Live Blogging: Notes From IBM’s PureApplication Systems Announcement


The PureSystems Manufacturing Team

IBM has drawn on its vast expertise of applications and workload knowledge and best practice from the field and its partner ecosystem, codified this information and packaged it into its converged offering called PureApplication System. IBM believes this is a major differentiator from other converged solutions and will deliver substantial value to customers. This was the message from IBM executives at a private analyst event prior to the IBM public announcement. What follows are notes from the session with points of view inserted.

My PoV: I agree. If in fact IBM has been able to package this knowledge into systems that can be self-optimizing then this is huge.

Beth Smith, VP WebSphere

IBM PureApplication System – and patterns of expertise tuned for the system.

Could have worked across groups and created ‘bundles’ but in doing so IBM would have given up much of the value it is delivering.

80% of the world’s apps workloads are covered in this announcement.

Three Principles:

  1. Built in expertise
  2. Integration by design
  3. Simplification by experience

–>Wondering to myself how IBM codifies all this experience. How does it get from the field to the product?

Built-in Expertise – capture and automate what experts do – Patterns of pre-integrated, out-of-the-box IBM and partner patterns for deploying application workloads based on proven best practices and expertise. Bottom line is reduced management and change management costs.

Integration by Design – deeply integrate and tune hw and sw. Integrated provisioning, zero downtime for capacity upgrades, system maintenance, software updates, integrated analytics to optimize db query performance eand adaptive db compression that can provide up to 7x or greater space savings.

Simplified Experience – Make every part of the IT lifecycle easier – integrated and open solution – single pane of glass view from the infrastructure up through the app platform designed to be up and running in four hours

Complete and ready to go systems – running in lt 4 hrs – 4 power cables, 4 nw cables. ;pre-optimized and pre-configured for app workloads.

Apps – Websphere and DB2 apps – platform as a service

Integration by design – optimizes the complete solution stack – all hw and sw components factory integrated and optimize. Virtualized across the stack for efficiency, unified mgt, monitoring and maint. Integrated and elastic app and data run times – servers, networking, storage, virt, mgt, development, mw, deployment, apps

IBM shows some proof points wrt to the value prop. – would like more detail on the before and after. It’s always hard to isolate the value contribution from the various piece parts.

Jason Gartner, Director, Java & Troy Application System Product Management

Business wants application that scales and connects to the corp database and scales.

Users have to deploy http server, app servers, caching servers, database, LDAP – so what businesses do is build to the peak. What IBM’s done w/patterns of expertise is it has encapsulated its knowledge and capabilities in patterns. Connects monitoring, security, OS, management – packaged as a single entity. When these components are bundled and managed as a single entity any changes effect the entire system, not just the piece parts.

–>So this would seem to address on the devops question. Specifically, if I spin up virtual dev instances, infrastructure configuration nuances can be set up by developers with visibility on the production infrastructure such that when I go to deployment the system is properly configured for the production environment. The key is X-training the dev and ops team so that the same individual that writes code can also configure the infrastructure.

Example – a Web app. Want to deploy the Web app, needs a database. Can use existing database or deploy new one w/ a schema. Let’s choose the latter – new DB. Slider bar pops us to determine response time required. If RT goes below the threshold system resources are re-allocated. This is carried through from dev to test to deploy.

My PoV: The ability to embed pattern knowledge into the converged system is a differentiator that will facilitate the move to devops and that’s where IT organizations will achieve hyper productivity. By capturing the content from IBM’s field and partner ecosystem, IBM is taking advantage of significant experience and intelligence that can be embedded into the system to simplify operations, reduce re-work and deliver clear opex savings. As well, this system can be a platform for ISVs and VARs.