Join AITSM Co-Op, A community for those who ♥️ AITSM (ITSM, ITIL, AI and Automation for ITSM)

AITSM - What No One Is Talking About (Part-2) - An Elephant In The Room

May 31, 2022

AITSM - What No One Is Talking About (Part-2) - An Elephant In The Room

AITSM - What No One Is Talking About (Part-2) 

An Elephant In The Room

How to measure health, and achieve faster remediation via AI & Automation effectively?

Special Interview With Neethi Upadhya, Associate Director, ServiceNow Practice at IBM

AITSM - What No One Is Talking About (Part-2) 

An Elephant In The Room

How to measure health, and achieve faster remediation via AI & Automation effectively?

Special Interview With Neethi Upadhya, Associate Director, ServiceNow Practice at IBM

In the last episode, we discussed the elephant in the room. Were you able to guess it right? 

Hold your breath. It is CMDB. Configuration Management Database.

We discussed more exciting aspects of CMDB in this episode with Neethi Upadhya, Associate Director, ServiceNow practice at IBM.

How can a powerful CMDB help speed up the detection and remediation process or prevent the problem? What are the three metrics to assess or measure the health of CMDB? What is an effective way to use automation & AI with CMDB?

We'd learn a lot from her plenty of examples, conviction, and point-by-point discussion from 18 years of rich experience. 

Don't miss the end, where she discusses an essential aspect of CMDB because of the changing landscape, templatized environments, and fail-over mechanism, to be explored in further episodes.

"If we can predict what is going to go down or what could go down and somebody acts on it even before the incident occurs, that gives you the nirvana state. Because then, your end-users are not affected. I think organizations are getting there, but that takes a lot of discipline into understanding, changing the way of working, changing your KPIs, and ensuring you give more importance to the CMDB."

"Having stale data or stale CMDB, it's worse than having no CMDB. Sometimes you can work without data. It is better than working with incorrect data. So, from the CMDB perspective, being dynamic and up-to-date becomes very critical."

Learn and Grow, With AITSM Show!!!


Anand Tambey: Welcome back to AITSM show. In the last episode we discussed about an elephant in the room. Were you able to guess it right? It is CMDB. We will explore more insights about it in this episode with Neethi Upadhya, Associate Director, Service Now practice at IBM.

So from the service desk point of view whenever there are certain incidents are reported and then tracing back to the CI, like how this process happened. So what are the critical factors here where CMDB could actually accelerate the process of resolution or kind of preventing the problem?

Neethi Upadhya: A good question I would suggest. See, a lot of times what we've seen from service desk perspective comes to a point of having somebody report a particular incident when it comes through the help desk or the service desk channel. Right. So what happens is I, as a user when I create an incident on the system, I have to tag the right configuration item or I have to tag the right CI or application as such.

There are two aspects that I would like to highlight there, is one being, see, what is the earlier ways of working when people go and tag a CI. A lot of times what used to happen, Anand, is when we go in for such incidents, we would see that the user's asset gets tagged because I, as a user, when I'm creating an incident, sometimes my laptop gets tagged, but the issue could be completely different.

It could be an application, which I'm not able to access. And it's not that my laptop has any issues, but then issues with the application not being up and running, because there could be a network mode or something, which is not accessible at that time. Right. Lot of time was getting wasted in actually tracing as to which one is the right part of the puzzle that you need to look at to fix the issue.

But then now people are going service-based, even if I am creating an incident from my laptop, I will still select the service, which is not available for me. So then it becomes easier for a service-desk guy reporting. And then in the backend, if you look at how CMDB is getting structured and how they should be structured, it's a service which is connected to the application, to the technical service to the infrastructure. And there are a lot of our newer service models, service mapping as you call it, right. That if you have the right service models built-in, right service mappings built-in very clearly, it can indicate as, because there's not just one incident that's coming to you at that point of time. Right? So there could be one incident coming from a user saying I'm not able to access this application. And if you have your monitoring tools, for the CIs, you have your monitoring alert also come and saying, okay, this network node is not accessible. And when that gets into the service model and you start looking at what are the points of failures then you can detect as to which is the core which is causing on this issue.

And that's faster to detect, faster to remediate also. CMDB with relationships becomes very, very critical. That again goes back to the fact that having static CMDB is no longer helpful. You have to have CMDBs with dynamic relationships, so that you know exactly what is connected to what and how is it getting impacted in, in times of failures and so?

Anand Tambey: Right. Insightful ideas here , like we don't have to actually worry about incidents and problem, if the CMDB, the infrastructure and all this foundation is built up on the correct grounds, correct relationships and with impact, whatever, like where it impacts most and the service mapping.

Neethi Upadhya: Absolutely. I think that the way I mentioned earlier, there has to be a paradigm shift from failures to actually ensuring that we have value and customer satisfaction is not impacted, which means the organizations have to start thinking about how these incidents can be caught before a user realizing that an application is down..

That's where AI operations, AI ops concept comes into very big use because the predictive part of it, right? Predictive part of AI into the service management aspects of it. It helps you because at that point of time, you can predict an incident before it actually happens for the user to know.

So if your systems, because you have, you have multiple tools then integration between a couple of tools and reading of the logs if we can predict as to what is going to go down or what could go down and somebody acts on it even before the incident occurs, I think that gives you the nirwana state, right?

Because at that point in time, your end-users are not impacted. I think organizations are getting there, but then that takes a lot of, lot of discipline into understanding, changing the way of working, changing your KPIs and ensuring you give more of importance on the CMDB. Because for AI ops to work, CMDB will be the very, very critical piece of puzzle there.

Anand Tambey: And it would to be interesting to know, like, What kind of metrices you track, to measure the health of a CMDB? And on what parameters I would measure it, like on what metrics, I would say this is working perfectly and maybe in some areas it would need some kind of improvements.

Neethi Upadhya: Yes. One is the completeness. So your CMDB has to be complete in the sense, as I said, once you understand which are your classes, that you have to manage and govern. You have to understand what are the attributes within the classes, to be able to say that my CMDB is complete, the data is complete in that. So completeness of a CMDB is one of the metrics that you have to measure.

The second is compliance. You will have certain norms on conditions that have to be met for the CMDB to be called compliant. So it could, it could be a certain license aspects of it. It could be certain security aspects of it. It could be certain data aspects of it that have to be there for a CMDB to be called compliant.

The third is, we'll have to look at the stale factor. Your CMDB should not be stale. So basically it should say that how frequently it's getting updated, how dynamic is your CMDB. So being stale is one of the factors that you have to measure and it should not be very stale. The data should be up to date. It should be dynamic. All of that. So these are a couple of metrics, which you have to have and the relationships. So the stale also comes from the relationship. How frequently is the relationship data between the CIs managed and maintained?

Having a stale data or stale CMDB, it's worse than having no CMDB. Sometimes you can work without data. It is better than working with wrong data. So this from CMDB perspective, being dynamic, being up-to-date becomes very, very, very critical.

Anand Tambey: Correct? A lot of insights are coming up with this discussion. So what is your view on the importance of CMDB and the automation, say AI-Ops or conversational chatbots and AI, how it is actually changing the landscape of CMDB and helping us to move faster and better in ITSM.

Neethi Upadhya: So from the automation perspective, If you look at it holistically, right? Automation can also be a part of the input into CMDB.

 The best way to manage and maintain and govern CMDB is to put automations around it where you have discovery kind of solutions from various vendors. Service Now has a discovery . We have the discovery for incidents, we're bringing CIs through discovery, an automated way of actually populating and managing your CMDB. That's one part of it.

The other part also, which is where the output of the CMDB will help you automate activities in the landscape. So if for example, you have, as we said, let's just continuing the previous example of failure that happens in your landscape. A particular server is down because of certain things.

Then you want to restart also lot of time having that details within your CMDB helps to trigger those remediation processes also, which is in itself a large conversation as suggest to how a CMDB scan help automation. But then if CMDB is done right, all our automation tools integrated into your system of workflows, integrated with all your automation tools can trigger remediations also using the CI data that is in the CMDB as such. So with that, automation is helping through CMDB.

Also, the way systems are changing, when your entire platform and entire stack is templatized and having that information also within CMDB, as such helps you not just manage, simulate, or create that stack in the environment as such, but also helps you create failovers whenever something goes down.

So that's a larger conversation, but then having the CMDB right can help you with a lot of automations, can help you with automated failovers, can help you with getting remediations sooner, faster, and get the right people identified to fix the right kind of an issue..

So I think with all of these aspects around also managing your security, vulnerability aspects of the CIs. So all of that being very critical to a business at this point of time, the importance that we give to CMDB must be increased in an organization

Anand Tambey: Right, right. So that is really interesting, like getting into that mode of templatization and then failovers.

So that discussion is like a kind of a larger discussion. So probably we would like to invite you for a few more episodes to discuss that. Right. So that would be another thing which people are not really talking much about and then taking much care about.

Neethi Upadhya: Absolutely, Anand! I would love to come in and have a discussion on those aspects. Because. they're very closer to the transformations that we work, day in and day out. And I think this is one of that aspects where we have to start looking more into and propagating the same, in the community as such, would love to collaborate.

Anand Tambey: Thanks. Thanks a lot, Neethi for joining our AITSM show. We are proud to have you. And then with your rich experience, we have learned so many things and so many insightful ideas are coming up. We'll also do more and more episodes and collaborations in the future. So a really, really great discussion we had today. Thanks a lot for joining us. Thank you.

Neethi Upadhya: Thanks for having me. Have a nice day.

Anand Tambey: Have a nice day. You're welcome.

Thanks a lot for listening. Please let us know what topics we could add here. And give us feedback. And if you have loved it. Please share it generously.

Thanks again. Bye-bye.

Neethi Upadhya Profile Photo

Neethi Upadhya

Associate Director- ServiceNow Practice at IBM

• Overall 18+ years of experience in IT industry in architecting, design, development, and delivery of complex information systems.
• Executed more than 20 projects of TCV greater than 50M$ in Service Management and automation for various clients across the globe, spanning across Telecommunication, Energy, Retail, Pharma, and Finance industries, bringing creative solutions to everyday service delivery challenges
• Managing a P&L of over 25M$
• Directed the design, development, and launch of highly profitable service offerings