now there we go we're recording so this will be available on the community link I will put it in the chat so this is beautiful so if you do have questions later mr. Steven Bell will be able to answer them directly on the community for today's session let's use the chat well you could go off of Newton interrupt mr. Bell I think it's circling around on your mic there goes that's better all right welcome everyone to this session MVP meetup the topic for discussion is implementing a CMDB I'll be presenting a couple of slides and opening it up for this course I'll throw in my two cents and then I'd be interested in people you know what your feedback is what you might think concerning this and what your experience might be so let's go ahead and get started I have to throw the safe harbor up because this is my opinion and you can take it with a lump of salt okay so our agenda is basically four slides total we'll be talking about CMDB architecture then the data feeds into the CMDB and then implementing the CMDB this is me I am a five year MVP with service now I work for Accenture I have almost nine years experience with ServiceNow and about almost 38 years experience as a developer so I've been at it for a while I also have main focuses in scripting and CMDB discovery and everything around CMDB so all the ITSM item and asset management side of the world I also do quite a bit with cloud management and I'm a certified instructor yeah I teach on twelve topics now so this is slightly out of date and I'm doing a lot of a lot of different things right now so very much into the development on the CMDB world alright so first discussion see NDB architecture what should really be there a couple things on this first off where we open it up for talk is discovery is sometimes this perceived as being the CMDB it is a single feed for the CMDB so watch out for that the next one is is that ServiceNow has brought forth the see SDM and we are at currently 2.1 3.0 should be released in the next couple of months hopefully so we'll see a 3.0 rolling out on that then ITSM is a main user of the CMDB data as asset management and both of these are extremely important consumers of the CMDB and then I'd like to bring in the reconciliation engine which has been around for a little while now and is underutilized by people who have implemented the same 2bn discovery then of course discovery service mapping third-party tools and so on plus customizations to the table structure to the scripts and everything else supporting this and DB so this is a generalized vision of a typical see NDB architecture and I think that right now we can open this up I've got a couple of wait couple things I want to point out number one is the reconciliation engine the order of precedence and the D dupe functionality is a core feature now for many different feeds to go through to get to the CMDB then a little bit on the tail end of things is the CMDB health dashboard which is relatively new for monitoring things on the far side for discovery service mapping third-party tools and such I want to point out that mid server clusters are your friend especially in production this helps load balancing and failover so it's a very important and crucial component for deployment of the CMDB end automation of the sandy beach and then let's see what else anything else on this not really I was going to bring in a bunch of stuff on asset management I backed away because asset management is not used by every company and part of this has to do with automatic creation of the seem to be records so if you create an asset then you can set it up to create a CMDB record okay so based on the things that brought up so far anyone got anything they would like to feed in ask me questions on hey this is peg question sure go ahead how do you know how do you get your CMDB to the point where your IT organization considers it their single source of truth when it comes to AWS and all the other hardware that we would discover in data centers so we're running a Discovery Project right now and we know that this is just a smaller piece of the larger CMDB project but it seemed like nobody really knows when this will be finished or ready or single source so how do we know when we're there or what can we do to help push it towards single source of truth there's a couple of aspects to that thank you peg that's a very good question and it's one that I asked a lot the first one is the old 80/20 rule so when you when you're at the point where you can discover 80% of your infrastructure you basically arrived the data is automated for the most part you're bringing in as much as you possibly can without going into the detail problems there's the act of not classifieds alive not classifieds in discovery and service mapping is the particular C eyes without a pattern can happen to discovery as well where you're not gaining very much back for the amount of work you're actually pushing into trying to discover things so getting any percent of the way there is usually considered a pretty good rule of thumb if you can get beyond that you're doing great you want your accuracy to be very high so you want it to be as automated as possible you reduce the manual input and the spreadsheet uploads down to almost zero except for on things that can't be discovered like asset tags and you've gotten most of low-hanging fruit out of the way where a good portion of the networks are being discovered you're working with good credentials on everything but there may be a handful of devices out there that just are not credentialed the same way and may even have individual credentials so you get the idea there are these little ones and twos that you get back that take quite a bit of work to get them back and you've gotten all the easy stuff out of the way getting to that 90 percent or I'm sorry 80 percent rule of thumb awesome thank you sure hey Steven this is FEMA I see now I work for Pfizer and manage asset management applications in ServiceNow I do have a questions and maybe guidance from you how to go about managing asset records and creating CIS today we do create from set records we do have other sources where discovery creates CIS people creating CI so what's the best way to go about looking at your diagram it looks like that asset creates CI is discovery does not bring CI so what's the best methodology to create TI is to avoid duplications or data quality yeah this diagram was simply to show the feeds that could go into C and BB that are the common feeds I wasn't really showing the reverse of that into a semantic model category table out of the box is decent it's not comprehensive so it takes a while to build that up where when you enter in an asset it will actually build a CI on the other side in the correct table so it takes some time to put it in for both directions and get it looking right the reverse of creating assets is actually simpler because you really just have four right you got the you've got your basic asset you've got your consumable you got your software and your hardware so the reverse direction is actually pretty straightforward but when you go over to the other direction from asset management to the CMDB that takes some serious thought and work as to how that's going to happen and that all works through models and model category now but with that understood as far as the you know best practice is surrounding it again it's always I recommend implementing seem to be for asset management and hardware asset management before you even touch software asset management I recommend having discovery in place before you do anything with asset management I think service mapping probably can be fit in there and implemented somewhat at the same time but you really need your automation of the CMDB to be really nailed down good and that way you get your existing technology your existing CMDB Lisa could we have everyone muted please I'm getting some feedback from some people thank you that way you get the majority of your known topology into the asset management table so discovery is really important or some sort of discovered tool service mapping is really important to get there and relationships down so that that's understood and then you build out your model category so that assets are being created and then you can look at new things so how do you want to do this do you want to go from procurement to asset management to creation of the CI automatically from asset which by the way a lot of people do but you can also end up with the service team taking it and putting it into the seem to be for the first time and having a backward create the asset I think this is upside down going in that direction but we also run into a lot of companies that do it that way I prefer the procurement side anyway I hope that helped it's you know I get I get this question a lot there's a lot of variables to it it did help absolutely and some of them we do today with automation was just from outside perspective what is the best way to do it versus how we do it but thank you very much manual entries got a problem right now in Orlando and may have had a problem in New York I've just checked it out recently and it failed for me so not sure what's going on with that I'll just be blunt okay does not utilize see our reconciliation yeah and it should yeah there's a problem there it has an entry and reconciliation for manual entry but it doesn't work so I'm not sure what's going on with it right now I haven't had a chance to turn in a high on it but I bumped into it for the first time in front of a client a couple days ago and it was very embarrassing watch this the leads me to suspect that the uploads may have a problem as well and I gotta go check it out let's see when I saw a man Yantra is one of the out-of-the-box choices reading your finger when I saw the many Huntress one of the hollaback choices for Discovery source uh navel II assume they are even handle the precedence for manual input versus integration yeah it doesn't oh you did raise the high okay yeah I was going to slam one in yesterday and it didn't happen yet yeah sure no it's busted and I mean it's busted so don't use man you know don't expect manual entry to work folks when you're doing the reconciliation it should but it does not another question from the chat area what advice you could give that company that is doing CMDB all manually ah got an hour that's a nasty one yes there are two aspects to this number one I would normally say implement a record for the data upload inside the reconciliation and then put the manual record in there as well but since manual isn't working for reconciliation it doesn't mean anything right now but you might do it anyway just to have a forward look the next thing is is set your processes in such a way that you're reducing the amount of manual entry on an individual basis and goto uploads as much as possible another thing you might do is then implement some sort of data certification there's a data certification plug-in out there highly recommended it allows you to double check the actual ownerships and what it does is it allows you to package things up so that the existing ownerships of c eyes are then sent as a checklist to the owner and the owner has to validate them you'll have to have someone actually working those tasks to make sure that you know they're checked against the existing scene DB and that way you have a couple things one is you've got basically a type of inventory going on where you know that the person actually owns the thing and then that it actually exists still so if they come back and saying no we've never heard of this then you'll have to do some you know legwork on it to determine who's got it and so on but that's a double check so that you're just not uploading blindly you're also turning around in verifying beyond that there's not a lot I highly recommend going to automation as soon as possible getting people out of the middle of the updates is the best thing in the world I have had people upload things like users to the CMDB I knew we were types of consumables but I had 300 users and it seemed to be I've seen cans of oil in the CMDB I've seen crazy stuff in the scene to be getting people out of that and getting into an automation of the actual network with software and hardware is the next maturity step to every organization and you want that you want that badly so you can deal with it for a while on small stuff you know and have like just servers in your CMDB that's usually the first place to start when you're firing up a scene to be and then introducing people's devices you know like laptops desktops and so on and and then trying to keep track of middle amount of hardware and our software and then the final thing is business services business applications application services and so on getting those put into the same two being you're going to do that anyway since they're non discoverable and manual or automatic you're gonna put that stuff in but yeah there you go make sure your relationships are in place there are other relationships make sure that you have some sort of business service mapping going on you can do that manually so it's it makes a lot of sense that will go a long way towards getting you to a even though it's manually entered it gets you to usable seemed to be that your I kiss them group can use this the services crowd really is your number one consumer in this regard with a small seem to be there really isn't many others that will rely on it because there's an enough data there but anyway hope that answered your question it's a it's one I get a lot to you know it's like what do we do you know we've got all manual ok let me move to the next slide here and it may actually help so we've talked a little bit about the architecture and and I've given you some best practices around some of that but let's expand on the data feed thing because as you heard me mention there's the uploads and these are the manual uploads and the manual entry side and then I've also talked to the reconciliation engine and the order of precedence you can set up for golden sources this is huge I can't push this enough I used to write these kind of things before servers now came up with their own and integrated it with the environment and what I did is I set up a very similar voting record kind of thing where I had the highest order of precedence was my golden record and then the next one down the number two would be you know some other automation finally the last two obviously being your uploads in your manual entry service now came along in and built this so the precedence allows you to order the golden source and let's say for sec M for example SCCM is an agent-based system it can do provisioning but it also does a decent job of discovery especially Microsoft software and does a great job of bringing back additional information it's a little harder to get through discovery so it makes a good golden source for certain software tables inside the CMDB and therefore if you do an SCC em integration you can bring that data in and when it's written down in the seem to be in those specific tables you can set it up so no one else can write over anything it's filled in and that's a little idea between the precedence and Reconciliation stuff normalization usually you want to take things hill Packard's the worst I've run into take things from various manufacturers and you know if they've got 47 different naming conventions like EULA Packard cue pack hue-hue P Inc HP Inc HP h dot P dot you get the idea there's like whoever was in charge of the factory during the day anyway the problem with that is then it automatically creates a record for that particular manufacturer in the manufacturer table so all of a sudden you've got you know fifteen different HP's sitting there normalization allows you to take all 15 Creek call them aliases and turn them into just HP dot and that it becomes the norm for your entire CMDB you can also do that with the model table which is important too I've set this sort of thing up and it took me a couple of days to manually enter everything because what you get into it kind of cadence for putting in all of these and you've done the investigation know what they are you can get it pretty much cleaned up no time flat but there are also other normalization tools out there third party type things CI identification there's a whole raft of stuff around that and we teach this in service mapping pretty heavily because there's there problems there and also in Discovery classes we teach this because of the potential for issues and so there's there's a lot of information behind that and configuration to get it to working optimally and keeping down duplication but it serves now also has duplicate detectors so they go out they try to identify the CI if it looks like it's going to be a duplicate they'll mark it as a dupe it's just different enough that it doesn't you know that it doesn't update the existing CI but instead would create a second CI so serfs not handles that and then generates a task alright so now we got that all sitting out there anyone have any questions on this sir stuff she permanently mute everyone yeah okay I just you know wanted to make sure that we could actually on the it for you it's okay we just had people talking over or you know background noise and stuff like that so go ahead yeah no I we found that with the reconciliation engine that the that that's probably the most difficult part of doing any import is remembering to actually call that I almost wish service now it would set that up as a any import into the CMDB would automatically go through the IR e versus having to remember to put that it our your code true yeah that is that is one thing to watch out for them when you do this the manual uploads and you know needs to be carefully done I think your acid it's been a while since I looked did they ever set up the SCCM integration to automatically go through the IRA or that still we have to manually configure that as well I think you believe I believe you still have to manually configure it okay that's what I thought but I thought I'd ask we we hadn't finished our upgrade or when you know yet oh yeah certain things like Steve McCarty just put in serfs maverick conciliation or per the base platform their additional well reconciliation is part of the base platform so and what I wanted to mention on this Philip was that service mapping actually comes with some of the stuff pre-installed and they put records into the reconciliation table for you so that stuff you know there are some plugins that are now coming with reconciliation records that are part of the plug-in installation side so that's pretty cool but Steve to answer your question the service mapping is a pay for service and usually if you get the item as the item suite you get discovery and service mapping together but the other paid services anybody else on this about four minutes for the for those that haven't gone through the CMDB micro certification yet I learned something new while going through that yes over the weekend that you were just talking about it what was it I was going to say that the mat it was something with mappings Wow I have an I had it in my head when I it's gone I'll go look it up and then come back okay yeah I was going actually I've got a slide here that is recommended links and one of those is that seem to be micro cert so and it was informative I will tell you that oh yeah I highly recommend it that's why I've got it on the additional links here I've got a couple things extra there CS DM link and a couple other things to point you to going out and looking at what the latest and greatest is for process and also for supporting but that microsurgeon excellent it's a hands-on self-paced course and it's free and it's 20 hours and it's worth going through customs here let's see Patrick we're on Madrid is there a way to automate the deduplication process yes but I believe it's going to be custom for you I set some stuff up like that from one of my customers and then also we had to go through a like ten step recovery process because they had a significant number of duplicate records in their relationship table tens of millions and then also and it required some serious cleanup because the C eyes were also in the tens of millions and the ITSM side of things had connected into a lot of the duplicates and because Madrid doesn't do things on a magic in there it was it was real grief type thing so we had actually go out and mark them as duplicate ourselves and then get rid of all of the actual duplicates there is a trick to it we basically looked at it through fixed grips you know went through a fixed grip type thing and then went through and updated the records I'll Terry I'll be Terry asked where do we find the info on the scene to be micro cert I'm giving out the link at the end of this session on that community link as well where this recording will be yeah if you want you can post this deck anyway the takeaway since we're down to one minute here take care of architecting your seem to be solution build a diagram like you saw just a couple slides ago sit down and take the time in Visio or draw or whatever flavor and you know put it on paper and visualize it understand what you're gonna be doing and how it's going to be done identify your current and future data feeds because you're gonna need to know how to set everything up utilize that reconciliation engine it is best tool also get automation in their discovery service mapping any other third-party product that can help you out and get the human out of it is a best practice and then I've got these as the links configuration management fundamentals is that micro cert CS DM 2.1 is a really great place to start go out there it's got all the process of procedure for CMDB and it's it's got it all laid out including roles responsibilities and everything so very nice document also goes into some of the discovery and service mapping side of things that's it thank you very much for attending this session thank you very much thank you all have a great knowledge