 Good morning Paris. I hope you are all awake. Enjoyed the big party yesterday. I think it was a great occasion. A very hardly welcome from my side are Michael Michael Keenlem. I'm the CEO of IT Novum and I'm gonna give you some ideas on how to combine an open-source approach with software defined storage of course for supporting a very smart storage platform for OpenStack. You probably has not come across IT Novum yet. We are a German-based consultancy company focusing pretty much on open-source. So we have the vision to help customers build and run a data center 100% on open-source. That's our core mission and of course we're focusing very much on things like software defined data center and all these things. But as this is the OpenStack Summit today I'm not talking about data center in general. I'm talking about OpenStack and how we can support with different approaches OpenStack. One of the things which I'm gonna tell you about is open source and the other is software defined storage. I'm starting with the later one. Of course you want to use a smart software defined storage when you're just building up your own clouds. Why? There are a lot of reasons for that. Of course you want to have the elastic performance, the scalability, which is way easier with an SDS approach. You want to have multi-tenant service operations and you want to combine a lot of your storage approaches into one unified systems. But the most crucial thing when doing that is you need open APIs. If you look at the legal C storage system they usually do not have an API which does not help you to integrate your processes which you need when you're deploying clouds to the storage. But with OpenStack you have the choice. There are a lot of APIs which you can use. Of course there are some APIs right now like the current Cinder approach for example which is not very powerful but I think things are going there. But at least you have the choice of your underlying storage system without a vendor login and it's your choice if you want to go into a legal C support into a commodity approach or into an SDS approach or of course you can combine both. In terms of open source storage yes there are a lot of tools available right now. I think the real challenge is to pick the appropriate tools, combine these tools in a proper way that these are working on an enterprise grade and you do not have too much hassle and administration of these tools. You know the data growth will be gigantic in the next couple of years. Everybody's talking about 40-50% per year and the question really is if your IT budget can cope with this growth. Are there any opportunities to make it way cheaper than your legal C approach? Yes and with open source you have the choice and you have the technology available they can do it way cheaper. There are a lot of cost savings on the way. The next major thing when you decide to go open source is that you have no vendor login. If you look at the storage approach from legal C vendors you see they drive huge margins by bundling stuff but you're locked in. You have no choice to work on the APIs. You have no choice on really think you need this module to do that functionality and open source can drive you to that. It's not only about cost saving but adding a lot of functionality to that. One of the things which is quite important here, the automation and the abstraction of software from hardware especially with open source allows you a lot of functionalities to really use commodity hardware and build it up on the software defined level to an enterprise grade storage. So if you look at that framework picture you see on the left side there are a lot of services and tools which are available right now in the open source space. But if you want to use that on an enterprise grade level there's still a lot of work to do. A lot of work in combining a lot of these and believe me is there even dozens more and you need to combine that on the back end that everybody is working tightly integrated and smoothly and of course if you look at the red block here you need a common API. A common API which on the one hand provides all the the graphical user interface with all the function but also an open API which helps your cloud approach to address all the underlying functionality here in a proper way. And if this is done so in a smart way you can address all the functionality with the API and that's a major advantage of this approach. I'm going to talk you through a couple of use cases for an open software defined storage approach in an open stack environment. There are basically four there are a lot more but you know time is running out quickly so I just concentrate on these four. How to manage your commodity storage nodes in an open stack environment. The next is the automated provisioning of resources through an open API and the open API is quite important here. The third one how to provide your storage infrastructure with a high availability approach and the fourth how to monitor your storage config without too much manual work. So let's start off with the first one the easy management of commodity storage. You know if you're using commodity storage it's quite easy you get the the hardware you can collect a lot of tools if you're using about Linux as a standard platform and a couple of tools but then you end up in a situation if it's about to scale on a large level you want to apply a thousand or ten thousand of servers you'll leak a lot of petabytes of storage there's a lot of work to do from the administrative perspective. So what you really need there you need something which we call a storage operating system which is way more than a pure Linux platform to help you in doing this. At the top you see some screenshots of things we've done already to combine the things but I get later to that. So what you need really is a single point of administration for all these tools and you know when you're working with 30, 40, 50 tools on the underlying platform you need a tool which helps you everything to administrate easily and then you have the choice basically to compete with all the legacy vendors and providing all these nice features you need and you're quite used to when you run with one of the major vendors. The second one is about automation of provisioning processes. You know when you look at OpenStack especially as a Cinder module you see that the Cinder API and you see here in the green box the current API there's not much functionality you can just create a volume you can create a snapshot but that's basically it. If you look at the storage you're used to and functionality in the storage you're used to you see there are a lot of things you really want to have when you run an enterprise-grade cloud like mirroring, high availability, replication, consistent snapshots of applications like databases so all these things which are quite nice to have but this is currently not supported by Cinder so you have to find some other ways on the one hand using Cinder on the volume control to get the target but then there's a lot of work to do like high availability, mirroring, clustering all the stuff which is not done by Cinder so you need one more thing to take care of this you do not want to make it manually you think of some automation processes and again the open API is very important here if you have an open API to your storage system it helps you to comply these processes very easily. I've already talked about high availability if you look at Cinder and its native form you do not see much about high availability and of course yes it really depends on the type of seller for your cloud but high availability could be quite important even on the storage level if you're running an enterprise-grade cloud. High availability as you might know really depends on your setup, your policy, how of your employees, your distribution as well and of course if you want to go on a local or just a metro high availability scenario you do synchronous or asynchronous you have the amount of cluster nodes which is important so really helping you to pick up the right tools for that and of course you know if you look at the deluxe platform scenarios there are a lot of tools available like DRBD, clustering pacemaker which can really bring you a lot of value in getting to the high availability scenarios but what you need to do them you need to configure each of these tools individually which of course you do not want when you scale up to a large scenario. The next thing if you want to run enterprise-grade clouds is you need some kind of storage monitoring you really want to see if all the volumes are filling up if there's some some issues around that and you do not want to have a look at that if you look at OpenStack and Cinder there's no automated way right now in that so you need to add these functionality as well and if you're familiar with open source you know that there are tools available like Nages for example which basically helps you in monitoring all these things so the next one is if you just want to look at the old open source framework you need to combine a lot of things to really get to the enterprise-grade level for example doing SLA management SLA reporting on that as well so that's what we basically already did combine a lot of open source tools into one framework which is open as we are an open source company we called it OpenAttack and it helps you to build all these enterprise-grade storage functionality four different cloud platforms on 100% open source so you get all the features which we combined into these OpenAttack platform a unified storage with all the unified storage approaches you already have like different protocols and all this standard stuff a smart software defined storage layer which is fully accessible we are an open API providing you monitoring high-villability application snapshotting all the stuff and the cloud storage think with the specific cloud AP and the object or functionality as we are an open source company yes of course there is a free community addition you might have guessed that already but of course we offer other services on top of that so you see currently Cynder is just not enough for enterprise-grade scenarios but the good news is there are a lot of things which you can add on top on a top of an existing open source framework you just need to do it in a very clever way you need to combine a lot of things which are there and then you can provide a very high-end software defined storage on top of open source which works perfectly for cloud setups and as well for open stack and of course we have a Cynder driver so if you just want to have a look at it just go to openattack.org and hopefully you just like it and most important feel free to contribute there's a lot of things which you can add a lot of things which you can modify for yourself but of course communities are always living a lot of people are contributing I would be quite happy if you do so if you need any other things like services go to openatic.com which is the commercial version thanks a lot for attending in the early morning session today if you have any questions just feel free to ask me right now or catch up with me later as I'll be around here we have integrated already Seth from the API that you can deliver different type of the storage lunt which is then on the back end realized with Seth we have it not on the GUI in the open source version right now this will be probably in three or four months time they have this available they can have all the the rapid user interface to administer different things like the Cynder the Swift integration and then the back end Seth but the most important thing we want to have it on a on a general graph user interface approach one graph user interface to manage all the different storage which you have to support your OpenStack cloud any other questions you're probably still asleep okay if you do not have any questions thanks a lot for this enjoy your day hope it will be a nice day for you it was a lot of good ideas a lot of talks and see you soon thanks