I have been lurking on the INUG – watching the exchange between product management and end users. I cannot help but think that maybe IBM needs a bit of vision from an architect that is not Developer centric.
Doesn’t a product start with a vision of “something” that is supposed to address peoples problems?
The problems that are occurring have been happening a long time. So, what is a User to do?
Firt up, your problem as a user, hasn’t gone away. Support quit responding and your problem exacerbates exponentially until you’ve painted yourself into a “lose – lose” situation. Because nothing is getting done via PMRs and IPARS, your managements frustration levels are increasing on a moment by moment basis.
What if? What if we built a generic Netcool replacement or for a better term, Event Management dashboard. This app would be done in HTML, a database, and a bit of Perl. What do you say? We can do this interactively as well. Tack on comments, requirements, or even samples of code. In the end, lets see what WE can come up with.
In the next couple of days, I’ll lay the foundation for daemons in our new EMD function. Keep in touch!
My team built a netcool omnibus replacement for my company over eight years ago with mysql, <500 lines of perl as a server daemon, and a php/javascript/ajax web front-end. We built an operations self-service admin rules formatting interface right in Remedy, a system that every employee already had access to that gave us row-based access.
Over the years, we added replication between instances and global virtual desks with some follow-the-sun functionality.
We also built a custom C++ (linux server/windows mfc client) OpenView replacement. This system has a built-in pollers and can be configured to use external pollers like NerveCenter.
What exactly are you looking to do?
I’ve been thinking about this quite a bit as of late. (A couple vendors are looking at doing this again…
Lets visit this in detail. I’m thinking the end-to-end solution consists very broadly of event management, database, and portal.
Event management handles pollers, collectors of all kinds, finite automata and is scalable, distributed and async. My thought on the system side is to use C with PERL extensions for the user base. I thought of opening this up but getting the same solutions in different scripting languages might get too muddled.
Database is similarly scalable, distributed and async. Any SQL enabled database should be made to work. By having async collection times, this gets around problems of flooding the WAN.
The portal side includes workflow enablement from various practices with swappable jargon (ITIL, BPM, etc.) The portal should leverage JAVASCRIPT over JAVA with jquery most likely due to speed but extjs and others could be used for the admin. Perhaps WordPress, Drupal and/or Joomla for the top side with event list plug ins.
Of course the dividing lines should fall along where the industry is verses following any top down academic plan. Ideally, other OPEN source tools should be able to be plugged into the other pieces
I like this idea. 😎