ITIL

Hi,

We're using Semantic Data Theory and Semantic Data Modeling concepts to design and build a very light-weight and low cost implementation of a CMDB that rapidly renders itself as a fully functional and highly organized web-site with advanced visualization features (trees/graphs/charts/etc.). The semantics are being used to automatically control the organization of data and to generate constructs like the CI Catalogs & Inventories, the CIs themselves, and Relationships between CIs.

I'm wondering: Is anyone else using Semantic Data Theory and Modeling for their CMDBs and, if so, what kind of challenges and successes have you had doing so?

Also, if you are, are you using OWL(WOL)/RDF semantic web standards or baking your own?

Thanks for any assistance you can offer.

My Best

Looked it up

Never used it as far as I know

I build DBs based on the information needed to be in there

Whether it is using this theory / model or not, shrug

UKVIKING wrote:Looked it up

Never used it as far as I know

I build DBs based on the information needed to be in there

Whether it is using this theory / model or not, shrug

Hi John,

I'm assuming that using or building upon "the information needed" is the traditional method of relational data modeling, which implies that for every CI Type, I need to add a set of new and dedicated tables. Because the number of CI Types is theoretically endless, the traditional form of modeling does not scale because of the following pattern...

Every time I add a new CI Type and assuming I'm building a full-featured CMDB Web Application, I have to...
- Add a new elements to the database
- Add new Data Access and Value Objects
- Add new elements to the Application Business Rules
- Add new elements to the Presentation Layer of the Application
- Add new elements to the User Interface
- Add new relationship definitions
- For every application layer, above, integrate all of the above for the new CI Type to all previously existing CI Types, based on valid/acceptable relationship definitions
- Create new CI Type related Catalogs, Indexes, Reports, etc.

The problem with this is pattern is that it clearly doesn't scale well, as every new CI Type or Category requires a new/more work (ultimately translated to "money") to truly make the CMDB "rich" with usable features related to that new CI Type and all the CIs of that type that get loaded into the system.

Semantic Data Theory and Modeling creates a far more "generic" model that:
- Deals with Types of Things
- Specific Things of a specific Type of Thing
- The harvesting or specification of any relevant relationship between any to Things (even if both sides of the relation are the same Thing)

We've gotten the application to work using these concepts (although we've also gone the route of totally eliminating the need for Databases and App Servers, which is a whole different conversation). It took us about 7 months to get to Alpha state and the last two months have been about making sure that we can see a full 360 Degree view of any CI, traverse any relationship between any two CIs, and create many different views and reports. The goal is to start layering advanced visualizations over the data, such as Visual Matrix Mappings, Trees, and Graphs (a.k.a Network Diagrams). The entire output renders itself as a Web Site that looks feels and acts like a fully functional Intranet. This is how we maximize the user experience.

I'm wondering if anyone else has gone the route of using Semantics as their baseline and seen the same sort of productivity we have. And, if not, what were/are the issues they encountered, hopefully so we can use their experiences to head off any unforeseen issues we might run into.

BTW, If you're ever interested in seeing it, just PM me and I'll gladly set up a WebEx session to walk you through it. Would certainly appreciate any professional feedback.

My Best

IF4IT

You asked a simple question about Semantic Data Theory

I answered.

What was not needed is what you will / would do as sort of validation

I use tools - Ticket systems, etc. I use the db structure within the tool if it meets my requirement. If it does not meet my requirements, I dont buy the tool

I dont develop the DB from scratch when I can purchase one that has been made

Whilst I develop for myself DBs in relational databases, that is completely different

UKVIKING wrote:IF4IT
I use tools - Ticket systems, etc. I use the db structure within the tool if it meets my requirement. If it does not meet my requirements, I dont buy the tool

I dont develop the DB from scratch when I can purchase one that has been made

Hi John,

Yes, have gone that route multiple times as well. It's always been a slow and expensive route. The challenge always seems to be when a customer wants to do things like import CIs that are not "technical" in nature, such as: Business Requirements, Use Cases, Test Cases, Glossary Terms, Capabilities, Industries, Business Segments, Countries, Locations, etc. Do you ever experience these types of requirements or challenges with your customers?

Also, may I kindly ask if you believe there are any 3rd party DBs or Application that can be purchased which allow you to quickly and affordably handle such complexities?

My Best

Most IT SM tools have a Asset Register or CMDB built

I have used several

each one is different and has it benefits and drawbacks

I prefer the one that meets 80 % of my requirements