“Atlas for Lotus Connections is a social networking application that allows users to visualize their current network of contacts and see how they can extend that network to tap into valuable resources and trusted experts across an entire organization.“

http://www-306.ibm.com/software/lotus/services/assets.html

I think this is just one example of how the social networks which have been parading around the web for the next couple of years are likely to manifest themselves in the enterprise.

For a long time vendors have been modelling the relationships between configuration items (sometimes entire systems) and their business communities using formal organisational data.

In the new age where it seems to be becoming acceepted that a business ‘edge’ may be gained by allowing the rapid formation of dynamic communities of like-minded individuals, perhaps the future of understanding and servicing the needs of our clients is within understanding these new communities and their technical dependancies ?

That is a very funky idea.  (Hi, Dave!)

A lot of the ideas behind social networks make a lot of sense for people collaborating within the company - groups that you can easily join/leave, the ability to freely create accounts and so on.

Things we’d need to think about:
1) Licensing - if new users could easily add accounts for themselves then licensing per user might not be very applicable.
2) Permissions - if users are creating their own accounts then issuing them with perms to see what they need to see needs to be done rather more carefully.

Hi Simon,

Why do you feel accounts would be a factor ?

- I’d have thought most systems within an enterprise would be using some kind of integrated AD authentication.

The picture I had in mind was that of enterprises forming communities for competitive edge made up of what would traditionally have been either organisationally or geographcially diverse individuals.

These communities will be looking to accomplish a common business process, which in turn would require the support of certain business systems.

Maintaining visibility of the dependancies between the different types of communities that are being formed in large enterprises and their key systems is a a huge challenge, not just for traditional internal support organisations but also for configuration management products as those I’ve worked with don’t expect businesses to be multi-dimensional.

How would these communities form?  Would the product’s admin create a new group/community/channel, or would a user decide that they were going to create this group?  I can see the advantage of users being able to create their own groups.

I guess the question that I have is around permissions and security: most CM products are only used by one or two users.  What we’re proposing here would be to make a CM product that a LOT of users could access - would there not be issues with people having access to potentially sensitive configuration information?  The running versions of SSL libraries, for example?  Would the product need a really rich set of permissions to block people from accessing certain data items?

What we’re proposing here would be to make a CM product that a LOT of users could access - would there not be issues with people having access to potentially sensitive configuration information?

I’m not sure we’re on the same page - I’m not suggesting that the CMDB should interact with all the users, just that it should be able to model the relationship between a layer of technology, the business processes it supports and the communities which are reliant upon that process.

I believe that most large enterprises will find a way to model the scope of communities, this being mnecesary in order to deliver common software packages or services, a flexible integration for this layer of the CM system that allows it to be populated by an AD or Notes group or by some other type of list server would therefore be requisite, as would I think some kind of understanding that the members of a community will most often (but not always be) be relatable (is that a word ?) to the records in an HR system, this would probably be useful when trying to model chargebacks.

It may also be useful to be able to model communities on multiple levels.

You can probably tell I’m well into space cadet mode at this point…

As an example, the Sales & Marketing community would most likely be made up of client facing individuals across the business, with no consistent relationship to organisational units but likely with an identifiable relationship to a list or an AD global group, this community all rely upon the sales and marketing process, which is made up at the technology level of the CRM system and likely some other systems related to pipeline and financials.

An understanding of the relationship of the community to the underlying technology allows the impact of change or outage of an individual CI or to a system or even to a relationship between systems to be understood in the context of how well it supports tjhe business.

The ultimate goal may be to make that understanding even more sophisticated - the business at certain times will have a higher reliance upon a certain community, (at end of quarter or end of year the sales or financial communties may become more important), wouldn’t it be interesting if the technology management systems understood these changing needs…

Ah, I see what you’re saying. 

Capturing and maintaining the structure of a community in a product would be very time consuming - we can’t run scanners to see which people the other people tend to communicate with! We could model dependencies between users and systems by scraping access logs (where these are available/accurate/exist), but only at a person <-> system level.

I suppose that you could hook the system up to your AD system and point it at certain lists and workgroups that people are members of to get community links. 

Is it reasonable to say that if a member of a community has a dependency on a system then the whole community does?  If somebody is a member of communities a, b and c, and that person uses system Z, then all three communities depend on system Z?

It seems to me that the best way to find out what communities people are part of is to adopt the social network approach, and let people join the online communities that they’re interested in.  If you give them the normal social network functionality then they’d probably use the product, and this would give you a way to spot the groupings between them.  How then would you hook these groups into dependencies on systems?  Tricky…

Appreicate that this sort of thing would be difficult today, I was more thinking the day after tomorrow…

my thoughts are that Microsoft and IBM will continue to evolve their collaborative technologies into community hubs, if they do their jobs right then it should allow with relative ease the type of loose integration you are referring to.

The Tideway angle would be the management of the CIs, their relationship to systems and perhaps the systems relationships to underlying processes.

I’m not quite sure where the mapping of the relationship between a community and a business process should sit but its a key part of the end to end puzzle.

The vision I had was that in the future, an understanding of the linkages between a community and its supporting technology base would lead to a possibility of agile provisioning, either of servies or of capacity.

For the third section, that of provisioning and management of workloads I think the new combination of technologies caused by the acquisition of Platespin by Novell looks quite interesting.

An example - The sales community undertakes a recruitment drive to grow by 20% (across various organisational groups), an understanding of their relationship to the CRM system and the underlying technology allows for IT to proactively provision additional workloads in a timely fashion, providing additional capacity.

Dave Gomm - 03 March 2008 04:53 PM

“Atlas for Lotus Connections is a social networking application that allows users to visualize their current network of contacts and see how they can extend that network to tap into valuable resources and trusted experts across an entire organization.“

http://www-306.ibm.com/software/lotus/services/assets.html

I think this is just one example of how the social networks which have been parading around the web for the next couple of years are likely to manifest themselves in the enterprise.

For a long time vendors have been modelling the relationships between configuration items (sometimes entire systems) and their business communities using formal organisational data.

In the new age where it seems to be becoming acceepted that a business ‘edge’ may be gained by allowing the rapid formation of dynamic communities of like-minded individuals, perhaps the future of understanding and servicing the needs of our clients is within understanding these new communities and their technical dependancies ?

This is a very intriguing idea.  I happen to agree that this sort of functionality is perhaps one of the future use cases for a product like Tideway Foundation.  As the use of distributed systems continues to proliferate in firms, the complexity of the n-tier applications that run on these systems is exponetially becoming more complex.  It is virtually impossible to personally know about all of the touch points of an multi tier application.  The concept of tribal knowledge only goes so far in today’s complex IT environments.  If there were a way to assemble a map of the virtual communities in a firm, then when it comes time to troubleshoot or (in the case of Foundation and BAMMING) and application you would have a better idea of the individuals or teams that support an application.