How to get Started with Big Data

Dave:  Rackspace releases its OpenStack private cloud software for free

Rackspace today released Rackspace Private Cloud Software, thus offering organizations a free download of the same complete version of Essex OpenStack that the company runs in its own hosted private clouds.

But Rackspace's ambitions here are clear: To entice cloud-curious companies to kick OpenStack's tires at their leisure and on their own equipment, with the hopes of getting them to sign up for Rackspace's support services.

By Rackspace's reckoning, services are going to be the big differentiator among the various versions of OpenStack that companies like Red Hat are rolling out.

Chris: Big Data Programs: How to Get Started

It’s an interesting article listing some obvious and some less obvious standard practices for innovation around big data programs.  I think the discussion around this kind of topic is worth having, because a lot of organizations are working through this right now.

A shift in data management is under way. For the business world, the days of the relational database as the de facto data management approach are coming to an end. Instead, enterprises are plunging headlong into a dynamic new era of data explosion and innovation. As a result, many elements involved in data management are changing –the data itself, IT’s approach to the data, the people working with the data, the technology options and the skills needed to support new applications. There are new opportunities to gain insights from data for business decisions and success, but with this comes the need for change. Organizations everywhere must implement big data programs, but with so many moving pieces, it is unclear to some how to proceed or where to start.

1. Define the business drivers
2. Discover the data and it’s location
3. Plan for iterations and expect some to fail (unlearn traditional relational db techniques)
4. Provide access to the data, not just DBAs any more
5. Leverage new tech - NoSQL, Hadoop, NewSQL
6. Plan for success - transition to production and transition to operations
7. Publish / subscription capabilities - wrap your data in hardened SOA
8. Take a hard look at roles - data scientist team for example may be new
9. Monitor industry, keep lab going testing new capabilities as they come every month

JP:

I want to pick up the mantle from Dave’s Infoworld blog about the 3 missing pieces in cloud computing. I’ve been asked quite often lately by customers about what can really be achieved today; especially after delivering the cloud journey vision. There’s still a lot of scripting and the tools Dave mentioned are not well-integrated with available orchestration and automation tools limiting policy-based cloud decisioning. Moreover, most of the cloud managers today want to own the resources available to them, which takes them out of the pool for non-cloud apps. e.g. difficult to share full storage pool with AIX and big iron.