 Hello. Hey, everyone. How's everyone doing? Yeah, last day ready to leave get home burned out Yep, sleep. Yeah So my name is Greg Lockmiller. I work for NetApp I'm a technical marketing engineer been with them for nine years and various roles from Rack and stack storage designing storage putting applications on it working in the emerging technologies and things like open stack Part of my focus within NetApp within open stack is really just to for lack of a better term Evangelize Manila, you know share the information about Manila Work with the development team and work with customers about what Manila does and what it can help and as well as Actually digging into code and seeing what breaks or doesn't break. So Again, I'm Greg. Yeah, I'm Thomas Thomas Leach. I'm also working for NetApp. I'm a cloud architect based in the mirror So, you know today's session Manila 101 and then use cases We're gonna go through a little bit about what Manila is try to give you a basis Talk a little bit about that talk about what's new in Kilo for Manila and then we'll get into some use cases and we if some of you were here earlier with Thompson Reuters You heard a little bit about their use case. You might hear some repetitive information there But talk a little bit about where it fits in what people are doing with it What we see with customers and how it helps solve problems So As we mentioned, we're also gonna have a demo. Sorry. I forgot to mention that so we'll have a demo of Manila So you can see a little bit about through horizon how it works and what can be done Again, I'll talk about what Manila is give a couple of brief slides there We'll have a demo and then we'll we'll kind of get into it You know why of the shared file systems and within OpenStack and what it does there so If you were here earlier again, this is a this is a slide where and I'll just kind of give a little history a Two years ago a little bit more than two years ago now You know there was a group of people in the community. It said, you know a lot of a lot of customers want NFS They want sifts. They want that shared file systems in the cloud and into this is really just to kind of iterate in 2012, you know, there was a survey done by IDC It said, you know, look there's 65% of the storage delivered and sold for all vendors, you know And that was for shared file systems And so there was a group within the community like I said they got together and proposed made a proposal to the technical committee and said Hey, you know, how about a shared file system project within OpenStack? They recognized a gap in both the technology being available as well as a solvent of business needs so So that's kind of what started it all off Manila is based out of or off of sender so it gets its roots from sender I think if you know sender pretty well, you could almost and certainly understand Manila. It's got very similar architecture It's got a same configuration type things items with a Manila comm versus sender.com And you'll hear about some terminology that has carried over from sender changed the name a little bit But the concept is the same, you know just on another note here just real quick You can read those bullet items, but there was a an annual report done by IBM that you know Where they're predicting in the future about the number of applications and the percentage of applications being cloud enabled Going to the cloud and even moving Those traditional applications that you wouldn't think that would be cloud-enabled things like an SAP or databases You know the things that continue on as the maturity and stability of things like OpenStack get better People are putting more of those type of apps in there So just out of curiosity, how many people have NFS in their infrastructure today? Anyone? And from NFS, do you guys Do you like NFS because of its ease of manageability? In other words, I don't have to I don't have to create a line I don't have to carve up a line. I don't have to format the file system. I mean I find personally I was In my career used NFS quite a bit and found it very easy to say okay I can share this file system between three or four bare metal or VMs I used it for cluster databases that way. I didn't have to worry about different types of Application lock managers. I just let the NFS lock manager handle it So I felt like NFS was pretty you know pretty significant for certain applications as well as ease of management So that you know, that's some of the reasoning behind trying to bring NFS to the cloud And who has recently heard of the AWS announcement? Do you guys know that AWS now offers an EFS? so we've had Manila for in the community for two years and Just recently AWS has come along and said hey, we've got an elastic file system So I think that adds credibility to the need of solving a you know Technical gap in a business need for even those folks that are within a the AWS spacing I need to have that and and so again within the private cloud or being a service provider using open stack We obviously you guys are here, you know over the past two years You've seen quite a bit of maturity and improvement amongst it and and so you see a lot more Enterprise type customers going to it and then bringing in NFS to it as well. So real quick What is Manila? It's a multi-tenant secure Excuse me shared file service Think of what sender provides to block services Manila provides for shared file services. So I can do NFS or SIFs It does offer some security services. I can provide Kibara's or you know LDAP things like that. So You can provide lay a level of security You can use it again for both NFS and SIFs and it provides the shares to VMs or bare metal and As what Manila does is will provision a share and a share can be based on your protocol in a Fester SIFs And then you can allow access or deny access to that share So I can say for example, I create the share and I may say, you know Thomas You can't have access to the share so I can deny Thomas's host access or I can grant everyone in the room And we can create one large clustered file system We actually see that type of use case in universities and other in high-performance compute spaces where they want to be able to have You know one file system they drop everything in they run their test do the analytics and leave it there So, you know just different types of use cases there and how people use it, but again, it's a multi-tenant secure environment It's a lifecycle management of NFS shares. So I can create it. I grant access to it I can create snapshots and then I can say, okay, I'm done with it Let's delete it and break it down and we're going to talk a little bit later about the different types of deployment models But you know one of the one of the first things that next thing I'd really like to Ask is if Thomas could show us a demo, you know instead of just sitting here talking about it and getting depth by PowerPoint like all week You know, I'm gonna give a demo or Thomas will give a demo of Manila and we'll talk a little bit more about Details of what's coming in kilo and then we'll get into the use cases Bear with us while we swap here. Yeah. Thanks, Greg. So We did a little recording here so this is a familiar a horizon dashboard a lock-in and Manila actually provides an integration with horizon. So it's very convenient to use as you can see here We we currently don't have any shares Deployed in our environment. So the first thing I would like to show you so there's a new menu item and a new kind of overview Of shares here. It's very similar and reminiscent of how volumes look like for Cinder. So let's see If we wanted to see how this looks like on a net-up system for now I guess some of you are familiar with net-up and the system manager So this is a view on the storage virtual machine and as you can see there currently no so-called flex volumes deployed so when we now go back to the open-set control plane and Quite have a look at the creation dialogue for a new share. We can give it a name I think that's very familiar. We can give it a size and there's also the possibility to actually Assign a share type to to a share. So in this case We have a default share type and we also have a share type which is called SIM provision You can give it any name basically, but this should reflect what we kind of tried to achieve here And I'll get to get get to that in a second so the shares now to be deployed in the background and If we have a look at the share type again, this is something that the cloud administrator is able to define and to Create so I'm going to look in as the admin and There's again the shared types and as you can see here We have a default share type as I said and it's in provision share types and there are So-called extra specs that you may be familiar of from Cinder as well. So this allows you to unlock driver-specific Capabilities in this case we have defined a share type called SIM provisioned and we enable the SIM provision flag for the Flex volume on a net up. So any other driver or any other vendor driver could potentially also expose different capabilities And you can actually reflect workloads types or bronze silver Gold types of services in this regard So as I said in the background we created the share. So that's now available As you can see here When we click on the share we Like with Cinder we see the details and one of the nice things here is that It has a you ID obviously like almost any energy and open stack and the share actually got created on the back end So there's now a new flex volume, which has a share underscore prefix and the same you ID reflected And it's also as you can see SIM provisioned. Okay, so that got also transferred to the back end So the next thing we wanted to do is to mount this actually in a virtual machine that we deployed earlier So for that we need to assign or allow the this machine to access the share through NFS So we would like to add the IP address of the client to the access rules list so we do that so as you can see there's a Dropdown in your item and you can actually manage the access rules for that So currently that's empty And as you can see here, we can add a new rule. So in this case, it will be IP based if you are deploying such shares You can also actually use the SID type to actually assign allow users to access that there's also an access level so that can be either read only or read write and for now we will use read write and then we add the IP address at the rule and we should now Be actually to be able to mount this share from the client VM. So let's do that next We will in the details of that share will copy the mount location lock into the machine and as you can see is this is your typical mount and mount NFS command We supply the export location and mounted to mount share so let's CD into that and Create a new file called hello world So that's now sitting on the share So what Melilla also allows you to do is not only allows you to create new shares But there's also a nice feature called snapshots. So that's a point in time copy of a share so let's do that next so Just like with scenery can do a volume snapshot. We can do a share snapshot Again, what this means that's driver specific in the case of net app What what this translates to is that we do a flex volume a snap the volume Sorry a net app snapshot of the volume So that has an idea as well and as you can see in a snap manager once we refresh the view at the bottom You see that there's now a new net app snapshot of that flex volume created again the snapshot is created and One of the nice, you know next step that you can do is you can actually create a new share from a snapshot So in a sense this allows you to do clones of a share Depending on that point in time copy. So as you may be aware, you know, if you want to do parallel testing for instance on Where a database is involved that you need to Go from a clean state. This would allow you to create in parallel multiple Environments that you could run your tests against so let's do that next create a new share from that snapshot We'll give it a name So that chair got created and it's now in a sense a copy of what we created previously To prove that point on the back end this means again There's a new flex volume which is actually a clone of that snapshot. So We would to demonstrate that this is actually working we will also allow the client to access that new cloned or copied chair and We will remove the file that we created earlier from the original system of sorry from the original share What does it actually mean if I allow a client to access a share the export rules on the net app system will be Adopted so that you can see here that the policies got Got adjusted to what we entered in the open stack Horizon so we again copy the mount location of that news share Mount that on the client as well But first of all we removed now the hello world from the original chair and mount the snapshot or the share from the snapshot And as you can see the that file is still there because obviously we we snapshot it the previous state of the share and This was kind of like a demo and I think it's quite easy to see how convenient it is to use from from the horizon dashboard and that is actually, you know Integrated very well and in the standard workflows that you might have in your environment, and I would like to Invite Greg to to discuss a little bit more into detail how the architecture looks like So we'll get back into the slideshow mode We're with us So let's talk a little bit. Excuse me talk a little bit about You know what Manila is and again this may look quite familiar to you Manila is takes its roots from center. So you have the Manila scheduler who decides Okay, where do I want to place this particular flex? I'm sorry share, you know is it'll look at multiple back ends it'll Schedule those based initially on capacity or capabilities and when I mentioned capabilities Thomas talked about Extra specs and volume types so you can again provide extra specs and share types and then if there's a capability that you require or a different type of Storage feature and then the scheduler will take care of that and then obviously the API where you can talk via rest API at restful API calls horizon or you can we do command line There has been a couple of projects done where Groups have integrated to the Manila API As well as from the Manila API service to so you know some ways to integrate there pretty quick to do that too Actually be quite honest with you No different than what sender and others that uses the the rabbit MQ For example here, and then the Manila share process and there's a Manila share process for each back end that you define There's the management process and then a Share process for each back end so One of the things that you can do if you're Manila back in very similar to sender is that I can have multiple back ends defined They can be arranged to talk to the same or different storage arrays that you support behind it I can have multi drivers. I can have a net up driver EMC driver and a cluster FS driver for example Defined in the three different back end so it's not where I can only run Manila on one storage array or one vendor But you can run it on multiple vendors so it can fit whatever your infrastructure has and how you can support it too And we're I'll show you some of the vendors that are available here in the next slide But again, so this is a very this looks very familiar if you're if you've seen sender and you send her quite a bit And the configuration is much similar too So a couple of things real quick So we've these are some of the new drivers and the current driver set and this is what came out with kilo I've kind of earmarked those that are in the stable kilo release Initially for the past couple of years There's been the net app cluster data on tap driver a cluster FS driver as well as an EMC Isoline and VNX driver and then just recently that there's has been Within the master branch and some of the split branches, you know the other drivers, but now within kilo These drivers are available So if you went to get hub wanted to see what was there in the master kilo branch Yeah, Matt the kilo stable branch you could see that these particular vendors have now provided drivers for Manila So we think that again, you know kind of out of you know Trying to promote Manila that other people have seen the need for it to solve a business as well as technical gap So you got the other drivers there, you know, I won't go through them You guys can see I'm just a real quick the HDFS one is for Hadoop to just share that with you So let's talk a little bit about Some things that are new with Manila. I won't again. I'll go over two or three of my won't read These are some of the bigger things at least from a net app perspective some of the things that we've been doing and working With I'll share with you guys a brief discussion about deployment models of Manila that was community driven and the network plugins that are available now Talk a little bit about the storage service catalog We do have heat plug-in which is it's not quite upstream so you can get a heat plug-in And maybe you want to you know deploy a stack and maybe I need ten Manila shares eight cinder volumes and you know Two or three VMs things like that a couple of things that's going on in the future. That's not in kilo just to share with you You know some of the future pieces of Manila as a community driven project is a replication with men within Manila as in share Mount automation so one could be able to define how shares could get automatically mounted within your VM today. That's not possible Again, it's you know just the progression of what the project's doing and how things are moving forward So one of the things we'll talk about here next is along the kilo release is deployment models so There's a couple of different deployment models that I'll be brief here I can do a single SVM or multi SVM and and so you're probably saying okay great. What's that really mean? Well Earlier releases of Manila from ice house four would say I as a share in a Manila deployer would create what they call an SVM it's effectively I'm going to go create an NFS server and It would build up that NFS server and break it down based on the share network and it was integrated with neutron So if you went to a Juno or ice house release of Manila, that's what you would get You would have what we call the multi SVM model where when I went to create a share It's going to build that NFS server for you Now with what we're saying is it's going to manage that SVM creation breakdown in the SVM management, right? Single SVM is a new feature with kilo where the idea is to simplify the deployment Make it a little easier for those that might want to do it a demo or just put it on a Dev stack deploy it with on their Laptop with vagrant and things like that so you can have a single SVM and what we're saying there is have a pre Configured NFS server. Maybe it's a storage virtual machine for net app Or maybe it's the reference implementation that comes with Manila and it creates the NFS server for you so and what we mean there is that'll be configured within your Manila comm it won't manage or Create or break down those NFS servers when you run Manila in the single SVM mode So those are that that's a couple of things that are new with Manila in the kilo release The other things that are new that are quite significant are the different plugins for network like I mentioned earlier The network plugins that initially with Manila. It was the neutron plug-in, you know great I've got an ml2 agent. I'm gonna also have a layer 3 and Be have that integration to neutron and we found Both in the community and as well as it net up that not all customers wanted to have Neutron they didn't want to have to have you know build out all of the layer 2 activity now Don't get me wrong most people have VLAN tagging and everything in their network anyway, but from you know Being able to deploy it easily or don't want to change their architecture or their infrastructure. So the community got together and said well They built they allowed network Network plugins for the Nova Network plug-in We've heard where people still wanted to use the Nova net type activity and then finally the standalone plug-in too So I can have a flat network standalone plug-in network plug-in for Manila to use and not worry about neutron And it when with each of these methods you can you know for example with neutron Neutrons got to define networks the IPs and things like that how it's been built and how the layer 2 connectivity and in Packets are designed, but you can also what happens is with Manila Manila would go grab network information from Neutron you'd have to define that in other words I would tell Manila hey Manila create a share network and you're gonna use the neutron net and neutron subnet And then that's how Manila would build network connectivity for the storage virtual machine on the back end So if I if I were in the multi SVM with neutron, I would create see if this works there we go I would create an SVM within my storage and I would use Neutron information for IP addressing and the segmentation ID so now with Kilo, we've taken that Still have that but taken it the next step further where you can use standalone plug-ins and define some network elements that way Or use Nova networking so those are some of the changes that's been Driven into Kilo been a lot of work by the community within the Manila team and the core team there The other thing that I'll talk about is Thomas gave a demonstration of the the share types and and this is somewhat like a storage service catalog that you can get with with sender and use of extra specs and with the Volume types and sender we have extra specs for the net app driver, and then you can create share types Obviously share types are most certainly not just net app driven But what we're trying to share with you here is the ability to surface up net app technology Through the abstraction of Manila and have different ways of provisioning shares that maybe your Architectures or your standards define or that they require to for your tenants The share types, you know just an example here. You're gonna have database or analytic type share types archival Maybe maybe you have like I mentioned earlier Maybe you have a bunch of different back ends and as long as there's a driver You can define a share type for a specific back end and so for like an archival share type Maybe that's the cheap and steep So the other cheap and deep storage right, you know, it's something that is You know, you know very large footprint very cheap It can be whatever we if there's a driver for it then you can define that back end and use that driver Or you can use things for like database. I mentioned earlier about you know, we're starting to see more traditional applications go into the cloud and so a Lot of guys that I worked with and customers I worked with like to have an FS for their databases for manageability quick backups quick recovery Persistent file systems and things like that. So then you can use maybe to find your back end there to use Maybe a higher end back in back or a storage array as the back into things like that So, you know, it's it's the storage service catalog is meant to help you as a cloud administrator and the storage administrator Provide options to the tenants or maybe isolate a tenant to go down a certain path. So he doesn't interfere with other things, too so One of the things again, we all think about the cloud and open stack and the control plane and it's about being enablement Right self-enablement and Manila's no no different there It is an enablement of the end users and being able to provide that self-service and those that heard earlier sessions You know you empower those end users you allow them to take care of their business and not wait on it Processes for example, I can still limit and not create runaway Tenants for I can give them quotas You can define access rules form as well, you know the space quota number of snapshots number shares quotas things like that So you can really tie some things down to the tenants, too I can make the share types that we just spoke about be Public for everyone or maybe per private per tenant or maybe I can share them, you know for a couple of tenants things like that I Mentioned earlier, too about integration and your rest API integration if you already have an automation infrastructure Or some sense of configuration management of a structure you can work with the API's and integrate with the API's as well so I think there's lots of different advantages and why you bring share file services into You know the private cloud or being a service provider using open stack, you know one of the things too that I'll mention later You know, we're also working to bring Manila into distributions and granted right now We're only a what used to be an incubated type terminology and tag and that's changing but Manila will be a Released in Liberty, you know, that's the expectations. They'll get picked up by other vendors by like the Red Hats or the Susie's or the Morantis and the things like that, but we're already working with those guys so they can generate RPM so if you want to go to a You know an OSP or a Susie cloud, you know, you can have those RPMs for a kilo version of Manila to work with too, but you know, we've talked about different things We're getting ready to get in use cases talked about what Manila is and how you can use it But one of the things that I'd like to Thomas to share with us is he worked with quite a few customers in Europe And so he's got a nice perspective about, you know, you know, what what's different about Manila? Why is that different than object storage or why is that different block storage and you know have Thomas share with us? A little bit about what he's got there. Yeah, sure Greg so so one of our largest customers that I will talk about in the next slide actually came to us and They saw that they could use block storage obviously in in open stack They could also use object storage today in open stack and they had used both for certain applications But one they were missing one thing shared file service for that matter and the reason behind that is that Block storage, it's you know, it's very kind of like a low level obviously. It's very fast from a latency perspective Object storage on the other end. It's, you know, very high level. You can use metadata association But it has it lacks usually in terms of latency Both obviously have valued use cases and benefits but also some drawbacks and shared file services or shared file systems seem to be kind of like the middle ground where you have a you know, you actually trade a latency for Manageability and so this customer is actually Sorry, excuse me, so this customer is actually a dodger telecom Who have a use case for shared file services in the cloud and we did an evaluation of Manila with them in the last release so The key idea here is that they run Email service that they want to onboard potentially an open stack and the thing is that If you consider the subject and the mail body, that is something that customers Want to have access to in a very snappy manner? Whereas there's a read, you know some tolerance towards the attachments those can be fetched with a reduced latency So the key idea is to use shared file Services NFS NFS back ends for that matter to store the email the index and the Subject and the body of that mail to be quickly fetched if you are interested in this specific evaluation where we looked into the Applicability of Manila for shared file services this talk Was taking place yesterday and it's already on YouTube and you can check check it out and see how a customer from a customer endpoint Or view this they see that one of the other use cases would potentially be you know making use of the Snapshotting and creating a new share from snapshots capability to use that for parallel testing One In another session, I think two days ago here someone someone from the community Demonstrated how he was using Manila together with heat and trove to actually spun up new databases Which were backed by shared file services or NFS back ends? this is one of the conceptually you could use as for continuous integration in a very Beneficial manner if you consider tests which may be dependent on a certain database state What you usually traditionally end up with is that you run a test on a certain state of a database? That usually changes the database in a certain manner and then you you know once this test is successful You actually reset the database and then you run the next step next test So this is kind of like a sequential test and if you have a very large test suite this can take up You know can arbitrarily long so from minutes to maybe hours or whatever have you so what if we could now? create let's say 10 Similar shares of the database use trove to actually spun up 10 versions of the database Simultaneously and run all of our tests in parallel So this is something where you could potentially use Miller through the API benefits Combined that with Docker which would be a very lightweight. So this is one of the ideas where you could use this for continuous integration I will have Greg to talk about a few more use cases Thanks, Thomas. So There we go. Sorry So we've you know heard a little bit about what Thomas is seeing with customers in the use cases with Dutch telecom and continuous integration And that flows in well with some other use cases too where people would use continuous integration and in advance that and say well I need a NAS as a service We already know of a couple of people a couple of customers wanting to do NAS as a service One of them is actually using Manila. They were here earlier today to speak about that so think about if you as a developer or you as You know a deployer and say well, I need I need three or four file systems and the time it takes Traditionally, that's I'm not paraphrasing this for every customer But the time it takes to get that done versus being able to do it self-provision So you're able to do it yourself versus having you know waiting on the traditional it process So again, you know We heard earlier today about a customer that wanted to bring in that that whole it provisioning process for NFS into the customer You know be self-service be quicker to market being able to develop Software faster be able to bring applications to bear faster, you know quicker to market So I'm not going to read through all these but I think you could see some of the benefits of what we do today with OpenStack Which you can do with Cinder or Nova, you know, you know great I can give a Nova commute compute layer pretty quick to so when they get the VM now you can provide them with some shared file Services or a clustered file system as well just like you can with Cinder so Again, it's that flexibility that speed the market and providing those tenets and those project users You know the ability to secure that too. So again, we talked about neutron so you can have that Neutron subnet just for that tenant or you can have a storage network that everyone uses so lots of different opportunities for Improvement and how they can be self self-serve and self-management to so again the NAS as a service Used in many different ways as well as the complete life cycle of that share too. I create it I use it and I can destroy it. So it's the complete life cycle. They're just like anything else I think one of the things I'd like to point out too is that I keep hearing from some customers and people here is The integration with to do so there is that you know the Hadoop FS a lot of times we hear people that say what you know, I need Hadoop within the Within the OpenStack so you can also use that the Hadoop FS driver with Manila and support that as well So again extending those traditional IT service back into your private cloud or as a service provider You can build SLAs and provide NAS as a service too. We think that's you know added credibility again I mentioned it once already is that There's a lot of different in the community a lot of different companies coming together as the core team Those who used to be competitors are now Collaborators within Manila and we also obviously see that when AWS came out and said great You know, I can give you an elastic file service to where you know OpenStack's had it for almost two years another use case that is Is the integration with Manila? How many people have deployments or environments where maybe they write rest API integration into Cinder or Nova or Neutron? Anyone do any of the Python development work to integrate a Couple okay, so again We found just with the Manila API being very similar and if not almost exactly like Cinder that it makes it quite easier to Integrate with so in a couple of examples And I'll talk about a couple of things here that one's prototype and really just coming off the ground How many people are familiar with SAP and? Anybody use the SAP's LVM by chance the landscape virtualization manager? So that's a tool that allows one to build You know clone up and build workflows for SAP environments and those that know about SAP large-scale applications Databases file systems and so we've done a proof of concept with a prototype that says allows SAP's LVM tool to communicate to Manila and take in transfer You know the commands from SAP's LVM to say oh you need a share or you need to create a clone So it would create a snapshot create a share from snapshot and Manila's terminology and to present that back up to SAP So they can complete within workflows and integrate some of the SAP LVM workflows That's just one example. You know earlier today. We talked about the integration to net apps workflow automation And then I spoke earlier today. There's a group that's got a Home homegrown Python configuration management tool where they've they work with ironic So they'll talk to ironic for bare metal and they've done some API integration there And now they want to talk to Manila as well possibly so you know a lot of different Potentials but the point we're making here is it's got the API its rest API like anything else an open stack And it's it's actually quite easy to work with and then I guess finally Thomas spoke to it quite a bit but database as a service and it can be application as a service You know just think about if I want to be able to quickly spin up applications provide files shares and file systems You know being able to quickly provide that clone that so when I say clone that think about what Cinder does I can with Cinder and you saw in the demo I can do a Manila create share Manila create snapshot and then I can create a share from snapshot So think about the use cases are kind of more so outside the box than a lot of people are accustomed to and that I can Make a known point with valid data go through use cases hit a milestone Maybe something breaks. Maybe I've got to redo it So I'll come back to that known point with my Manila shares created from snapshot and things like that So I've I've preserved my data. I've done a test and then you can come back and use that data as well So that can be an application. It can be just file systems. It can be a database or it can be a large-scale application too So just you know another piece of things that you can do and really you know trying to give you the you know The art of the you know empowerment right be able to empower your end users to take care of a lot of work And not necessarily be dependent upon an IT group or you as a developer or you as a deployer Being more capable of you're providing different levels of SLAs for your customers to so these are some of the things that you know That we could talk all day about different use cases and what happens and how customers are using it But just like anything else in OpenStack and within the cloud. It's all about trying to Enablement be quicker and be more efficient too So then finally, you know a little summary about what we've done here today And hopefully you know you've got a little bit of what we've spoke of today You know the key a little bit about the key concepts of what Manila project was it's about you know shared file systems being able to allow self-management of those You know great I can do NFS and CFS protocol within a cloud now and hopefully that helps solve some of the problems that your business units or you as a You know an IT person needs to solve or a developer needs to solve and one of the things We didn't get into a whole lot, but it is multi-tenant. I can use neutron to provide You know tenet level networking I don't have to share that so if there's so security or isolation that you need you have that capability with Manila as well We talked about the use cases, you know dev ops Customers using it for the NAS as a service think about database as a service for applications or file system as a service Whatever the case may be and then share lifecycle management You heard Thomas talk a little bit about you know That the dev ops you in testing and coming back to a known point and a good point or restoring back to another point in time But I think one of the things that you know sharing with you here is also the benefits of Manila You know maybe for those that have the ability to say I need to do things quicker bigger and faster with my customers And your customers could be your business units those customers could be true customers that away from your company But empower those guys to take care You know being able to use file systems one of the things that I heard earlier was that if you heard Justin Deppman talk You know the you know the one point that he made out is that you know his business units are his customers And they were used to file systems. They knew how to mount it They knew how to have a persistent files system storage They were able to you know use that to their advantage and if the system crashed They didn't have to worry about recovery because in fast and sifts being a persistent file system So those are some of the things that we hope we kind of shared with you kind of stir the thought process We're not going to be able to give you the whole nine yards about what Manila is and in a short period of time But you know instead of death by PowerPoint We wanted to give you a demo to to kind of help see how it works and what it's about You know it's always good to be able to associate some true activity around it versus just hearing about it from PowerPoint And then finally some resources here so my understanding is that these slides will be available with as attendees here and You know you could probably say within 30 minutes the Video of this session is going to be posted on YouTube too as well But there are some within a net app perspective We've got some blogs around open stack and Manila One of the things if you wanted to get started with Manila is to use the deployment and operations guide So if you did went out and just did a Google, you know open stack deployment operations guide net app It'll tell you a little bit about our driver and how to configure that as well as sender and some other things too We do have a couple of technical reports You know ha deployments using that app storage OSP 5 on that app storage. We actually have a Suzy cloud reference architecture that just came out and Just two days ago We put out a technical report or a white paper around the use of Or that's it that Suzy cloud 5. I'm sorry, but two days ago We've have a a technical report about running SAP within open stack with Manila as well And with some of the integration I spoke of earlier, so You know a couple of the resources there We had a couple other sessions here so you can go to the YouTube channel search for Manila There'll be more than just ours here Thomas mentioned one earlier or someone did the you the trove the heat and Manila and be on the lookout for more Around that and maybe building app catalog around that just think about you know Some of the application catalog and container things that are going on now and building those type of application and service catalogs and Bring it together more and more projects to make things like that work. So it is the last day We do have a booth downstairs, but please take the moment and run by there if you need any other information or details And then finally do you guys have any questions? Yes, sir, please Yeah, if you could please could you stand up and use the mic it helps for the audio I just want to explore the snap the snapshot feature a little bit. I Know with the net app you you can lock the block the the blocks for the snapshot. Is that what actually happens when you do a snapshot? Yeah, so we're using that app as a back end Yeah, so if you have net up as a back end and you're using the net app driver it's going to leverage the net app snapshot technology as well as flex clone technology and then Just to take it one more step if I do a Manila create share and I use the net up back end It creates a net up flexible. So obviously when I do a snapshot It's a point you can be used for point-in-time and it also I'm usually worried about protection. So if I had replication on the back end with net app already Would it then automatically be replicating in my protection too? So right now, that's one of the extra specs that we want to implement Let me ask for Clinton is the snap mirror extra spec. That's not ready yet. Is it in Liberty? So in Liberty you'll have the ability to say here's an extra spec to define Data protection and so it would find a it would create a Flex ball or a share and then have a snap mirror relationship. So You know that's kind of where that is right now Any other questions? Very good. So, you know, thanks for your time I know you guys had lots of choices in the in the session catalog And we do appreciate y'all coming by and spending a few minutes with us to hear about Manila