Deriving insights from your unstructured business data (files and objects) has become an obvious and ubiquitous business requirement. Typical legacy deployments or deployments not planned for having analytics as an integral part of their architecture tend to end up with silos of of unstructured data repositories. Such deployment face various issues:
– addressing issues of deploying separate clustered (like HDFS clusters) where the file or object data has to be explicitly copied ,
– addressing issues of extra storage has to be explicitly provisioned
– addressing issues of provisioning and maintaining network bandwidth consumption for regular data copy has to be planned
Swallowing bitter pill of:
– delays in able to get instant analytic insights from file or object data
– higher TOC need to accepted to be higher
– administration overhead of multiple Silo’s of unstructured data storage
Why would one want to get into this mess ? Would it not be nice to have a storage system that supports unified file and object, has inplace analytics support via Hadoop connectors, performance well, is scalable , has ability to seamlessly tier to other object stores or tape and is software defined. It sounds like a No Brainier !
Here is a IBM Spectrum Scale offering (rated as the number one filesystem in 2017) which address all the the above. You can also download a free trial VM to take a look at it.
Here is a paper “A Deployment Guide for IBM Spectrum Scale Unified File and Object Storage”
and here is a presentation which insights on how Spectrum Scale can help with being the platform for hosting unstructured data that supports unified file and object, has inplace analytic support via Hadoop connectors, performance well, is scalable , has ability to seamlessly tier to other object stores or tape and is software defined.
– Sandeep Patil