 All right, we'll go ahead and we'll get started. That's a little loud. OK, my name's Nate Zeeman. I'm with the IBM Cloud team. I'd like to welcome everyone to the enterprise journey to adopting OpenStack, real world stories. Today I have the pleasure of introducing our two guest speakers, NTT Data, one of the largest Japanese system integrators, as well as SAP, a well-known enterprise application company out of Germany. So with that, we'll go ahead and kind of get started. I'm going to start with Kokasan, who's an IT architect for NTT Data. He will be speaking about a project that they're using OpenStack with an adopting OpenStack for Kileen. So with that, I'll go ahead and bring Kokasan up to the stage. Hi, my name is Atsuhiko Ga. I now I speak in Japanese. First of all, since November last year, I've been working with a famous beer company in Japan. It's a private cloud company that uses OpenStack. That's what I've been working on. First of all, I'd like to talk about the background and explain it to you. First of all, this paper shows the environment of Kileen before opening OpenStack. As you can see, there are five data centers, including the backup centers. On top of the data centers, there are mainframes, such as HPX, AIX, etc. There are various environments. The number of servers will appear later, but it's about 2,000 servers. From three to four years ago, there was a part where I was working on some of the virtual contracts, but it was still about 7 to 8% of the 2,000 servers. One more thing. In fact, about half of the 2,000 servers started this project, and in 2015 and 2016, we were in a situation where we had to replace the EOSL. While we were working on it, we were thinking about how to replace the EOSL and how to use OpenStack. One more thing. In 2016 to 2018, we wanted to achieve a drastic cost reduction in infrastructure planning. I can't say the exact number, but we wanted to achieve about 20% of the cost reduction. From hardware to data centers, we wanted to make sure that the infrastructure and the infrastructure of the infrastructure and the infrastructure of the infrastructure would be in the right direction. In that context, we used OpenStack. That's why we have these three points. When I first started this project, I didn't even know about OpenStack. In any case, we had to replace the big servers with the low-cost servers. We had to cut the huge costs. There was no input in my head. From there, I used a lot of vendors. I had to put my voice on IBM, but we had to use EMC or Red Hat. We had to figure out how to find solutions for the vendors. We tried to make the private cloud out of OpenStack. For about two months, I've been working on the vendor hearing. First of all, I've put in the keyword SOFISCATED. In fact, I'm also interested in the environment of Kirin's enterprise. When it comes to this open stack, I've been looking at whether this open stack technology is the enterprise level, or whether it's a technology that will be thrown away in the next five or ten years. That's what I'm talking about in the first place, but I found out that it's a very excellent software based on a very well-designed architecture. Next is flexibility. In fact, based on the cost-effectiveness, what I'm looking for in the private cloud is not only the influx of the IaaS like AWS, but it's also a set-up of the middleware or the installed environment. I wanted to make everything automatic around this area. I was looking at whether I could create a variety of things on the IaaS, but I found out that the open stack could be customized. The third point is vendor neutral. The software that we've been using these days is a direct connection. When we use this software, we often have to use vendor products. We don't have a hardware company, so we can use any vendor or any hardware to choose from. That's what we've been using these days. OpenStack is a one-stop product that has a direct connection. OpenStack's software itself is not a product with a hardware footprint. That's the third point. Next is the three points that we've challenged in this project. In the previous explanation, we've put out some key words. The first is full automation. As I told you earlier, it's not on the IaaS level, but on the Perth level. The second point is mass migration. After creating this environment, we're waiting for the sensor server. We'll explain a little more in our session from 7.30. In fact, rather than creating a private cloud on the Perth level of OpenStack, it was one of the challenges for us to achieve this sensor server. The sensor server in Kirin is an enterprise-level system, so we can use it immediately after we switch it over. That was a very difficult challenge. The third point is to achieve a drastic cost-effect as I've already told you. Lastly, I've talked a little bit about how we created it. I can't talk about the time, but at the end, I'd like to talk about how this project was introduced and how it worked. The top point is the flow of infrastructure based on the previous work. By using full automation from design to building and testing, we were able to cut the cost-effect of the 75% of the cost-effect spent on infrastructure. This project started in November last year, and the environmental construction was over in September of this year. It's been three months since December and December, and it's been mass-migrating. It's been about 30% of the cost-effect, but at the end of September, we were able to evaluate the cost-effect of the new environment that was created. This is the result. So, I'll give a little explanation about this project, but it was a good idea to create a private cloud at the level of IaaS only with OpenStack. And as I explained so far, we were able to implement the automatic operation of the Perth level or the automatic operation of the test function. The orchestration function is also connected to the standard function of OpenStack, but the level we were aiming for is very high, so it's quite difficult with the standard function. So, this time, we used the function of IBM Cloud Orchestrator which was created for OpenStack. So, rather than the function that was originally made on the physical server or on the virtual server, the orchestration function was created in a way that the standard function was quite high and this time, we changed the function to OpenStack so that the original product would be the one of the reasons we were able to use it. But, in fact, we were able to use it and there was almost no problem. We were able to turn the past into a year and I myself was responsible for the project but using OpenStack and using OpenStack was the most difficult part but in fact, there was no fatal problem with the test phase. I think it was a very stable and reliable product. Finally, there are three months left after this project but we were able to learn how to do it and finally I would like to end my presentation but well, yes, using OpenStack to create a private cloud I am sorry but it is not that difficult. Which one? OpenStack is the way and what we achieve this time in this project we were able to achieve a full-automation but without such a goal it would have been a different direction simply but because we were able to understand and understand what was going on we were able to do this. There are many things that I would like to talk about but OpenStack is not full-o-s-s for example, Hypervisor using ESX but it was not a goal so we were able to achieve the cost so for some risk hedges it is not pure OpenStack for example, using IBM's ICO and Hypervisor using ESX and so on and so forth Thank you Thank you, Kogasan. That was very enlightening, I appreciate that. Next up, we have Marcus Winter Marcus is with SAP Thank you very much So, welcome everybody My name is Marcus Winter I'm with SAP I'm based in Waldorf, SAP's headquarter and first of all thanks, Nate Thanks for IBM for the invitation to speak here about our journey towards OpenStack with the cloud management products that SAP is working on for our customers So, first of all SAP is the leading vendor in enterprise applications we build software that power the world and in the end there are some very important depending industries running on SAP software and one of the facts that I really do like is that for example more than 60% of the beer producing industry runs SAP software so in case anything is wrong with our SAP ERP and enterprise solutions the world beer supply would drastically go down and so this is how important our software is to most of our customers and we ask them very early what does the cloud mean for them we have been shipping software for more than 40 years SAP is quite an old company in the software industry when we look back software has not been around that long and so we ask them what do you need from us from SAP in order to run your enterprise software and we got various feedback and one important thing was that of course they want to keep the stability that they have from their software that has been running for a long time they also have a lot of legacy software that they still need to keep up and running and they expect from us the necessary tools that in this new flexible world where a virtual machine can be spun up within seconds they have the necessary tools that their SAP environment can keep up with that flexibility that they gain now when you look to this slide I brought two pictures this is actually not very far from SAP headquarters so in case you happen to come to this tiny village of Waldorf about one hour south of Frankfurt Germany about 15 minutes away from SAP there is the place where the car was invented where Carl Benz and actually it was his wife who did the first drive where they started manufacturing cars and at these days building of a car was a very very manual process you can see in the left picture there were many many people involved in building a car and the cars were very individual because they were handcrafted when you look to the car industry today there is a very different picture there is a lot of automation there is our robots building the car and we went towards mass production and a very high volume and a high quality of cars and the IT industry also in the enterprise segment is moving towards a very similar direction so from handcrafted individual setups to a highly automated way and this is where open stack virtualization and what we are listening here at this conference plays a very important role when we look at the costs and this is a real example from one SAP customer I couldn't display the actual numbers but the percentage how is distributed you can see that in the blue bar almost 75 percent of the costs are operating the system the enterprise software and the administration so this means there is still a lot of handcrafted like in the cars so there is still a lot of room for automation and improvement and optimizing this cost block and this is what we try to address in my business unit so the task of my business unit where I'm responsible at SAP cloud management is helping our customers with the necessary tools so that their classical formally on-premise software can be automated and benefit from the flexibility that we gain in clouds when we look there is a journey happening on the left side of the picture of the individual physical server installations we introduced virtualization to the data centers to consolidate the workload onto fewer hardware then automation was added and we suddenly start talking about private and public cloud environments and in the enterprise segment where our customers are this transformation journey is happening a little slower than in the rest of the industry because remember the beer production if there would be no beer there would be a big problem so they are really careful about adopting new technologies and want to make sure that when they invest in something new that the stability that they have been building up for many many years can reside and from an SAP point of view this does not mean that all customers need to sit on the right side there are very good reasons for having physical server installations for having software run on virtualized servers and there are other very good reasons for moving towards a public and private cloud so what we will see in the next years is definitely that there is a hybrid environment that is very important to our customers and this is also important to SAP to support our customers in their environments in the future now we have created a product it's called SAP landscape virtualization management LVM that helps our customers on this transformation journey and automating their SAP environments in a central way so many of our customers don't have one single system they have many installations they have a production environment, a staging environment a quality assurance environment a development environment and they have distributed the workload over many servers so over many years our customers have built up big landscapes where there is a high need of automation and on the other side our customers also want to keep up flexibility when you talk to five customers you probably will get ten different answers how the data center environment looks like various operating system combinations various choices that they have made over the past years so automation is really challenging that's why we very very early started thinking about and listening to OpenStack because this was also feedback from our customers that it's an interesting and very important development to them that they can gain more flexibility without having a vendor lock in and having a community supporting what is happening out there in the area of cloud automation landscape virtualization management consists of three pillars it's the provisioning of SAP systems it's everything managing and what is important and what I will emphasize today here in this talk is the extensibility so we know that we are not alone in the data center this is fortunately and unfortunately for SAP but we know that customers are running more than just SAP systems so we know that we need to be very flexible in managing SAP systems in various environments this is also true for the product LVM that can use can be used in multiple different technology environments we work together with many partners in software industry so that customers whatever choice they have made of storage vendor or virtualization vendor that this can be automated with this tool in particular for SAP and this was also the reason to build open stack adapters for LVM in the picture on the right hand side you see a high level overview how this looks like so the LVM software itself has an interface that adapters can be written so that LVM can talk to different technologies and on the one side we have a virtualization adapter so to control infrastructure and virtual machines start stop virtual machines create virtual machines deploying SAP on top of them and then we have secondly what we call a storage adapter to simulate storage volumes this is important when our customers for example want to create a test environment from their production system to test out a new software patch or test out a new add-on to their software now these adapters need to communicate to the underlying infrastructure and therefore we have written to the different adapters so that the customers SAP customers can directly talk out of the management tool LVM towards open stack infrastructure and again open stack is not open stack the situation at customers can be very individual so we here teamed up with IBM and tested our adapters against IBM cloud which is an environment that is accessible for us and that is accessible for our customers so we can make sure that the functionality that we have implemented in the product is working and that our customers have a reference where they can also test out these adapters and make the necessary adjustments we also decided that these adapters will be released in a open source way so they will be available at GitHub we just completed this for our event last week at SAP TechEd we made the announcement that these adapters will be available and we started working especially over the last half year together with IBM on testing those adapters and of course in the beginning things are not always as easy we also realized that there are challenges in getting such an open stack environment up and running because you can see there are multiple components you need to know how to configure it and we also started building up our own internal open stack environment and challenged our IT guys I think in the middle they gave up and then we had to motivate them again to keep us supporting so things are not that easy and how can we access it network configuration issues in the beginning so there were several hurdles that we had to overcome and there is also room for improvement and I think that this will be also challenging in our customer environments in the end this is I think a very good starting point and when I look towards our enterprise customers we are a little ahead of where most enterprise customers are today they are looking at open stack they are testing it out I was very happy to hear in the previous presentation that entity data is already using open stack in an enterprise environment and moving to that we see that these projects are happening now and that's why it was very important for us at SAP that we are also ready for those environments and that our customers can use open stack and we are part of that journey with this being said I think that is an important step for us SAP joined also the open stack family a while ago it is also part of our philosophy to give our customers choice so this fits very nicely into our strategy and we get positive feedback from customers where we talk that this is what they wanted to see that they have the possibility to build up their own environments and use these open source adapters and open source style of automation that this can be influenced also for their particular needs my name is Marcus Winter if you have any questions this would be the right point of time to ask I will also be available later after the session and thanks for having me here so the question was about a hybrid environment where things should remain physical, where should things remain in a virtualized environment this really depends on the workload SAP is investing a lot in a new in memory technology for example SAP HANA are together with Intel and other partners to really get the last few percent out of compute power so that you can browse huge amounts of data in seconds and customers are also investing a lot in this hardware so here they are a little reluctant to move that into a virtualized environment because when you have a two terabyte in memory database a hypervisor that maybe adds five percent or a little more on performance impact this hurts so there is a tendency where in these environments where a single HANA database consumes up an entire physical machine that this is not being virtualized on the other side we see that those customers are virtualizing the same database but in their test environments where there is not such a huge data load and where they need the flexibility to quickly spin them up and put them down so I think it really depends on the workload and also on the size of the landscape and we see that there is definitely a trend where our customers have been virtualizing a lot in their data centers and SAP is one of the last bigger baskets where they are now also after they have made a lot of good experience with the data centers where they are also now moving into projects to bring up SAP on the same cloud stack as their other applications any other questions that is not the case thank you very much Nate back to you