hello my name is lauren miller i'm a solution consultant here at servicenow specializing in it operations management or itom and today in this brief video i'm going to be talking about the machine learning approach to service mapping within servicenow so as you'll see here on the next slide there are a few other approaches to service mapping as of the rome and san diego releases so we will have two other videos diving into the tag-based approach to service mapping as well as the more traditional top-down approach to service mapping but in today's session we're going to be looking at the machine learning approach to service mapping in particular so just to start out why would you use the machine learning approach to service mapping and then we'll go into really what it's doing and we'll take a look in my environment so first and foremost the machine learning approach to service mapping is really giving you the fastest time to value when you're creating those service maps and the reason for that as we'll see is the machine learning is running in the background constantly as discovery is populating your cmdb and therefore it's actually able to generate map suggestions for you so candidates as we call them for maps that you have in your environment that you need to actually go ahead and intentionally map within your servicenow environment so rather than with more traditional service mapping methods you have to go in and establish that entry point for the application to launch the mapping launch that top-down discovery in this case you're actually going to see a list of map suggestions that discovery has already identified it's seeing the trends that may suggest an application and all you have to do is click to actually map that process that it believes is in your environment and we'll talk deeply about how this works but as you'll see this works best for if you have a lot of homegrown applications that you've built natively or if you have many applications in general maybe it's just too many to actually keep track of keep track of their entry points and map more traditionally so you may be a good candidate for this approach but in general we'll take a look at boiling what this brings you for any of our servicenow customers so how the machine learning approach is service mapping actually works so basically we're using predictive intelligence to actually map those applications so as i mentioned as servicenow discovery is running across your environments running across your on-prem data centers your cloud environments it's already pulling that ci configuration item data to populate your cmdb but in addition to that data it's also pulling in many of your running process data any of the connection points between your cis that traffic data that's connection connecting those configuration items and so as those particular data points are being pulled in we have machine learning algorithms that are running over that collected running process data for example and it's clustering that data into similar process groups so what this allows us to do and again this is running in the background automatically it's clustering that data and it's creating the candidates that may be mapped so it suggests it's looking at those similar processes that it's clustered and it's suggesting candidates for service maps that may not have been created yet this is what we call application fingerprinting now on the more granular level that same logic that same predictive intelligence that's running over that running process data is also suggesting candidate suggestions for existing maps so as you can see at the bottom of my screen here at the individual map level if you have maps that are already existing and need to be expanded or maybe there's a configuration item on an existing service map that's been swapped out or changed added it's going to automatically use that predictive intelligence at the individual map level to suggest any additional connections to that map that may be relevant to the existing map you have it's going to show you that that confidence level and allow you to add that suggestion to your existing map as well now here's the actual configuration to get these items in place and then we'll do a demonstration so you can actually see what this looks like in action so first couple things if you're not seeing these options available in your instance already you'll want to go ahead and verify one your licensing with your account rep that's first and foremost make sure you have service mapping active in your environment but the first step as you can see on my screen make sure this particular machine learning connection suggestions property is set to active and your environment i believe by default is it is set to inactive so make sure that's active so you can get these things connection suggestions flowing in and then two there's a bit of an optional piece here so as i mentioned and as you saw on my last screen as we're going to see in the demonstration as these connection suggestions are coming into your existing map there is an option as of the roam release where you can either set a setting to have those connections be audit added automatically to your map based on particular rules that you set or you can go ahead and not enable this particular setting so that you have to manually review those connection suggestions that are coming in from the predictive intelligence and make that decision for yourself so there is an optional setting there for you to either have those suggestions be added automatically based on rules or if you'd rather you can go through them manually and make the decision based on the intelligence of your teams managing those services step three once we get all this data flowing in we want to regularly review our application fingerprints as i mentioned those are the actual full unmapped applications that predictive intelligence is thinking hey i think i found this based on the running process cluster that we've identified so you'll want to review that list so you can decide if it's actually a service or an application that you want to map based on your environment and then of course regularly running discovery to keep not only your maps current but your cmdb as well pretty much a standard practice here at servicenow so for our last few minutes here i'm going to go into the demonstration to show you what this looks like in my environment okay so i've switched over to my demo environment here and where we're starting out is actually in my discovery home module so once you've activated that machine learning setting that i've mentioned you've upgraded accordingly to get the machine learning approach to service mapping modules and things that we've released you should see this application fingerprints widget on your discovery home page so i'm going to click in here to view my suggestions and this is going to be for that entire map level so this is as i mentioned discovery is running and it's seeing clustered process data that may suggest an application in my environment that i may wish to go ahead and map so go ahead and click into my suggested applications list and you can see i have quite a few i have quite a large environment that's being scanned by discovery and this is why if you have many applications if you have many applications that you maybe haven't been tracking um accordingly this is a really good approach to see everything that discovery is picking up on so you're not missing anything in terms of applications that you may have running so i'm going to go ahead and search for my trucks example here so as you can see this is my list that is populating based on that clustered running process data that discovery is pulled in and we have machine learning running over all of that data so i'm going to click into one of my examples here so you can see what an application suggestion looks like in my environment so for example we can click into this trex preprocessor suggestion now in this case this is a suggestion that i have not yet told to map but as you can see a lot of data has been populated for a map that i have not even given an entry point i haven't given it any data to intentionally map besides discovery running so there's a few things to note here as a prerequisite for discovery to actually pull any application data in and actually map a process it needs three things one is a process classifier which basically says okay this running process that we found belongs to this particular application so kind of closing that gap here besides the running process classifier it also needs a cmdb class to put this data into and it also needs a pattern to actually intentionally pull in specific attributes about that particular process that particular application now for anybody who's actually gone out and created a pattern within servicenow you know it's no easy feat and so the main benefit of having these application suggestions populate automatically is you can actually see all three of those things have been created for us out of the box for each of these suggestions coming in so that classifier rule that i mentioned which is pairing okay this running process belongs to in this case this tracks application we have that already created we have the pattern name that's pre-populated out of the box for us and the cmdb class that if i go to discover this application it will populate this data into this particular class that's already been created for me so it's really simplifying the process of creating a new route to discover any applications especially for those homegrown applications so if i want to launch this particular application i review it yes i've worked with my team here this is an application we have running and wish to map i would simply go ahead and discover that application but for the sake of time here i'll actually show you an example of what this looks like when it's already been done so as you can see i've already gone ahead and mapped this particular example it's still listed in my application suggestions that discover application box has gone away because i've gone ahead and mapped this application and i can now review it as part of my application services which we'll review in just a moment so as you can see as machine learning is running over that collected running process data over those traffic connections that discovery is bringing in it's suggesting these application candidates and really reduces the time for you to go ahead review these application fingerprints and map these suggested applications as they're relevant to your environment so we've just reviewed how the application fingerprinting aspect of machine learning approach to service mapping actually works for suggesting entire maps entire applications that we want to surface map into our servicenow environment and track that way i'm now going to show you what this looks like on the individual map level for pre-existing maps or new maps that you're going to create how can that machine learning be applied to those so we'll stick with our same trex example and i've just switched over here to my mapped application services page which is showing me all of my applications that i've gone ahead and service mapped either traditionally or through machine learning applications fingerprint module that we just looked at so i'm going to click into one of my existing application services this particular map that i have in my environment here and now that that's loaded we can really see this looks like any other service map but we can see from the entry point all the way down all of these particular ci's this service map was created from that application fingerprinting module that we were just on so from one of those application application suggestions that was launched based on how discovery and based on one of those running process clusters that was identified so i'm going to go ahead and open this box here which is called my connection suggestions rather than application suggestions as we saw previously and so for each of my maps i'm going to have this box and this is basically showing me connection suggestions again from that cmdb traffic data from those application fingerprints it's running over that process data the same way and it's looking for those connections for example hey we see there's a process that's connecting two servers and you didn't have that second server on this particular service map and that's feeding that predictive intelligence trainer to then give us these connection suggestions at the individual map level so based on these suggestions we can see the actual ci that it's recommending we add to our map here we can see where it's recommending that connection live based on the process data it identified some additional details about that ci and then the confidence level based on that predictive intelligence trainer how confident it is in this connection suggestion that it identified and then of course any action that's been taken on it we at the bottom here we can see a few connections suggestions that either me or my team with the correct permissions i've already actually added to this map from this list but if there's a few additional ones that have not yet been decided on i can go ahead select that and either maybe it's not a good fit maybe it's discovering it it's not actually a relevant suggestion to this map i can exclude it from this list or i can go ahead and add it to my map now go ahead next time i run my discover on my service map it'll add that suggestion as a ci on my service map as we can see here again really reducing that time to value to keep our maps current and to map entire applications that are coming from our machine learning connection suggestions thank you for your time and please don't hesitate to reach out to myself or my team anyone at servicenow as you have any questions or concerns as you navigate your service mapping journey with servicenow