IBM PureData System for Transactions

Kelly Schlamb, DB2 pureScale Specialist, IBM

By now, I’m hoping that all of you have had the opportunity to hear or see something about IBM’s latest addition to the PureSystems family – the PureData System.  Generally available in October 2012, this new system comes in three different models that have been designed and optimized to handle the different types of workloads that your organization typically needs to run: transactional analytics, and operational analytics.

If you have seen something on PureData before now, you have likely seen the following things mentioned:  built-in expertise, integration by design, and simplified experience. There’s a lot of great information out there talking to these points (and I’ll provide some links below) so I won’t spend too much time on them. However, a key take away is that these systems are “simple”.  That means simple to order, setup, configure, manage, upgrade, etc. These systems are hardware/software integrated, shipped in the rack, fully assembled and they just need to be powered on and connected into your corporate network.  Deployment is hours, not days.

Having a DB2 pureScale background, the system I’m personally the most interested in is the PureData System for Transactions… a platform that provides transactional data services, with DB2 pureScale at the heart of it.  For those of you who aren’t aware, pureScale is the DB2 feature that provides extreme scalability and availability in a shared data environment.  It could just be my background, but I think it’s very cool technology.

In a previous role I was actually one of the pureScale developers and development managers, but in my current role I’m now working directly with clients who are in the process of bringing pureScale into their IT infrastructure.  This has given me a unique opportunity to both work with this technology at a low bits and bytes level, but then also see first hand how clients are using it to power their core business applications.  And now with the new PureData System for Transactions, I’m really excited about how it is going to open up further opportunities for our clients who need this type of extreme availability, but perhaps were a bit hesitant to take on the task of standing up a pureScale system themselves.

The PureData System is designed to significantly improve the time to value for new application deployments because the system is already pre-configured and tuned for OLTP workloads. I was recently at the Information on Demand 2012 conference in Las Vegas and there was a lot of buzz around all of this. I had a lot of conversations with folks who were very excited about getting their hands on one of these new boxes.  They told me that more and more of their applications are becoming business critical and high availability has become a necessity.

For more information on the PureData System for Transactions, check out the following pages and videos:

IBM PureData System for Transactions

IBM PureData Systems

IBM PureData System for Transactions Tour with Tim Vincent :

IBM PureData System for Transactions :

Before signing off, I just wanted to quickly share a quote I saw in an article – just to highlight the kinds of things that analysts are saying. The analyst here says that the new solutions will enforce Big Blue’s current market and thought-leadership position: “That isn’t a bad thing, unless you happen to be one of the myriad companies traveling in IBM’s wake… But over time, we expect PureData and future IBM solutions to inspire what amounts to a template for what enterprises will come to expect from transaction processing and business analytics solutions.”

4 Responses to IBM PureData System for Transactions

  1. Brendan says:

    We have a workload that first needs to handle a lot of updates and then secondarily run a set of queries against that. DPF would seem provide the most throughput given separation down to the I/O level. However my impression is that IBM is pushing PureScale for transactional loads of any kind and de-emphasizing DPF to for anything except warehousing. True/false?

    • Kelly says:

      Hi Brendan. The PureData philosophy is that different workloads are best handled by systems designed and optimized specifically for those workloads (e.g. transactions, analytics, operational analytics). In the case of the PD transaction system, it does utilize pureScale and pureScale is primarily aimed at transactional workloads. Of course, rarely do you have a pure (pardon the pun) workload. Even a transactional system typically has a batch component to it. And perhaps there are some queries that need to run against the transactional system. That’s all fine with pureScale (and with the Pure Data System for Transactions). However, if the workload is primarily complex queries, analytics, scans involving large amounts of data, etc. then an MPP type of configuration (such as DB2 DPF or Netezza — or the corresponding PureData models) is likely better suited for it. And if you look at the PureData System for Operational Analytics, that’s a system that’s built to do analytics and large queries, but also allows for continuous ingest of data into the system (and that could be inserts, updates, deletes, merge, etc). Batch updates followed by large query workloads is not what one might classify as OLTP, but you might have left some details out. We would need to understand more about your update and query workloads to determine what type of system would best suit your needs.

  2. Joe says:

    I’m wondering if PureData for Transactions could be used in an Geographically Dispersed Cluster (2 PDTX – one on each site)? If not HADR with DB2 10.5 might be a solution – any ideas when PDTX will support DB2 10.5

    • Kelly says:

      Hi Joe. GDPC is not a configuration that can be done with PDTX systems today. A refresh of PDTX to include DB2 10.5 (including the ability to do HADR) will be available in the not-too-distant future. Looks like a common friend is in the process of setting up a call so we can talk more about this then. 🙂

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: