Tuesday, June 17, 2008

HP marries change management and problem isolation functions into an automated data center efficiency partnership

Borrowing heavily from its Opsware, Mercury and Bristol acquisitions, HP on Tuesday at the Software Universe conference announced products and services designed to automate and coordinate two thorny aspects of large-scale IT operations: change lifecycle management, and problem isolation and resolution.

While once mundane and esoteric aspects of running monolithic data centers, today's scale, complexity and far-flung fragmentation from services orientation have elevated managing change across webs of servers and components to a top priority. What we're really getting at here is making IT perform like a mature, refined and managed business function, not a near-sighted firefighting brigade.

In many cases, operators and IT executives are forbidding that requested changes can be made to services and applications for fear that the changes will stir up hard-to-locate and tough-to-remedy glitches. Such unintended consequences can be scattered across thousands of distributed servers and IT network devices. It's hard to enjoy the fruits of service oriented architecture (SOA) investments for business agility when the IT infrastructure is too brittle to accept applications-level change readily.

So in anticipation of dampening further hamstrung agility advances due to "brittle environments" -- particularly as SOA, virtualization and cloud computing come into play -- HP's Business Technology Optimization (BTO) and research teams have assembled what amounts to IT change confidence enhancement tools. [Disclosure: HP is a sponsor of BriefingsDirect podcasts.]

A recent study by The Economist research unit showed that some 50 percent of service outages in data centers was due to a change being made to a service. Then these problems were hard to isolate given that no one knew how the change impacted the distributed systems. From there, some 68 percent of companies responding said the applications issues were being tackled manually.

Obviously HP sees a huge opportunity here for making modern data centers behave more like assets and less like liabilities, at least in the eyes of business managers as they seek process and strategic initiatives changes.

So not only does the confidence to change services and processes freely amid SOA-supported processes need a boost, the ability to manage change requires automated lifecycle depth and breadth. Uncoordinated and manual attempts to manage change amid widespread complexity can actually make the problems and their resolution harder. Next generation data centers require an end-to-end services deployment and change management capability that maps to service workflow orchestration, services management and datacenter automation activities.

So the HP BTO and adaptive infrastructure engineers have designed HP Release Control 4.0, which identifies "change collisions" and creates a managed services approach to change. The solution not only manages technology, it manages the people managing the IT via providing advance visibility into change impacts and organizing teams so changes are coordinated.

The value does not end at the proactive stage of change management, but provides the tools to identify change-related issues over the lifetime of the services, said Sharmila Shahani, chief marketing officer of HP Software. "This provides a proactive, real-time and automated way to manage the change lifecycle," said Shahani, in an interview.

Additionally, HP has announced Business Availability Center (BAC) 7.5 for improved problem management in complex datacenter environments. Using new technology from HP Labs, the product helps isolate runtime problems before disruptions by allowing fast and visual "drilling down" into operations data regardless of scale and complexity.

Other new product releases here at Software Universe this week include: HP Client Automation Center 7.2, HP Storage Essentials 6.0, and HP Service Automation Reporter 7.0.

What's interesting to me is that HP is using he change management-focused Release Control functions in association with the BAC problem resolution functions, getting into a data center dance of efficiency. As the change piece and the problem identification piece are used in unison, a "closed loop" approach to datacenter performance amid constant change becomes possible, said Shahani.

I did a podcast interview just yesterday at the HP Technology Forum event in Las Vegas in these issues with Duncan Campbell, HP's Adaptive Infrastructure program leader. Have a listen. More podcasts from Software Universe are here.

Further burnishing the datacenter efficiency shine, HP has also updated its configuration management database (CMDB) system to embrace federation and ITIL v3 principles. Universal Configuration Management (CMS) 7.5 allows for many versions of configuration data from many sources to be used in unison for improved visibility and access for what's going on in as many of the systems as possible in nar real-time.

HP's latest CMDB does not force all the config data into a common CMDB, but rather uses connectors to other CMDBs for true federation on a meta data level, said Shahani, to provide a hub and consolidated view of all components within a large distributed system. Universal CMS 7.5 arrives this month.

HP is targeting the aggregated view of all systems elements value from the new CMDB at the burgeoning use of virtualization across datacenters. Virtualization promises utilization efficiencies and automated provisioning of services and applications support, but it also adds complexity as support infrastructure and application instances can pop in and out of use (existence?).

What I especially like about these new products is that they can increasingly be used in association with SOA governance and SLAs to begin to get to a true services lifecycle approach value and benefit. Used in association with HP SOA Center (including the Systinet repository), architects can integrate design and governance demands with change management, problem management and federated systems config data for a whole significantly larger than the sum of the parts.

It's just these kinds of complete services management capabilities, increasingly automated, that will make SOA pay off big dividends and paves the way for use of private cloud compute environs for enterprises and service providers alike.

No comments:

Post a Comment