Discover how Coca-Cola Hellenic Bottling Company transforms its Technology Platform | BRKFP75
thank you welcome all to this Microsoft ignite session I'm Jack the global sales team director servicing the Coca-Cola system based here in Atlanta with me today is Ben earlikam head of enabling technology at Coca-Cola Hellenic one of the largest Coke bottles in the world headquartered in Switzerland Sven and I have worked together on an Azure migration alongside Intel venomics and other partners and it has been a fabulous fabulous partnership hi spin it's great to have you here thank you Jack hello and thank you for having me um happy to share with you and all our listeners our Asia journey and the transformation of our technology platform before we jump into this I wanted to share with you quickly a bit more on who is speaking to you and what is the business we are in that you will understand better why we have to done specific choices on our journey so I'm as Jack mentioned working for um Coca-Cola Hellenic which is the beverage business selling products of the Coca-Cola Company exclusively in 29 markets mainly in South and Eastern Europe Ireland Nigeria and Egypt we offer a diversified portfolio of beverages in a growing industry you see the main share are sparkling drinks water juice energy and recently also we are having an increased focus on coffee and an increased footprint which we were very happy about besides that we also partner with other beverage businesses such as Monster Energy urgenton Brown foreman and Campari and sell their products obviously taking care of our people and our environment is a key priority for us we are a team of 36 000 employees and from a technology perspective you can imagine these are all our users um together with consumers and customers which are accessing our websites and e-commerce offerings this puts a lot of pressure on us as an I.T organization and also to the platforms which we are operating they need to be robust and as well operate 24 7. and there is always high expectations to Performance and availability maintenance should be minimized and tolerance to disruptions is low especially if customer facing processes are uh are involved so in a nutshell what you should take with you key priorities For Us customer and consumer centricity employee experience and operational productivity that's imperative for us and driving our decision making besides that and all the business and tech there are humans behind all that so these are the people which are driving the Asia migration and on a day-to-day basis and to also help preparing these slides so um thank you to that team and please feel free to reach out to them part of them are answering uh your questions in the chat so feel free to chat US during this session um yeah now the technology background um we are not new to Outsourcing I mean we have a long history starting in 2010 and we changed also providers which um we we did a bit a big lift and shift back in 2017 um which was obviously part of our DNA when we did this shift we we went to a new monolithic monolithic architecture and which was good at that time and filling our requirements um you will see on the next slide that today this provides us some challenges uh and was basically the momentum for change um while we did this Outsourcing Journey we started also our first Azure uh steps uh with Asia native services in machine learning with Azure data Lake and jack happy to talk about this or invite my colleagues to talk about this in another session so here we come to the challenges um yeah we have as I said the monolithic architecture which are basically split into two machines with the database scheme which you can imagine from one big database provider unfortunately over time these things were potentially in the beginning beginning proportional but as we have shifting workloads they became Divergent and not proportional anymore that locks Us in today and um having this architecture on aex maze is also very difficult to uh to to migrate to change because it's not an open technology and not a cloud agnostic technology obviously it is created as I said the momentum for change and you can imagine that we're expecting something else from our current uh from our new environment and that's the challenge we put out to check in the team we were looking for flexibility and scalability we wanted to handle resource Peaks we wanted to spin up and spin down workload for prototyping for Deaf work stuff which is rather difficult today and with um fixed workload um we also have still you saw it potentially on the other slide still country computer rooms 120 so quite a footprint which leaves room for optimization as well because they were grown and built in different times with different workloads and came partially to us with merchants and acquisitions so room of optimization there um not immediate pressing technical need but room for optimization especially when Hardware refreshes are are in the game so speed and automation was something else we wanted to leverage you will hear more about that how we went about this and obviously we wanted to have zero trust built in in the our new cloud and Cloud infrastructure while we were doing that obviously we were also using um Intel technology for RS for implementation with Jack you know it's not yet on Azure it will come later trust me but uh but uh well that was also a transformation which we ran in 2020 and to go live with with S4 on on Intel so let me continue uh with what were our driving principles when we were choosing uh when we were choosing um Asia so we wanted to have a platform which is uh portable between different Cloud providers we wanted to have our Geographic footprint covered um and Etc so you can use this for your reference um in in uh in in the in the post read of this session I'm not going to read you all through this slide foreign so what where we got a lot of help from you and um yeah you for sure appreciated the discussions on how do we group our workload um because it was quite cumbersome for us to get a business case if you remember Jack uh uh and we were wondering what is the right approach to that and brought it for up and down and right and left um what we what we needed to start with were obviously lending zones and network connectivity uh so we we built our Asia gateways we built our uh Express routes uh we also changed our VPN solution into a cloud VPN with that scaler so there were several Investments which we needed to do in order to be cloud ready and then we have as I mentioned before based on the challenges with our big Legacy platform or monolithics platform refactoring is our priority number one but you can imagine it's also the hardest to do we require involvement of our uh of our platforms and product managers who changed the applications which they own to serve our business with a different database and changes in their applications then as I mentioned the second priority were the optimization or is the optimization still of our country computer rooms wherever we want to avoid Hardware Investments um through the move uh to to Asia and um then last but not least we also have commercial of the Shelf already running on Intel and open Platforms in our Outsource data center today but still their opportunities moving them into a container-based environment in Asia or changing databases Etc then um what you see here in my if you don't measure you don't know how you're going and if you're doing well um so what we put out to ourselves is this kind of I don't know how I should call it a waterfall diagram of course and workload um so by the end of this year we want to have produced all these green boxes you see we are optimistic we have not moved all of them but we marked them already green so the green boxes are supposed and to be moved by end of this year and you will see on one of the next slide that we are in a good way and then obviously we will continue this journey until all the gray boxes are hopefully diminished by end of 2023 24 latest um that's that's our Target we started the journey we were wondering what options to take you can imagine that um an experienced team which had great experience with big bang that's where we were running to and you can imagine all our Consulting Partners technology Consulting Partners saying yeah big bang is the way to go to what we had to realize is that in a in a world with products um uh and business driving the priorities for us as I explained to you in the beginning well suddenly Big Bang was not working any longer uh we we decided for approach which was called wave approach today we could call it potentially increments and Sprints uh where we wanted to focus and and give or give the gear a bit to the business and to the product and platform owners because if I ask them to refactor they need to have a voice they need to say when is the right time to serve business requirements and when is the right time to work on technical debt so and when they can combine it so this is why we came to a wave approach or an increment approach which goes very well with our scaled agile framework so um yeah this is what what we decided for and that's what what is delivering us um the biggest value it's it's sometimes um uh not bringing in the first look it would not bring the speed if you think I over a long time I think it's making us also faster because when the people go about it they they really execute and they execute very targeted they own it um also it allows us to to onboard continuously learnings and improvements because with every workload you migrate you learn something and this is what we can reflect um on on the goal and also it allows us to be selective about Native Technologies um like kubernetes which we can use for dedicated workload which in a lift and shift environment you wouldn't have the room and the time to do so so for us that that was the right choice obviously everybody has to make a different choice or his choice based on on his environment I'm stating that what was working for Coca-Cola Hellenic or is working for Coca-Cola Atlantic at the moment um if we continue to the next slide uh you see what we achieved so um from a lean perspective I would say we are here having a continuous flow of delivery um and this is just a reference picture I think it's not fully complete there are other stuff which we managed to complete even on top I mentioned to you that the the data Journey which is not reflected here so all the green spots are are migrated these are country computer rooms as you see it's a specific um of the Shelf Solutions and you see also increasingly since maybe no end of last year when we started to move with the support of Economics uh the the first sap Supply Chain management system in a in a POC um to Asia a lot of sizing had been done there with great support from Intel and trying to find out the right sizing for the needs we have here you see then we fellowed with with more development and Quality Systems the first production system of solution manager fellowed in August this year um in parallel keeping consolidating country computer rules um if you wonder why Netherlands is here it's not part of our territory but we have an office there that's why you see the servers from there I think about the I noticed the CCM module is painted red can you share more about that uh should I really mention about red stuff okay uh yeah I mean as I said we want to experiment and try and we were a bit ambitious here um when as as I explained we went live with S4 you you know that and then we had some challenges in application servers and we thought let's try to put um application servers for some heavy jobs into Asia but due to network latency we didn't manage to get it work if somebody tells us the secret we will we will make it work uh but uh so far that that did not work but that's part of the journey as I said we are learning and uh proving yourself many people might be on the cloud for that for some time already for us there was no it worked but there was no business case which we could make work back then that's why we scheduled it for this year I think now we have a business case and towards the end of the year we'll also try side tricks in the cloud what you see here I mean is deal with the different boxes don't necessarily show you the workload so some systems are easier to to to to to do and as you see sdo it's it's it's the small boxes another one is the other ones but it's a major refactoring effort uh which the team drives greatly and it's it's also bringing us a lot of new technology which we're which we are learning all this you cannot do without a great partner environment and a partner ecosystem we obviously have lots of knowledge in-house but not everything and the team as I said is learning onboarding new new skills um but uh without Intel Microsoft and venomic we would not have been able to make it um binomic was helping us with this with scripts um to accelerate the and get us going very fast with the proof of concept as I mentioned before but also with deploying further sub workloads and we continue the collaboration here and as we have more sub workloads to migrate um Jack your team helped us with uh with great support in uh in um designated Engineers I think is the term you're using for uh for your support model uh your unified support where we get um really good help in the moment we need it because um yeah we'll come to it sometimes you face issues and then you need skills on the ground which you didn't plan for and right sizing I think is great help is of great help and better performance is what I mentioned in the beginning we need and here the collaboration with intel was extremely and is extremely powerful in value-adding Goods so wave waves rather than Big Bang Yeah that's I I think I I transferred it very passionately that this is one of our uh main learnings um but uh there are other learnings we we have uh that uh the solid design of Landing zones from day one is something which Jack your team was always preaching but you know when you do it uh you're learning it and um you know what what it means in terms of getting policies right getting identities right well you can correct it but if you have to correct it you have to stop to it right and and and then move on so it's slowing you down uh database migrations is a favorite topic of mine as you see we went into a very uh diverse environment from a very up on one provider focused in the past uh so we're leveraging among others postgree uh we are leveraging increasing the mssql and when you do these migrations you have tools they automate the process I mean you can just do that with the automated tools however don't underestimate that these tools don't resolve proper designing thinking and respectively performance testing um uh we we we had to learn that even if you have good user feedback when you send hundreds of users on a database it behaves differently this sounds very logical but it's very difficult sometimes to create the right testing Automation in a lab to to have all the different situations um uh constructed in a test and case which you require and in order to ensure that you have proper performance so um we we keep a focus for any further refactoring because that's that's one one thing where we where we can always be hit CIS hardening security you can imagine in a company like ours is extremely important I guess for all of you here in the room so we had learning here with CIS hardening dynomics was also here helping us very quickly we improved the scripts and the templates used and now any further systems are hardened as we need it and I may have read it but and learned it and people told me that networking and networking and again networking is key uh especially in a hybrid Cloud where you have still on-prem stuff where you have country stuff um you need to be really and you have as you saw our geographical footprint different isps involve different FTP SFTP providers um so so so networking is really key to make things going and especially with a lot of security policies in between and certificates and Azure key World playing in which is all there for good reasons and to secure but to make it go you really need to design carefully you need to configure carefully and obviously also test it as I mentioned foreign last but not least whatever you design architect you may have a problem and and here you require a team that owns the problem and the resolution and you saw the number of Partners and that was just a subset of Partners which I showed you it's very easy the people to say ah it's not me it's him uh it's her it's uh that if if you have a problem if you get the people in the call if the people say ah well I can contribute here I can contribute here you have half of the solution already because you have not only the skills but you have the people who just want to resolve it and obviously they need to resolve it fast when I say fast I think there is this difference between responsiveness and speed you can be fast but we have reacted too late so it's important to be on time when the issue appears and to get the things uh delivered this agility I think the other thing with the team spirit um I mentioned and well I don't want to get give the impression that chaos and all the time messing around is is this retrution but uh discipline and structure in all the approach and also in a responsive and agile environment is is key in order to speak the same language and to get the um get the things done in a disciplined way and to have a continuous flow as you saw it in the in the plan before thank you Sven this has been a truly amazing exposure on your Azure migration in the co-clinic context for our audience we have the call to action here there are some links from Intel and from venomics that we'd recommend you check into I'd like to also thank Sven for sharing his insights today on the Azure migration and thank coconic Intel venomics and our other partners who've made this migration so successful
2022-10-16 09:40
