BlueXP is now NetApp Console
Monitor and run hybrid cloud data services
So obviously intimate group um a lot of Googlers we do have customers and NetApp. Um so we just make this very interactive and ask questions as we go along. Certainly feel free to do that. We can take the conversation uh whichever direction uh we'd like. And then the other thing that uh we have uh towards the end is we do actually have some recorded demos. We were going to do live demos but yesterday we tried that and it didn't work. And so you know the old saying is two things happen when you do a demo a live demo and one of them is not good. So we uh so we experienced that yesterday. So we did the recorded demos uh that gun is going to run through at the end. But uh just quickly by way of introduction so Sean Darington the product management team at Google focused on our storage solutions and been working with NAP for the last uh 3 years on uh on our offerings and we've got some new announcements and exciting news to share uh that well we already shared on August 24th. uh Enuna here is the senior director PM on NetApp uh counterpart and then just by way of introductions for others we have other Google PMs here and um uh sales business ops and partner engineering and I won't ask the NetApp people to introduce themselves but uh just to kind of give you perspective about who's here um so how many people have heard of or are familiar with NetApp volumes so if you're not which is fine uh you can actually get started today with NetApp and Google Cloud even as and manage your storage on your own through a couple different ways through cloud backup service uh where you can have ontap uh replicate backups to uhthe star of the show. So uhyou couple different ways you can get started with NetApp on premises leveraging the cloud in a couple different ways. Onap is something that we've actually had in Google cloud since 2018. uh this is a self-managed uh fully functional APIdriven opportunity for you to leverage which you know on premises in the cloud you deploy it yourself on compute PD uh block storage and you can size it appropriately however we have also introduced net volumes this is now a Google managed service so you may be familiar with the cloud volume service which was something that we introduced in 2019 and it's a netapp managed service that was available through the marketplace um and very successful but what we've done is we've extended this relation ship now to have NetApp volumes be a Google first party fully managed offering. Uh and this is something that is um you can think of cloud volume service plus a lot more. So the same capabilities that you have been familiar with in terms of the protocols SMB NFS dual protocols also uh the opportunity of having a managed service as opposed to cloud volumes ontap or CBO which are the self-managed right but now the difference is obviously it's a Google managed service. There are additional things that we've added uh to this partnership uh and that you can now take advantage of particularly around what internally we call it Google horizontals. So billing you now get one bill for everything that you're using within Google. This is just another line item. Um Cec uh this is now integrated into our KMS. uh you can um uh leverage that as well as billing, auditing, cloud monitoring, those types of things are now part of this service that from an operational perspective, they're very beneficial. Uh and um if you think about isolating problems in your environment between slow performance for an application with cloud monitoring integration, you can see the entire Google cloud infrastructure from network to compute to storage now as opposed to having multiple dashboards and multiple inputs to try to understand that. Uh and depending upon your application needs uh you can choose the appropriate service uh for this. Uh so there are a couple different options. Premium and extreme are actually generally available now. Uh standard is beginning to roll out in a week uh in 14 regions. So you can see at the bottom different price points of 20, 29 and 39 cents a gig, but you can also see the corresponding performance levels that differ from that. Right? So standard starts at um at 16 and then extreme goes up to 128 megabytes per second per terabyte. And then both pre and extreme cap out at four and a half gig per second for throughput. >> So the regions of which they're located are the same regions. If you're familiar with CBSPO, it's those 14 regions. And then uh that's what we're rolling out for standard performance uh in next week. And then the other thing the other benefit that says CBS++ is that next quarter uh in Q1 we'll be rolling out the standard pre the standard service to all 38 Google cloud regions. Okay. So bless you. So you'll have this notion this capability to choose the appropriate service appropriate location. We also introduced storage pools uh as part of the netapp volumes release. This is a little bit different uh thinking where now you create a storage pool and that's what you actually get build at. So if you create a 100 terabyte storage pool, you're going to get a 100 terabyte bill for extreme. And you can create multiple storage pools and assign them to different performance levels. From there, you then decide to create your volumes as small as 100 gig. Um and uh then you can assign those volumes to a storage pool. And then that'show I mean that's how it works. You can also reassign and move volumes from one storage pool to another. Yes. So the uh it is kind of like um instances where you can create storage pools and it's there. You can also programmatically either through the UI or through the APIs um because this is another thing I didn't mention the API integration where you can automatically grow and shrink volumes within a storage pool. So if you're if you reach 90% utilization you can grow it above as long as you have enough capacity in the storage pool you can do that but then you can also uh dynamically grow that storage pool as needed. As far as discounts, um, for this is the other thing that with a single bill, if you're a Google customer and you have an EDP discount, that EDP discount applies to NetApp volumes. Okay? Uh, and that's a different way that it's worked in the past because cloud volume services through the marketplace, which is agreement and purchased from NetApp, but now that it's part of a Google service, you have that integrated account management uh, with that. uh and obviously the partnership is very strong given the netup relationship we have on the product management side but that's also on the engineering side and the S sur or the support side. So if you have an issue with support of NetApp volumes you just still continue to call Google and we handle everything behind the scenes. There's no ticket handoffs or anything like that uh anymore. Um the other thing that I'll mention is the backup service. Uh this is also coming uh this month. So I guess we're almost at the end of the month but uh in October here and this is something you're going to be able to do for uh any volumes. So as you think about different volumes and different performance levels for your application you can also think about the level of protection for those volumes. So for any given volume you can take snapshots the same snapshots you know with net space optimized immutable clone that for testdev workloads. Cross region backup allows you to have an async replication copy of the file system in two different regions uh for regional availability as well as backup. um backup. It works the same way as what you'd expect with uh cloud backup service or stat mirror where you can actually set a retention period uh and that's off in the behind the scenes. It's actually sent to a GCS bucket, a cloud storage bucket. You don't manage that again because it's a managed service. But now you have different levels of protection. Obviously, with a backup, you're going to have a longer recovery time objective than if you're doing cross region replication, having two file systems on each side uh asynchronously replicated. You can choose between a 10-minute RPO for replication, hour or daily. Uh, and there are corresponding price differences for those three RPOS. But you now have the option to again protect the volumes the way you want, provide the performance levels that you want for a given volume. Uh, little bit of an architecture here, architecture slide for what this looks like behind the scenes. Uh, for you in the console, you simply enable the API and begin to use it. Uh this is just more for background but the opportunity to deploy net volumes in a given project is something that you have the ability to do with PSA uh and then you just peer different projects um and you have onetoone peering for that you can't do transitive peering for projects so keep that in mind uh with your architecture but we do support VPC peering uh for that lots of the common use cases that you've everybody here is familiar with NetApp same use cases in terms of Windows file serving Linux file serving uh and also the dual protocol support, right? So, these are things particularly the dual protocol and SMB is something that Google did not have ahead of this relationship with NetApp um from 2018 to now NetApp volume uh in August. So, this is something that um we'll go through a couple of examples here for Windows uh actually Linux- based snapshots, right? Uh so, we'll have therecording for that. But this is something that if you're thinking about uh Windows file serving, there's an opportunity also to think about uh global file cache or edge cache. Now uh this is where we have a lot of customers moving Windows workloads, Windows file serving workloads to the cloud. And now it's a managed service that you can take advantage of. The global file cache is a little bit of a hybrid. So you still have to customers you'll still have to manage that individually, but it's an opportunity to move away from a capex filer model on premises. It's also an opportunity where a good example is um a large pharmaceutical company. They had uh 83 different locations throughout the world. uh a couple of data centers but then a lot of distributed locations and they had a whole bunch of seven mode filers they've been using anduh obviously that's reached end of life and it support as of last year but they wanted to continue to use netapp so they actually deployed that in Google cloud and now they have it deployed in six regions two in each major continent Europe America and APAC replicating across for backup and now they're not managing filers on premises but yet their users have that same Windows file ex experience like it's local wide area file locking protection, easier operations, etc. So, lots of opportunities to take advantage of uh of that workload. And then obviously also on the Linux side of things uh with NFS 41 uh andor dual protocol, you have that opportunity uh with Windows um uh we integrate with active directory as well as managed AD. So if you have AD on premises, you can uh integrate Net volumes with the on- premises AD. If you're deployed in Google Cloud as well as if you deploy managed AD or I should say consume manage AD from Google that integrates as well. If you want to do dual protocol you can do that and then integrate with um Kerros and then you can have both file systems access well both file systems access the exact same data.The other big uh use case that we see obviously there's you know the last few decades NetApp has had a little bit of success with VMware uh and we're leveraging that uh with Google Cloud VMware engine. Uh if you're not familiar with this,is a Google managed VMware environment. Uh so it's everything you know and like about vSphere in VMware, but it's managed by us. So you can easily lift and shift and just migrate applications from on premises into the cloud without necessarily having to refactor them. They don't run on Kubernetes or GCE comput instances. they can certainly run uh in VMware and you have a like for like uh you can also think about um hybrid models where you have a DR deployment where you have an on-remises VMware environment replicated to GCVE uh in the in Google cloud and that can be a failover scenario if you have if you're looking at data center exits or something like that. The use case here is not only for the ingest serving so like VDI and Windows file serving uh that type of thing but also NFS data store support is coming this quarter for this. So when you purchase Google Cloud VMware engine, it comes with block storage and you have uh number of vCPUs, memory and compute and as you scale the number of vCPUs based on the number of VMs you need, you get additional chunks of storage. You can't scale that storage independently. So if you have storage hungry VMs, you need to buy more vCPUs to get more block storage and that can become uneconomical. With NFS data store support, you can now have NetApp volumes be an NFS data store to augment the block storage data stores from VSAN. So you can have your tier one VMs running on the submillisecond block storage and have tier 2, tier three consume the NFS uh data storage. So it's a way to rightsize uh that environment for that. Um the other thing is uh I'm gonna I have these slides in here for snapshots and recovery more for you to take away because we do have a demo. I'm not going to go into a lot of details on this as well as this works exactly as you'd expect with uh all of your filer experience on premises. Um there are a couple things tonote here. Uh we do talk about ransomware because with uh netup snapshots they are immutable uh which is a good thing as well as regardless of the size of the volume because these are um space optimized you can actually recover any size volume in a matter of seconds. Okay. and you can roll back to a given prior point in time. If it's an individual file recovery or the entire volume, you can revert and recover that as I said in just a few seconds. Uh same thing with volume cloning. This is uh advantageous obviously for testdev workloads where you can take a clone of your production workload, do the test, throw it away when you're done, rinse and repeat with the next iteration as you go through that uh that cycle before uh deployments. The last thing um well a couplelast things before we get into the demos here but this is an example for the cross region replication um the 14 regions that we have today you can pick and choose what those 14 region the source and destinations are uh so you have flexibility there the other thing for backups this is what's going to be I mentioned in preview uh in the end of this month differentapproach different mentality but different price point as well uh so for thinking about how you want to protect uh your data you can again do that on a volume basis locally remotely. This is an example of storage pools and how you can think about assigning a storage pool to a given service level and then how you actually can create multiple volumes. But keep in mind as you provision storage pool, that's what you're actually build for. So if you're not using and consuming or assigning volumes to that storage pool, you're still paying for the storage pool. So leverage the APIs, leverage the infrastructure of the code. Uh Terraform will actually be coming this quarter as well. Uh so you have that opportunity again to just have more hands-off. So if you don't want to use a cloud console, you don't have to. So I know we're standing between you and beers uh after the session, but I did want to kind of give a brief overview. So we have two things. Uh we can do the demos, but I also wanted to introduce um I like going to introduce Anthony.So uh I happen to be a custom as [clears throat] I can tell you we deal a lot of the technology of GTP uh GCP is a hyperscaler choice uh we migrate our applications for corporate IT so for all of the backend application our financial like that's all a realist application is in front uh we utilize heat feed for those things um we also bronlyvaluable for valuable for um let that give us a lot of flexibility when it comes to that they can stay even for other hypers scalers from an acquisitionor whether we're thinking of coming off print data center and then moving it in. So the elite to have a great deal of flexibility about boot type of workloads also to vary the performance. This flexibility allowed us todo a lot of things very quickly. We standardized our data products across our infrastructure both one and many state.So this becomes extremely valuable for us to do that. actively right now taking applications from DC from an acquisition that we name and one of the challenges we think is a lot of this stuff and then dated operating systems and very dated hardware. So we had first thing going to need a lot of changes in order to get into format where we could actually move the threshold. So having the flexibility in the cloud volume as well as on tap and other solution gave us the ability to not only manage this data very easily but also because of the backup of the came recovery opportunity that you get we able to ditch a lot of the bills that we ran into in the legacy and guard so it became very valuable. >> Yeah great thank you so much Anthony. Yeah, like Anthony Lloyd has been a strong supporter for Net Poliums. Thank you, Anthony. Like one last question, couple of questions. You know, we want to have chat with Anthony in the morning session. So,how do you see like now that we are uh the first party service, how do you plan to use it differently or especially your teams different use cases?>> Yeah. So for us one of the things we're really trying to be with you robot acquisition one of the challenges you have is many cases you don't necessarily keep facilities and so for us one of our role is to get off as much as possible except in our key data center will be though we don't tend to but so being able to have cloud-based solutions give us tremendous flexibility where we're closing sites we get the data out and get up but the other thing that we're really trying to do is also move from the traditional little cap that's a desperate model 3 and you're sinking money into a bunch of hardware at the end of the life cycle hardware you got to figure out how you dispose of it pay to get rid of it which the problem then you got to deal with all the aspects that how do you impact your car footstep but you got a zero emission issue at 2035 I'm got big fluid data center they rather be dealing with so we're trying not only those data showing now but we're also trying to find ways to move fromappex conversation to opex conversations. So three things that we guarantee a tech ref we fresh we reduce our carbon footprint diminishing the father credits forthat is leasing model but we also take the operational support and hurricanes and that workload and do it to our fership which mean we don't have to do things we don't have to bash and you know that so it makes it muchsimpler from an operational course perspective now that the pricing model has become tetative you know per gigabyte basis is it really is a teasy conversation to have when we're looking at saying, "Okay, do I go and buy, you know, at le 7 7100, you know, or do I go and get me cloud-based solution andnow I got more flexibility. I don't have to worry about sitting on prim and I don't have to deal with all of those things.>> One last question before you we a couple of product guys, you know, my PMs are there from the Google team as well. So, what is like one feature that you want us to do soon? >> [laughter] >> [laughter] >> [laughter] >> We're gonna get it again. We're gonna have to do it again. I'm kidding. So, we made it through that. That's the first feature. [laughter] feature. [laughter] feature. [laughter] >> I get the other features. Obviously, you know, now you got AI and everybody talking AI because at Felus was, you know, was cloud three years ago, then it was digital, now it's AI. But you know implementing secured governand ethical control of AI so that if something happen to your environment my environment doesn't get compromised cuz see that's that for me is number one my data has to be protected my environment has to be protected so those things are foundational and I know everybody's talking about yeah we're putting AI we're put that's right but if you can't show me how you going to protect my environment without.>> Thank you, Anthony. Thanks a lot. Thanks so Thank you again. >> Yeah. >> Yeah. >> Yeah. >> You know, you know what? I I'll take some. My wife will take them foryour team, too. >> Yeah. You get an extra bag for that. [laughter] Very cool. So uh you know Mr. Anthony Loy like you would have seen has been part of our launch collateral. Thank you. So for some of you haven't seen uh it in action. Um as Sean was mentioning we actually tried to do live demo. So what happened is there's a little bit of a screen resolution issues. You know all everything ends with what you're going to say and what you're going to show. So I want to show you um a volume creation. Um you know some of you like you know Kevin clinging and all have been there with forever but I'll kind of correlate to um how you do onrem versus cloud. So the most important thing is especially for those of you like who have been super familiar with the onap API um I'm showing the through the UI the same thing can be exercised through G-Cloud SDK or REST API. So when you're operating at scale you know I'm pretty sure Anthony would be the first one to say you're not going to use ur you know you're going to do like you know uh everything is automated through either like know whatever playbooks etc right so I'm just showing you like you know for the for those uninitiated like but there is a massive amount of automation possible through net app volumes so the key things you know I want to show you is um you know as kind of Sean alluded to uh it's all about you know you create a volume uh it's self-service service the way that you can enable self-service is through AM the identity access management right so identu authentication like you can configure roles we have some default roles you can create custom roles and every time you create a volume you know it actually goes to a pool you know what uh Sean mentioned pool again those of you have been um in our untap lens pool is not an onap aggregate pool is just a logical entity right it's not an aggregate you don't need to think about oh my aggregate size is this I need to cram my volumes you don't need to worry about it like you know we actually take care of all those things but so pool kind of gives you still a logical space to think about how many applications you're putting in what pool especially if you're serving your end user teams you can think about how to charge back them etc you know whatever your use case may be but you don't need to worry about capacity management of volume mapping to pools that's very words. But no, fantastic question. Uhthere is actually thequestion is arepools uh just a container of sorts or does it have any performance implications? So when Sean showed three service levels, you know, premium, extreme and uhstandard they're all tiled to the pool. So what you're going to see the way we think about is like you know uh you may agree like you actually have different types of workloads. Some are capacity sensitive, someare performance sensitive, right? Or cost sensitive I should say. So based on that, you know, you're going to put like the appropriate service level. So you're going to create a pool for a service level and you can create multiple volumes of that service level. But I'm actually going to show you the later part of the demo where you can take a pool from one service level to the take a volume from one service level or one pool to the other pool when you really need the performance bump for the short duration. Thanks for the question. So what we're really showing is like know youyou're uh you got to create some pools. Uh they know you're let me see if I can stop uh pause. So I already sh stop at the pools where I think that maybe what caught your attention um you know how you can create multiple pools and the key thing I want to show you like you know uh again going back to your onap expertise u these things are built onQoS policies for those of us know who know like inside out these are like on provides a lot of QS policies so you can just create uh you know multiple QS policies um and add pools to I mean a pool is a kind of an abstraction on top of it. Yeah, this is where I think I wanted to show you like if you see here you know you create like you know pools across multiple regions and there are different types premium and extreme right. So extreme is 128 megabytes per second per terabyte of capacity. Premium is 64 and the standard uh which we are ging is 16 megabytes per second that also obviously have like the cost like you know appropriately adjusted. So you can choose like know one or more volumes per pool. You know there is there a lot of flexibility for you. Now when you create a uh when you create a pool or when you create a volume now obviously you're going to look at the share name that you're going to pass it to your application nodes and you're going to specify a capacity. The minimum capacity is 100 GB and the max capacity is actually like know 100 terabytes. Right now again goes back to the onap limit. you know the max volume you can create and very soon you know we're working to introduce like a pabyte volume especially for the EDA workloads so you're going to see some new functionality you know just to give like a directionally speaking and that's pretty much it you know now uh you're going to get the volume you know you can create either a protocol type NFS or SMB um because it's on tap uh every share we can make enable dual protocol both NFS and SMB or SMB3 we support both NFS v3 and v4.1 you know uh everything that you know how to use um and there can be multiple pools and you take uhnow if you create it like know it's going to give you an IP address right like you know uh in this case it's actually aqualified domainname you can take it and mount it to uh a Windows application the best part is like everything you're doing here uh it can be everything can be self-service you don't need to be in path between your user and you I mean your infrastructure like that's what you expect from a cloud service you know there is nobody calling I mean someof the customers like you know uh big customers put an app on top and create maybe a self-service on top you know if they want to because everything I'm showing can be exercised through API rightand it's thewhole thing is fully integrated into the Google IA there is no net app proprietary thing here and onceyou mount a volume, you know, especially for SMB3, you also can use your Microsoft uh ADLDAP integration. You may have groups and know what you let your users to do. It could be the we support an on-prem Microsoft uh active directory or you can also use Google cloud managed ad. We support both. I mean I'm just like you know telling you like you know it'sa directory trust me like I'm going to create a file [laughter]it'slike you know it's like it's a straightforward exercise right you know pretty very simple to do and on top of this now we're going to show uh some I'm not able to go fast but if I can oh I'm making this mistake How do I do that? Okay. Sorry for that. Yeah. means like it's a federated AD. So theway we work with is you provide the AD right like you know and uh your IDP could be anything uh who is the IDP that you're integrating to but that will be the responsibility of the ADso we're not going to federate you know you can federate yourself from the back and that's that right >> yeahI actually you know if you see here uh the active directory policy it may be too small so we actually create you can create active directory policies there. So at the time you tell like know who is your what is your ad server and you process it. Yeah, great question. You know, I love like when people are super familiar with onap and deal with us. The this is very interesting question. You know, you as a user of GCNV don't need to worry about SVMs.This is very important concept like when Mr. Anthony Lloyd was telling like you know you don't need to worry about patch management. You don't need to worry about SVMs as well because you're not managing on tap. You're basically getting a flex wall. Right? I'm speaking on language. You get the flex wall. You don't need to worry about aggregate. You don't need to worry about SVM. We take care of it on behalf of you. Um you know Sean actually showed thearchitecture slide. If you recolct there's something called a tenant project. There's a client project on the left side and a light blue the tenant project. The project actually maps to an SVM on the onapp. So that's how we introduce because a multi-enant service. So every one of you get your own SVM but you don't need to manage SVM right. So in the back end basically what's happening is because we know your project youraccount etc. We tie uh the active directory policy to the SVM we created on behalf of you right that gives you like a lot of flexibility now this is like the thing it's because it's SVM per cluster so you can actually go across multiple ontop nodes we manage the capacity for you right so you don't need to worry about SVM and similarly uh Sean also showed uh thecross region replication you don't need to peer to on top clusters and you don't need to peer SVMs on those uh onap clusters we do whenever you set up thecross region replication, right? You know, uh basically we kind of in a way you have to unlearn some of the onap on what do you call uh um uhthecore units of management. We're going to do it, but you're going to get the same functionality though. The data more is still the same old snap error that you're going to use in the behind. Okay, I think almost done. I'm going to show you like you know u you get u now this is where the adlap would come into picture like you know who accesses what you know you can create groups pretty much everything the same okay so now I'm going to show snapshot restore I think this is just a two-minute demo by the whole thing is only seven minutes other than my plates so what I'm actually showing here is uhthis is a flex wall you know you're going to create a snapshot u so that's what we're doing you're going to we already have bunch of snapshots I mean you guys are all like on tap experts so this looks so trivial to you but for those uninitiated we are showing you like bunch of pictures let's say you know you happen to delete it like know the manual errors. So these are these snapshots are local right now. advantage is you know if you want to make a human error you know you really want to oops go back you're going to restore from the local snapshot what we're doing Sean mentioned as part of the presentation right we are enabling backup so the snapshots now you can actually store it on Google cloud storage right you don't need to configure a bucket you don't need to manage the access controls we're going to do everything on behalf of you right so in essence you know you can keep the backup snapshots for three years you know However, longer duration that you may choose that is another protection for you. For some applications you want to use cross region replication but as Anthony was telling you know that could be cost prohibitive for like all the volumes you know for mission critical applications you may want to do but backup you know it comes at a lower cost it gives you the same longevity you know and the thing is the snapshots everything is incremental so you can actually clone a volume from the backup snapshot right yeah so same thing you know here you have you know uh hardly snapshots you can revert it shows back you know you're going to get your files back right the if you observe right something like you know the storage admin is not doing anything like everything is uh can be done through an application owner based on the policies that you let in your organization yes you can yeah because it's on the window side Okay. Uh this is actually a pretty cool demo. So this addresses uh the question that he was asking. Uh how can you u you know scale your performance right? You know first of all you create a volume um here I'm going to show I think this volume is created with the premium if you see here the premium pool right. Um it so performance is attribute of uhtwo variables. One thing is uh the service level. The other thing is the total capacity. So for example, you know, it's thevolume is about 1 TB. So it's basically clocking about 64 megabytes per second if you got it. So now say for some duration, you know, you want to increase the performance. I mean increase the capacity to two terabytes. Now you'll actually it's going to take a while. But the key thing is now you can actually see like you know performance linearly scaling to 128. All right. The key thing is like know the application didn't stop it is nondisruptive right because you're not really mounting this where like know the on top QS policies are basically in action this what you would do on prem this exactly what you're going to offer through our customers but now you say oh you know that's not sufficient for me you really want to double up because that's like you know say you're creating reports under the month you can now go to change the pool from premium you can actually change it to extreme right now you can see like you know linear scaling. So you're here it goes up and this almost done. So it's wrapping up. It can actually go up to uhdouble the so the performance scaling is extremely linear and uh you know it's very simple. Exactly.Yeah. It's the same flex wall going uh using you know different onap QS policies. I think yeah that's the my last slide any questions yeah I think I if your question is like you know is there a buffer cache that's happening ontap controller you know are you getting the benefit so there are two things you're going to observe right as part of theservice level that we deliver for each of theperformance we deliver for each of the service levels irrespect of what whether there is a buffer cache or we fetching the directly from the drive or the aggregate we're going to throttle it or we're going to allow you to scale to the performance that your capacity and service level allows you. So it's not like you know in the case of your like you know on-prem on top you know you get like know whatever you have because you're greedy you know because your application is greedy in the case of a cloud service it's a multi-tenant service you're not the only volume right you know we have like much like lot of volumes so we basically use the thrott >> yeah no go ahead yeah I was say so we have a >> across all services so like cloud storage or persistent disc or hybrid disk or netup volumes or file store we have an SLA associated with and that SLA is associated with the performance of that. So it's we don't wor it's a big difference between thinking about NetApp filers on premises and thinking about NetApp in the cloud and that you don't have to worry about the HDD to cache ratio or how much uh you know flex cash you have in there. So it's a different way ofconsuming it.removes some of your bells and whistles if you will, but it's delivering a different service for you to consume, right? And so thecash out ratios if you get a benefit of one great but there's no we don't disclose who gets it and who doesn't because asgoing said it's a multi-tenant architecture that you just can'tthe key idea you're not managing the on system I think Android is the that's where for the most customer that's actually good read you don't need to worry about the complexity so in other words you don't get to see it however at the volume level you get enough information through cloud logging and cloud monitoring. So when you give an application you see how the application uh performing uh then you have other demos now we can refer you to uh offline and similarly uh I think one of you are asking whe also any action that's performed on the volume is actually logging in a test so you want to get everything into cloud log you can see a particular user doing a part action they could be changing service level they could be changing the capacity they would added another uh meth policy right because that's what you expect you know for some plans we for monet can you keep my data secure so me as a provider will keep my data secure but you know you also have responsibility on who that's what like you know so you want to see the auditing something what to happen so you have like audit reports and this is the standard set of Google operating principles and it's fully integratedright you don't need to worry about on tab logs anymore you basically look at the cloud lots and it comes at the cloud All right. Well, great. Well, great. Well, great. Well, thank you everybody forattending and uh enjoy I guess happy hour now and then therest of the show tomorrow. So, thank you [applause]
Learn about Google Cloud NetApp Volumes, a fully Google-managed file-storage service built for enterprise applications and how you can instantly recover from a variety of outages, local, regional, and even ransomware.