Transcript of Basic ERP and MES Integration

all right integrating Erp and mes a high level overview take the zero [Applause] [Music] all right so it's been a lot of discussions lately about Erp integration I would say if there is a if there is one topic everyone wants to talk about um wants to understand better and you are leading or participating in a digital transformation initiative with an organization that one thing that everyone wants to understand better is how do you integrate Erp into um your digital infrastructure now there are many opinions out there okay so what I'm going to talk about is how we've been doing it since 2013 give or take um you know how we architect these Integrations how we unify Erp Master data with manufacturing execution information with raw events okay how we the approach that we take to doing that um if you go I have a call in an hour with um you know the the product owner of sap manufacturing who is and we're going to have a conversation about um you know is sap digital infrastructure for example um my opinion is it's not actually it's pretty easy to prove that it's not but sap is a suite of functions if you go to sap what sap is going to say is the best way to integrate Erp with MES is to basically use sap for as much as you possibly can so um you know use sap uh btp with PCO and um Leonardo and S4 Hana use all those things together um to basically you know build the pipes the problem is is that a solution stack is not a digital infrastructure okay and this is I mean Rockwell has run into this every basically everyone who's tried to do this right Rockwell Schneider Siemens who's tried to beat the digital infrastructure they run into problems and the problem is they don't have a solution for every manufacturing issue in an organization so invariably there you end up putting smart things in your business that aren't owned by the solution provider who has basically said use our stack as your digital infrastructure so there's that argument there's the argument that hey Buy digital infrastructure which you can't you can't do digital infrastructure is something you create when you connect together all the smart stuff in your business right the other approach is you create a agnostic infrastructure based on technology and in our case we use the unified namespace architecture um we use open Edge driven report by exception and lightweight protocols to pipe everything together we do get use gateways to convert Legacy data models and information models into a unified structure for the business we use functional namespaces informational namespaces and definitional namespaces to kind of to structure a unified namespace and put all the data together in a way that it is contextualized semantically hierarchical that is I know how to get to the data point that I'm looking for and it's all normalized that is uh I can then plug that infrastructure that digital infrastructure into a machine learning ecosystem and we can look for patterns in that data because every data point has a relationship to every other data point okay that's called normalization all right so what I'm going to do is I'm going to talk about how it is we actually do this in in advanced mes boot camp which is taking place so this is session two tomorrow um May 25th if you're watching this on May 24th what we're going to be doing in the lesson tomorrow is we're going to be integrating an existing manufacturing execution system with a odoo Erp system now why are we using odoo well we're using odoo for a series of reasons number one odoo is the most open Erp system and it has all of the Erp functions in it so um things like production orders bill of materials um manufacturing steps etc etc all right so what we're gonna what I'm going to do is I'm just going to quickly take you through what does this process look like from a very high level and talk about what is the relationship between Erp data and Manufacturing execution data which is probably the most common request but the first thing I want to do is I want to read a comment from one of our Architects I was he and I were messaging this morning on teams he's going to be on this sap call that we're doing here in an hour and I want to read you know what he said this is one of the Architects we you know works on all the biggest implementations in the world and he you know he came you know he did it for 10 years before he came to us so in and this is my experience what he's summarized here is exactly my experience over the course of my career so he said all of our active projects have some form of integration with an Erp so we have some 230 clients we generally are working with between 8 and 10 actively at any given time all of our active projects have some form of integration with an Erp not just sap in all cases and even historically in every project I've ever been a part of in his career Erp integration was custom built and limited specifically for only the data for the Erp so that is the integration to the Erp system actually was limited and constrained and the reason why there's a whole series of reasons but the answer is we only care about the Erp data that operations cares about when you when you're unifying in digital infrastructure to start so inventory levels order status production and consumption of finished goods in any manufacturing facility the needs of the operator the line lead the supervisor the quality Tech Etc go Way Beyond this but they leverage Master data from the Erp and so what and that is my experience as well if you look at successful digital transformation initiatives they start on the plant floor and they work their way up they are Strat this digital strategy comes from the boardroom the enablement comes from the boardroom but the implementation the he approach to digital starts on the plan floor and works her way up and a really good example is um if you look at Tesla for example Tesla built their own Erp system from the ground up um you know sap likes to brag that they are in um 99 of the 100 largest companies in the world the one company in that 99 or the 100 that they're not in is Tesla okay um and there's a reason for that there's also a reason why Tesla happens to be the most digitally mature organization in the world I'm not just pick I'm not picking on sap here the reason we've been talking about sap more and more is because there's been a request from the community specifically from our integrator Partners saying Hey listen you know representatives from sap are telling manufacturers that unified namespace doesn't scale which is literally Preposterous if you look at the 10 most digitally mature organizations in our data set which are 1381 companies uh all 10 of them well nine nine of ten have pure unified namespace and 10 of 10 have some form of unified namespace the reality is is that if you take the traditional approach of integration to Erp then what you will do is focus on transactions over events um and any events that you do handle you will have to find you'll have to Define you'll have to create sequential function charts put in business logic to handle all those events because in inherently Erp is transaction based it is not event based no Erp system is event based no all Erp systems are looking first and foremost before they ever insert a record um you know say reference data or historical data transactional records they're looking to coalesce things into a transaction they're looking to coalesce events into transactions because they have multiple columns there's think of it in a in database terms an event is the value inside of one column so it is literally the it's the value in a Cell one row in one column that in it in a nutshell in in general that is a an event what most a transaction is really when you take all a series of events and you coalesce them into one transaction then I'll have one time stamp when that record was inserted but the in the values for each of the cells may have changed at very at very different times and what you're seeing is the timestamp associated with that associated with that transaction is the is the the point at which it was inserted into the Erp system so this is why when you ever hear people talking about Erp they're they talk about Erp in terms of the context of near real time okay all right so real quick let's go let's kind of go through what these Integrations um basically look like so let's start with you know what do we want to achieve all right I'm going to try and do this here so um all right so if what I want to do is I want to build something like this okay this this dashboard here contains data from Erp in the upper left hand corner here contains data from the manufacturing execution system down here this is this is mes data there is a lot of information here that is from our Erp system and how do we put those together in one visualization that informs the business okay so question number one is how do we do that well the answer is we could do it a whole series of ways number one we could treat that window right there that informational window as a consumer of a result set that comes back from say an API an application programming interface the way that you build most software that actually isn't what we do here what we do is we create an intermediary layer which is a namespace that contains all the information that's on that window and that information and that namespace think of it as a folder that contains all the parameters that namespace is a coalescence of data from Erp and mes so in order for us to do that we have to have a namespace that contains all the Erp data and we have to have a namespace that has all the mes stuff and then we coalesce it into an informative namespace the reason this is important is because what for in terms of raw data we are not using apis to move raw data from one consumer to another what we are doing is we are consuming all raw data in the form of namespaces into one infrastructure and then we move the whole infrastructure around the business so if what I wanted to do was say initially all I want to do is have a pretty little dashboard that looks like this I can start with say one I I start with a common infrastructure digital infrastructure building a unified namespace and let's say it's 1 100th of what will it will ultimately be a unified namespace is living its breathing it expands it's always the current state it's the structure in the events of the business at any given time as as uh as functions are unplugged we don't use this function anymore that function is removed from the namespace okay um what we do is we say the first thing we're going to do is we're going to go ahead and collect all the the raw data in individual events individual namespaces generally grouped by the devices so where did that raw digital event come from oh it came from this manufacturing execution system okay we have a node call manufacturing execution system and it contains us all the raw stuff and then what we do is we point to that raw stuff as we coalesce in some place else in the other in in the namespace where we may create some functional namespace that's going to use a Raw event from the mes system plus a Raw event from someplace else instead of creating individual business logic which is the way that most you know sap does this sap uses their infrastructure to create custom triggers and logic to move events from one place to another well that's deterministic I have to know where that event is going in order for me to build that thing well no what you do is you say what is my function and that I want to build say oee that's using an event coming from a raw namespace and you just reference the Raw event right instead of creating a pipe in between that has business logic that moves it what you do is from your functional namespace you reference the event you want to use and then you do all of your logic within the function Okay the reason why is because now the function itself becomes part of the digital infrastructure if you were to go look at a standard Erp system say you go to you know look inside a btp at sap and what you wanted to look at was the actual logic I want to understand the sequence that this Raw event goes through in order to get to whatever manual you know manufacturing flow output that I'm looking for you can't see that that's inside of the monolith right what we say is build open architectures that allow you to even see the event so in in the case of this window right here because there's a functional namespace sitting underneath everything that's on that window is literally coming from a functional namespace and we're going to take the entire unified namespace and we're going to transmit it into say a data lake or we're going to put it you know where it's going to be stored in a historian then we have the ability to view what it is someone was actually looking at that is what was on the screen at any given time not just the Raw event we can actually go well what were they looking at what window was open what what were the values on the window at the time the event happened so you get to create that now be gets to become a variable that you can use in predicting failure or and or mitigation okay all right so number one if you want to build this kind of thing and you want to do it in a way where this kind of thing is part of digital infrastructure and it's not just a consumer then unified namespace is awfully important all right number two so this obviously is what we're we're trying to build here right so this is the visual representation where we've got you know basically namespaces in our nodes we unify those namespace in a semantic hierarchy Isa 95 part two so this would generally be Site Area line cell here and then we would publish Site Area line cell starting with the site node up into Enterprise and so then you would have Enterprise Site Area line cell and then you would have Enterprise Site Area line cell and then inside of your line you would have an Erp namespace an mes namespace that comes from those raw name spaces okay all right and then uh so this is you know a sketch if you're again we're going to go over this in the advanced settings boot camp tomorrow when we're integrating Erp you know odoo with MES but essentially this is a high level sketch of how namespaces are constructed if you're using Isa 95 part two but this semantic hierarchy is the important piece because if I'm say I'm from a different site but what I want to do is I want to analyze data at this production line at site a with this product this similar production line at site b i I have to have a mechanism to add an ad hoc way be able to get to the data I'm looking for okay and not have the domain expertise to understand where the raw events are for that production line that's in a site that I have nothing to do with and then the last thing is we want to do uh just take a look at the the debt the slide that kind of started this whole thing started the whole conversation about Erp and that is the uh our Trend so I I publish this I want to talk about this piece here so one of the things that I was trying to illustrate um in the LinkedIn post that actually started this entire conversation although we talk about Erp all the time was if you look at the trends in the industry you know the so in terms of you know Google analytics where is all the movement okay um the movement is in digital transformation unified namespace open architecture basically the architectures we're teaching and the declines are in uh monolithic platform and Industrial Automation and what my argument is is that is that as we're moving into the fifth Industrial Revolution companies that are safe from this this is the industry 4.0 distribution the 1381 companies the company you'll notice that what's happening is the companies that are you know just under the mean of the distribution and above are pulling away that's where we see these gaps here they're pulling away from the companies that are you know their digital maturity is basically on the mean or just under the mean and down to the the least digitally mature organizations you know this is so this is something like um 700 companies in this group right here okay and up here that these three bubbles here that they account for only 10 companies okay so um and all 10 of these companies here oh well nine of the ten are using unified namespace okay um and part of my argument was if you were if you're working in this space here in the automation space then you know your role is going to change pretty drastically because what's happening is the companies who haven't already automated are in this group here and they are falling off but essentially what's happening is companies who are very digitally mature they are iterating with new Solutions and iterating means I'm solving a new problem um in their digital infrastructure so I'm building new dashboards I'm you know building new business functions whatever it is I'm doing and they're doing that ad hoc and they're doing it in a way where they're enabling CI CD on the plant floor so that is the best ideas in the organization are coming from operators technicians electricians mechanics Quality Personnel it's the people who do the actual work the best Solutions are not being built by some you know engineering team that was hired you know by Deloitte or or you know one of the big engineering firms you don't throw them in an office and they just basically solve all the business problems on some huge monolithic platform no that isn't what happens what happens is they spend a lot of money they build a lot of things that the business doesn't use and the reason why it's the same reason that manufacturers who offshored most of their manufacturing operations their their products fell further and further behind the market and that was because the people who were solving the problems weren't the closest to the product and it's the same thing here the the people who actually know what the problems are in the business aren't close enough to the digital infrastructure to solve the problem digital transformation is about enabling the end user that is the person on the plant floor to solve the businesses problems you don't digital transformation doesn't happen from the top down it I mean that that's a Surefire way of you might be able to capture value but you may be able to spend 20 million dollars and yield say a 35 million dollar return so 15 million on a 20 million dollar spend but what you see is those companies who make that huge capital investment up front in these huge monolithic infrastructures right and they you know they those organizations see diminishing return over time at equal spend so the amount that they're spending the total cost of owner ownership the amount that they're spending every month isn't you know one it isn't a it isn't exactly what I spent last month but let's say it's 0.8 it's 0.7 digitally mature organizations each each month each year they're reducing the spend on their solutions by you know say they're doing 0.4 which is our goal our goal is with each subsequent phase in digital transformation you're spending four tenths as much in the next phase as you did this time so if you're an organization who's doing a digital transformation initiative and every phase is costing you the same amount of money or maybe just a little less or God forbid it's costing you more you have the wrong strategy the wrong technology and the wrong Partners okay that's just the simple fact of it okay all right so let's go all right so what I'm going to do is I'm just going to kind of show you how we coalesce together um data and information all right so what I'm going to do is go here so what I have is a I have a mqtt broker that's running in the cloud this is actually our um this is our community broker right so those of you who are transmitting your um projects like you have people from mentorship translating their projects up in our broker this is our common open broker but what I'm going to do is I am going to um I'm gonna just run a little script in our Erp system and you will see what we've done is we've added in this namespace that is mes and Erp data all coalesced together okay so Enterprise Site Area line and then what we have are three namespaces in our UNS we have one called Edge which is our raw events from the edge so in feed out feed waste and state they're all zeros we're not streaming anything and then we have our process data so we have our spindle speed feed rate tool wear coolant temperature vibration Etc okay we have our manufacturing execution namespace where we have our key namespaces of production Quality Maintenance and kpis these are all coming from the mes system in most cases these this is native so the mes system if it's say built-in ignition or it's built in Factory Studio ignition was the one I showed you earlier you know all you're doing is setting up an mqtt transmitter you're literally grabbing the mes namespace for whatever production line This is and you're literally transmitting it to a an infrastructure okay so we have a production name space that contains a series of parameters uh oee performance efficiency etc etc we have quality we have maintenance and we have our kpis okay including custom kpis this is a really important Point here one of the Hallmarks of a successful did successfully digitally mature organization is the ability for CI CD from The Edge that is and so imagine everything here from Enterprise Site Area line everything except for custom one and custom two these kpis were red and what that means is red is Enterprise so we have one common infrastructure if I were to look at our semantic hierarchy I would know which of the things are deployed from the Enterprise and which of the things are custom features built on the edge we call those blue so say uh an engineer um you know he has access to the infrastructure he's got access to connectivity and equipment and what he wants to do is create a new key performance indicator to test so he doesn't want to go through the process of you know going through the um data enablement team to have them build a custom kpi that he may or may not use what he has to have the ability to do is go to the infrastructure grab the data points from say production grab the data points from The Edge and if he wants to create a new custom kpi he or she can and that in that custom kpi is blue that is anybody who looks at our infrastructure will know by some indicator that that custom one is not an Enterprise kpi it is a kpi that is within the kpi namespace for this production line but it's ad hoc if you want to know the number one and the reason why that's important is that if I have a digital transformation team who's in charge of all the Enterprise Solutions and they own the infrastructure so they create the rules the minimum technical requirements for all the people working within the infrastructure okay so the person who built this kpi has been taught instructed knows the digital strategy knows the infrastructure and knows that if they're going to create a custom-made kpi that they're going to and they're creating it for that production line they're going to follow these rules they're going to go to the line they're going to go to the kpis namespace and they're going to create a kpi they're going to either call it something or give it a parameter say the payload is a Json and one of the key value pairs inside that Json is this is a custom kpi not Enterprise whatever it is however you define it in your MTR you have to give people on the edge the ability to solve their own problems using your digital infrastructure if you look the most advanced digitally mature organizations they all enable that and the reason why is because they want they're looking for a competitive advantage and that is they want to be able to solve problems at scale they that you can't scale solutions from the top down that's just not possible okay um even even in Greenfield you can do it initially but as a Greenfield site becomes more and more Brownfield it becomes less it becomes less and less possible to go top down in solving your digital Solutions you have to enable people on the plant floor to solve their own this is also the way that you conquer the problem of recruiting and retaining the employee of the future why is it you you can't you know Legacy organizations can't um recruit and retain gen Z or Millennials it's because that's an entire generation of people who were who grew up solving their own problems they've been they've had access to all of human knowledge through a single pane of glass since they could walk and then all of a sudden they go into a legacy manufacturing organization and the CTO is the CT no they don't have access to the Wi-Fi they don't have the tools to solve their own problems they say yeah I'm out of here you know look at your you know what your organizations don't have a problem recruiting and retaining gen z a millennial the most digitally mature if you look at the 10 companies that are the in that 1381 it's not even that if you look at two standard deviations above the mean it literally there's a direct correlation between their Recruitment and Retention of gen Z and Millennials based on where they are in their digital maturity score the more digitally mature you are the higher your Recruitment and Retention rate is it's just and that's a literally a linear relationship okay all right so tomorrow this is going to look very familiar those of you who are in advanced mes boot camp at iot.university tomorrow we're actually going to do this using odoo and our existing mes system and I'm going to go through the mechanics of putting this all together now there's a lot going on there there's there's you know this is just barely scratching the surface of you know what a you know this is a very big basic example of a unified namespace I mean it's very common for a UNS to at an Enterprise level to have hundreds of millions of topics very common it is incredibly common to have 400 000 to 500 000 topics at a site level at the site level so all right anyway hopefully that clarifies a little better you know the the approaches of unifying Erp and mes what I I haven't showed you the actual Integrations like how to integrate but what I've shown you is how to coalesce what I've shown you is we're we're starting with what our goal is we want to coalesce these things together and then we're working our way back to the connections in general Erp doesn't meet our minimum technical requirements out of the box so ideally let's say we are you know vanilla mqtt5 right so our infrastructure is based on mqtt5 uh what we would do is we would say you know very you know there's only a handful of erps that support mqtt natively why because erps are built transaction based unified namespaces are event based and here's a really good example every time the job status changes that's an event okay but if I wanted to take a snapshot of what the job status start time end time produced quantity rejected quantity operational availability quality rate performance efficiency and oee was at the time the job status or the the uh job status changes that's a transaction so if I fire this event and then I store all these things that's a transaction if I take it at exactly 12 o'clock that's a transaction if I take it exactly the top of the hour or every five minutes that's a transaction that's not an event an event is when something changed and when it changed so when a data point changed and what time okay all right that's my story I'm sticking to it thanks for watching like subscribe comment down below I'll see you in advanced mes tomorrow if not I will see you in the next one