AIR Wiki : InformationModels

HomePage :: Categories :: PageIndex :: RecentChanges :: RecentlyCommented :: Login/Register

Information models


There are several information models available for describing clients, servers and their networks. Below is an overview of the different models as far as we are familiar with them.

Please contribute if you find errors or know of other models!


Management Plane Models


The book "OpticalNetworkingControl" has the following to say on information models (section 13.4.1.1, page 415):
The bulk of fault management is centered on notifications, alarms, alarm status, alarm logging, and alarm correlation, and is orthogonal to control plane activities.

Management Information Base


MIB is one of the first information models. It is based on the object-oriented approach of describing managed objects. However, it mostly aims at describing management information that can be read out using SNMP, for example, "output queue length"(ifOutQLen) or "Address Translation table" (atTable). And its main use is in event notifications and alarms.

MIBs have been in use for quite some time, mostly because of the popularity of SNMP. It allows for vendor extensions and most implementations use these extended MIBs exclusively.

In theory SNMP and its MIBs can be used to configure devices as well, but this is almost never implemented or enabled.

Common Information Model (CIM)


CIM is an effort of the DMTF (Distributed Management Task Force) to define a common definition of management information. It is part of the WBEM (Web Based Enterprise Management), which also provides a whole suite of management applications.

It is a huge model, spanning over a dozen documents. It has been implemented, has a command line interface, mapping to XML, and a query language. It also allows for vendor extensions.

I tried to get a grasp on the model and to describe networks using it, but it seems that CIM does not contain enough expressiveness to completely describe networks as we would like to do. Furthermore, the model is so complex that it takes a huge effort to grasp it and to become able to actively contribute to it. And to make matters worse, new versions come out at a rate of once per six months.

DCML (Data Center Markup Language)


Standardising approach by an OASIS committee. It is an approach to _completely_ describe data centers, this includes infrastructure, but also servers, storage, power and applications.

It makes use of RDF and OWL to define and describe messages. Currently I have only seen a document that describes some global objects and the messaging format, but nothing yet on the model itself.

The website has not been updated since 12 August 2005 (and date before that was 3 November 2005).

Difference between CIM and DCML


(CIM) is a Distributed Management Task Force (DMTF) standards effort for describing overall management information in a network/enterprise environment. CIM defines an abstract data model, method for instantiation in XML, and mappings to other management and information standards like SNMP and the Lightweight Directory Access Protocol (LDAP). CIM's focus is on defining abstract models of managed objects to be used by monitoring and other traditional management systems. While CIM is quite comprehensive in this area and therefore overlaps somewhat with DCML, CIM is not well-suited to the data center automation problem that is DCML's focus. DCML describes not just the state of an environment, but also how to construct the environment and the policies governing the management of that environment, neither of which CIM describes. Where CIM concepts and data elements can be mapped onto DCML for use with data center automation tools, DCML references CIM to avoid duplicating work.
(Taken from DCML FAQ)

RDF approaches


The two models described below are models written in RDF. These models were created because of a need for either generating DHCP and DNS configurations or as use for (simple) topology description for a network simulation tool.

Both models are somewhat "naive" and do not scale to large scale networks.

Using RDF for Home Network Configuration


Naive approach to describe home computer networks so that DNS and DHCP configurations can be generated from it.


Network Topology Description


Seems to be abandoned.

Was used as Layer 3 topology description for eWAN, a high performance network environment emulator.


Theoretical approach


ITU G.805


Theoretical model describing networks and their connections. Is purely aimed at networks and connections, does not say anything about nodes in the network, other than the information that can be derived from the described connections.

Category

CategoryTechnologies

There are no comments on this page. [Add comment]

Valid XHTML 1.0 Transitional :: Valid CSS :: Powered by Wikka Wakka Wiki 1.1.6.0
Page was generated in 0.0305 seconds