a price-performance win.