CMDB Modeling

Recently, IBM, HP, Fujitsu and BMC launched a CMDB standardization initiative. The goal of this effort is to define a standard allowing the exchange of information between CMDBs. The standardization can happen at different levels, as an API, as an exchange protocol, or a data model.

Lets look at the current state of affairs in that space :

CMDB standard space

It might look a bit like an alphabet soup, so here is the explanation :

  • Sun has a product called N1 Service Provisioning System, which, while not really being a CMDB is able to model applications and associated artifacts in a language called CRML (CenterRun Markup Language). N1 SPS is more like a release management product coupled with a DSL, a Definitive Software Library. This markup language is expressed in XML using XML-Schema.
  • Opsware has a CMDB in their Asset Tracking Edition. Opsware is a co-founder of the, now, OASIS TC defining the DCML, the Data Center Markup Language. DCML is based on RDF, the Resource Description Framework. RDF allow the definition of a taxonomy of data center objects and their relationships. DCML was not designed from its inception as a CMDB description model, and is trying to play catchup now.
  • BMC is providing a CMDB, Atrium, built on top of a DMTF CIM model. This model could be exported using the WebService / CIM mapping: WS-CIM (now part of the WBEM Infrastructure and Protocols WG).
  • CA seem to also use DMTF CIM in their CMDB.

That's for the part about the CMDB vendors. Now, as for the modeling and the repository :

It will be interesting to see where the new initiative will land in term of technology and standard body.

Tags :

MIT Sloan Article on Grid Computing

In an article on Grid Computing appearing in Fall 2004 MIT Sloan Management Review, Heather A. Smith and Benn Konsynski explore the challenges of Grid Computing.

While taking an IBM centric view, focusing on Globus for example, the article raise several interesting issues :

... the internal integration that is achieved with technologies such as enterprise-resource planning (ERP) and customer-relationship management (CRM) is not going to help organizations evolve in the future. Hardwired technologies will actually ossify existing organizational structures, preventing the flexibility companies will need to operate in a network-centric environment.
... centralized command-and-control structures will no longer be effective.
Top Managers will therfore have to relinquish some degree of decision making control in favor of an accelerated ability to respond.
... some CIOs will find their jobs reduced in scope and will become IT directors.

I'm not sure about the last one, except maybe if we think that, in the case of global grids, each grid CEO become an IT director for the global grid ?

Hyperic HQ goes Open Source

Today, Hyperic announced the open sourcing of their HQ product.
Hyperic HQ is a management solution based on J2EE providing the following features :

  • Discovery
  • Monitoring
  • Inventory

They support an extensive list of platforms and applications.

This announcement is to put in perspective with the recent announcement of the OpenManagement consortium bringing together the following OSS projects :

  • OpenEMS : Open Source Enterprise Management Services
  • OpenQRM : QLusters' management solution
  • OpenSIMS: Security Infrastructure Management System
  • Enomalism: Xen Virtualized Management Console
  • Webmin : Operating System Management solution
  • NetDirector: Configuration management for network services
  • NodeDirector: User and Host management
  • Nagios : Host and Network Monitoring program
  • Zenoss: IT infrastructure monitoring software

The space of open source management products has really changed.