Welcome to the Episode 320, part of the continuing series called “Behind the Scenes of the NetApp Tech ONTAP Podcast.”
This week, Keith Aasen from ONTAP Product Management (keith.aasen@netapp.com) joins us to deliver an overview of the latest ONTAP release, ONTAP 9.10.1.
For more information:
- Automatic Ransomware Detection blog
- NVMe/TCP blog
- Behind the Scenes Episode 312: NFS over RDMA – NetApp ONTAP 9.10.1
- Behind the Scenes Episode 301: NVMe over TCP in ONTAP 9.10.1
- Behind the Scenes Episode 311: Introducing the NetApp AFF A900
Tech ONTAP Community
We also now have a presence on the NetApp Communities page. You can subscribe there to get emails when we have new episodes.
Finding the Podcast
You can find this week’s episode here:
You can also find the Tech ONTAP Podcast on:
I also recently got asked how to leverage RSS for the podcast. You can do that here:
http://feeds.soundcloud.com/users/soundcloud:users:164421460/sounds.rss
Transcription
The following transcript was generated using Google’s voice to text transcription service. As it is AI generated, YMMV.
I’m here in the basement of my house. And with me today, I have Keith Aasen to talk to us all about ONTAP 9. 10.1. I can’t believe. I haven’t already done. And ONTAP 9. 10.1 podcast. I’ve got a couple of the features Bap. Ami will be better about nine, ten eleven or sorry. Nine, eleven one coming up in the future. But Keith. What do you do here at NetApp? And how do I reach you? I first off a Super Bowl party here. I’ve always wanted to be on the podcast of always listening, been a fan. So I’m pretty excited to be here. I have been at NetApp for a while, a kicking around fourteen years, this spring, but most of that in the field. I was a Solutions architect and took forever. And then just this last fall moved over into the product management team. So I’m actually on the ONTAP, customer product marketing team or product management team leader specifically in customer Communications. So I do a ton of e b, cs represent the ONTAP vision and help customers in the field. Better understand what we’re doing and ONTAP product management. All right wage.
00:59 – 01:59
0
en-us
0.90
Cool, how do I reach you? You can kind of old school. Email is probably the best. That’s easy one. It’s Keith at NetApp, Let me tell you a little secret about getting out of the podcast Keith cuz you’ve been on this a long time ago. So first email me second, you say I have an idea for a podcast and third I say, okay, that’s how hard it is to get those guys. So anyone listening if you’ve, if you’ve been yearning to be on this podcast, if you have subject-matter expertise in something, whether you’re a customer or a field or a partner, feel free to reach out to Age baat cast. NetApp. Com, and I will answer your emails and we’ll schedule something up cuz I’m always looking for Content. I don’t like missing weeks. Sometimes I miss weeks because we don’t have content. So, yes, you could have been on this years ago, Keith. Well. Now, you’re gonna have a tough time. Getting rid of me. That’s good. Any time in fact, you will be doing this on a regular Cadence now, since you are the ONTAP guy. So, be ready to do this again in about a month or two, or three or whatever.
01:59 – 02:59
0
en-us
0.90
Cadence’s, speaking of the Cadence, ONTAP is famous in recent years, for, for switching at the cadences to the point where it’s kind of confusing for people say, you know, where we are, what release were on, why we change the cadences, why we change the release numbers. So today, where are we with the Cadence? And I know we’ve been there for a while but let’s just kind of level set their wage. I know it was changed it a couple of times and now the numbering system is all kind of funky cuz we’re back to three numbers or three digits and we seem to be skipping versions. It is a little confusing. What we do, actually we’re maintaining a 4 month lease per year, Cadence? If you can believe that for releases a year, it’s a it’s a maelstrom of software development. But as far as for most of the listeners here, we do 2 on Prime releases a year and to Cloud own new leases. And so if you’re wondering what happened to the. 0 releases, those are the cloud releases and then the. Ones are on Prime releases and they happen twice a year spring and fall and Ed.
02:59 – 03:59
0
en-us
0.88
The cloud ones are our winter and summer. So it’s sort of seasonal, the season release. Yeah, it’s interesting. Like the off Cadence really or not off. Cadence like the off releases where it’s basically just Cloud. Only backup going to be necessarily feature releases, but they are very important to the development of ONTAP because of the push to use more cloud services. Moving more to Cloud volume services, such as NVMe over TCP, and that sort of thing. So, those are very crucial to those particular aspects of the business. Yeah. Absolutely. And you in your new role, maybe can say more than I can. But, you know, people think that, oh, we just got ONTAP to run a cloud and we’re sort of, you know, done and dusted. But it is so much more than that, you know, there’s obviously the optimization for the hyper scalar infrastructure, but then you know where the real value starts to come in this type of ONTAP directly into the various Services, you know, the security services and directory services and different work flow mechanisms inside those Cloud, hyper scalars, you know, there’s dead.
03:59 – 04:44
0
en-us
0.84
Ton of of integration points to be had in those. And so that’s where those to you mostly development for those those zero releases. Go to. Yeah and the cloud releases if you’re not familiar with the ones that are waged not managed by you, right? So basically you have Cloud volumes ONTAP, which is like you can log in and you know to the CLI and administer yourself. There’s FSx for Amazon. You can do the same thing, but then you have things like a n f or you have TCP where it’s basically just as a service model and there’s a entirely different gooey. There’s entirely different functionality. There’s things that you can actually access yourself because they want to simplify the interaction as much as much as possible and just make it make it a basically a volume provisioning service.
04:47 – 05:45
0
en-us
0.84
Yeah, and the fact there’s actually, you know, there’s one of those three flavors now in the cloud, right? There’s those ones that are strictly as a service where you can access certain things and, and there’s customizations. We found those receive e o, which you sort of managed as as a, your own device with a customer managed, but even they’re, they’re getting tons of Integrations to be done. Unless I say, with the Direct Services and security services. And then there’s the, you know, there’s kind of the ones in the middle like, like the FSX ONTAP that’s available where it’s, you know, it’s a mix, right? AWS manager, some of it. And we managed something that the customer manager. So, some of it and so, again, that’s a different interface and different set of APIs. And so, there’s tons of customization to be done there as well. Yeah, I think FSx, has that sweet spot where most people want it to be down. There are people out there that just don’t want to think about it in the ANF things, serve a purpose. But there are a lot of customizations that you can do with FSx that you can’t do it, see as a service pieces, but I digress, you know, them.
05:45 – 06:45
0
en-us
0.89
Talk here to talk about today is ONTAP 9. 10. 1 and what’s new in there and then we can kind of tie it into like how it fits into Cloud. I know one feature that comes with this is not yet available to a CVS services or CVS offerings. And that’s the automatic ran somewhere. But I think it would be very valuable in those use cases. So let’s talk about that new feature and ONTAP night at 10 to 1 a.m. Sounds good. Before we start, I will highlight this business, the perk of you delaying. This 910 one did just pass. Yeah, the P1 release came at last week. So not only we talking about Iraq. It’s available, but it’s already in P1. So it’s there in ready. It’s on the truck ready to be used, but the autonomous ransomware protection. I think I agree. I think that is the coolest feature. We’ve added in in nine, ten 1 and I have a bit of analogy for describing. It shows kind of dive into that. Your audience is probably Matrix fans. I think we’re probably SVM.
06:45 – 07:45
0
en-us
0.86
Common thing is it not? I think it’s pretty much, you know, pop cultural or that we can talk about. People won’t be like too taken aback by it. So, so if you think about the, the early t v, a trips, we all had those screensavers of the Matrix running, right? The code rained on the screen, think about what ONTAP is doing behind the scenes when it looks at a filesystem, right? So if you have a file share out there or filesystem up there and users are, you know, connecting to it and they’re updating files and creating new files. That’s almost like that kind of rain of random, you know, code that’s coming down and that off Matrix screensaver. And that’s what ONTAP kind of watches. And it can do it, you know, very lightweight. It doesn’t take a whole bunch of CPU Cycles to do that. And anything about normal humans, interacting with the brake system. That’s going to be pretty random. I owe ONTAP also looks at something called file entropy, and that’s kind of looking at how the files themselves are actually being changed. And again, that should be for the most part time job.
07:45 – 08:28
0
en-us
0.85
Random that with what human interaction. But when something like a ransomware attack occurs, you know, that that’s a level of automation that sparked up. Now, it’s coming in through a user account. But it’s, you know, it’s an automated process that fiber up and, and, and it begins to change the files in in a consistent pattern. No different ransomware has changed them in different ways and, and it changes it in in, you know, different measures. But those are all done in a pattern and what ONTAP able to do now is detect this pattern, this change in the maitre. See an alert on it. And so, so the first thing it does is it triggers a snapshot to occur to give you a choice of recovery Point. Very close to when the attack actually occurred. Then obviously trip off some alerts to notify that the tax actually happened.
08:30 – 09:29
0
en-us
0.77
Yeah, it’s like one of those things where you know, you need to have a multi-layered approach to ransomware. And this one of those layers. It’s a way to detect and let you know that something’s going on. Something weird is happening in your database that you may not have otherwise detected yourself. Yeah, this this wouldn’t be the only, you know, it’s not you can go. Oh, yeah, open the floodgates and you know, click on whatever link you want. We’re protected wage by all means, you know, the downside is this this reactionary, right? Obviously, the better way is to prevent the attack from happening in the first place. And so it doesn’t preclude any of those tools that do, you know, scanning a general training of your of your staff. To not, not click suspicious links or open suspicious files, but this a great safety net. This an airbag, right? It doesn’t mean that, you know, if you put an airbag in as mean we get to I’ll drive like idiots. It just means if something bad happens, you know, hopefully things are a lot safer. Yeah, I mean ONTAP doesn’t have a lot of control over what happens before it gets too ONTAP home.
09:29 – 10:29
0
en-us
0.90
We don’t have any sort of like safety mechanisms that we can put in front of your network. So that’s kind of up to the administrators. But you know, we want to be there in case something does happen. Exactly. And what I, what I think is really cool about this, It’s I tell people we’ve trained ONTAP to look for rent somewhere and that’s really the best way of saying it is. We’ve already done a lot of the heavy computer intensive stuff to track ONTAP. So now just looks for those particular patterns and that means it can be done in real time. And with very little overhead on the on the ONTAP system. Mind you is not the only way to look for rent somewhere, you know, off of our, our other offerings in the portfolio is cloud secure, and it uses some of the capabilities and ONTAP to look for rent, somewhere in a very different way. Can you kind of talk about that a little bit? The difference is between, how you know, Cloud secure operates with versus like the automatic transfer detection with ONTAP. Yeah. Yeah. Sure. So ONTAP partners with classic, you’re into Cloud slime.
10:29 – 11:29
0
en-us
0.89
Is a space offering and so it’s, you know, it’s in the cloud and the perk there is. It can use a bunch more compute in the cloud and so, ONTAP can pass a bunch of data. And one of the things we’ve enhanced also in ninth one is false system analytics. I’ll talk about how, you know, we’re using that natively as well in a bit. But essentially we pass a bunch of information about the file system up to 2,000 secure and then because there’s a lot more compute going on or we have access to a lot more compute and it’s not in line. Secure can do a lot more analysis of that data and they actually take the approach of looking at the user behavior. And so they start to build up a pattern rather than looking at what’s happening on the file system. They look at, hey, what’s this user normally do, which shares they normally access where they, you know, what do they normally do from a read-write a pattern? And again when a ransomware attack occurs, it comes in as a user but it’ll radically change how that users accessing their files system. And so they’re they can detect. Hey this song
11:29 – 11:54
0
en-us
0.87
This person is suddenly, you know, doing a ton of Rights into a particular file share or they’re accessing a, a bunch of file shares. They normally never do and again we can, you know, it reacts and a person actually calls back to ONTAP and go. Hey, something funny is going on here. Trigger a snapshot give us that very close recovery point, but it also can take it things, a step further and option. Do things like actually even disconnect the user from shares to prevent a damage from occurring.
11:56 – 12:55
0
en-us
0.75
And it’s one of those features that I imagine. We are eventually going to keep bolting on new things too. And we’re going to evolve and make it, you know, overall better, you know have more functionality wage. Is that kind of a correct assumption? Yeah. Absolutely. It’s going to be it’s going to be ongoing every one of these releases is going to have new patterns, right? As as the rent Summers, learn new ways of attacking the systems that we’re gonna have to keep up with that, you know, a good example of it is, you know, historically always think about these things encrypting. The entire file, which is, you know, as a very slow process and we’ve gotten pretty Savvy about how about cash that. Now they’re getting pretty tricky around. Hey, let’s encrypt the first little bit of a file, right? We can do it really fast and it’s still enough that I really, you know, works the fiber that makes it really hard for people to access it, but they can do it a lot faster and those are harder to detect but not for ONTAP. We’ve, we’ve caught that pattern. Just you’re right. It’s going to be an ongoing thing. Whereas Transformers, find new techniques of birth.
12:55 – 13:55
0
en-us
0.90
Of damaging files ONTAP is going to have to be trained in each of those detect just as fast. Yeah, and you know, ransomware is always evolving and that’s going to be part of it as well. So you kind of gotta keep up with that sort of thing. So, you know, as far as the automatic rent somewhere future goes. I understand. It’s not something that’s built into ONTAP. You have to kind of unlock it with a license bundle. Is that, is that accurate? It is, it is choice of the security and compliance bundle. So that kind of makes sense. And so it’s enabled in there and then you do need to turn it on for, you know, given volumes. And there’s a bit of a warm-up. The world. Is basically for you to detect. Are you getting any false positive? Maybe going to file share that for whatever reason has, you know, some strange file entropy. Maybe it’s just it’s you know an application that’s interacting with the files something that might be triggering a bunch of false positives. We give you kind of a warm up. We’re ONTAP, you know, starts looking for that pattern, you can get ideas. It is at the end too sensitive or
13:55 – 14:18
0
en-us
0.71
Or not sensitive enough but gives you a chance to cap test out the analytics before you actually kind of commit it onto a volume. And then and then once you do with the kind of protected, you’ll get an alert if it detects on something and you’ll find these kind of special secure snapshots that have occurred when those actually trigger. Excellent.
14:19 – 14:46
0
en-us
0.89
So, speaking of secure snapshots, we also have another feature that’s been around for a while in this called SnapLock. And essentially, that’s a compliance replication feature that we have and, and understand that 9:10 to 9. 10:00. 1 p.m. With some enhancements to that as well. So, can you talk a little bit about the enhancements to SnapLock? Yeah, for sure. I know what you just, I always love when, when we build something for one, use case, and that over the years we took is it’s really good for something else. That’s happened on a number of occasions, right?
14:46 – 15:46
0
en-us
0.90
Yeah, absolutely. I mean, there’s been several things that, you know, like, I’ll give you an example. So, you know, this something that I always talk about, but FlexCache, right? So FlexCache Siz initially. Hey, we need a larger volume that can handle a workloads. And that was the initial use case, that wasn’t the end used case. It was just, this what it was initially designed for. Now. It’s evolved into many use cases, many different uses and eventually, we’ll be down the volume and ONTAP. But, you know, right now it’s, it’s still got some Niche. Use cases as well as a more, General use case. But, yeah, we have a lot of features that can do that. The one that I always think of is in the virtualization days engineering. Came up with a cheat to clone out virtual machines for testbeds. Right? And I think, one of the T Amy’s actually saw them do this, and I was like, whoa, what is that? Has like a, oh, we just do that for our tests. Like, would you guys want to use that for something like heck? Yeah, right. That became the, the rapid cloning utility which was, which was taking advantage, and ultimately found its way into Vai to do, you know, fast, DM, clones, but dead.
15:46 – 16:46
0
en-us
0.90
SnapLock is like that to me initially designed that with the use case of, you know, customers that have they’re in high compliance Industries where they had a need that, you know, I’ve got a set of files and I need to, you know, walk particular set of files for five years or ten years or thirty years. That that I have to prove that file is not only thing, you know, not changed but hasn’t been deleted or can’t be deleted for that particular window. And that’s kind of our SnapLock, you know, lived in its Niche and you know, we sold some of it and, you know, setting some customers used it but it was pretty stable. How, how many customers actually used it? But that’s, that’s evolved as. Hey. This an incredible way of protecting your data from ransomware, from compromise, administrator accounts, from Rogue administrators, all of things, which are on the rise for, for whatever reason. And so it’s evolved outside of just that compliance window for for basically, you know, almost any customer. Most customers want this level extra level of protection of birth.
16:46 – 17:46
0
en-us
0.90
I, you know, you know, I want to have the zero trust models, even my own and means I want to make sure that my own and men’s could do something damaging to, to my own internal files and maybe I don’t need to lock my data for for thirty years. I just want to protect it for for 30 days and have a copy of data is protected. Now, the rubble SnapLock used to be that you had to have a dedicated Aggregate. And the reason for that is we, we don’t trust things like the system place. Where it’s like, well, what if I just Advanced the system clock for, like twenty years or change? My ntp source. It’s like, well, no, we thought about that long to get that easily. So, you know, we used to have to build a dedicated aggregate where where the we would actually use Hardware on the drives themselves to keep track of time so that you know, short of destroying the truck if there’s no way of tricking the system into know thinking the time has passed when it hadn’t. But again with the rise of demand around SnapLock, we realize what it’s actually no harm. In fact.
17:46 – 18:08
0
en-us
0.88
Doing that for an aggregate. So it’s so in nine, ten 1. When you upgrade to 9:10, one, essentially we make automatically every aggregate a SnapLock aggregate. So there’s no longer a need for a dedicated, disk pool to SnapLock. You suck SnapLock volume anywhere in your existing environment. Again, it does require an extra license. And again that’s in the security and compliance bundle. So. So, you know, that’s a good good upsell for those to that particular bundle.
18:10 – 18:17
0
en-us
0.91
Yeah, this the ability to separate, you know, away from the Aggregates and make it more of a mobile FlexCache.
18:47 – 19:45
0
en-us
0.87
Absolutely, and the level of security is that this has zero trust model that used to be kind of exclusively in the networking space is finding its way into storage data and infrastructure. As it should write it is, hey, you love your storage admins, but, you know, lets you know, maybe it’s just a compromise set of credentials or maybe this actually you a rogue and but you won’t have a way to protect that data in such a way that you’re not even urine. Man. Not even NetApp, support could actually delete that data or remove that particular data from the system of caution is. We also in nine 10, 11 of those minor features. I know if you know Justin, we extend it. We had customers to saying, you know, thirty years not enough. We need to protect that data for 100 years so you can now actually took that snap law to lock data for 100 years, which is kind of mind-blowing. I can imagine this for calls. I accidentally locked my data for 100 years off.
19:45 – 20:45
0
en-us
0.90
That creative salesperson. Good news is you’ll be dead by the time it’s unlocked. It doesn’t matter. Exactly. Exactly. All right, so, you know, speaking of the data protection piece. We also have SnapMirror for S3 now, and ONTAP night at 10. 1. So talk to me a little bit about how SnapMirror for S3 Works. How is it different than normal SnapMirror for like an a source and volume? I’m sure. Yeah. Well, I mean, first off, you know, highlight, we are not announced S3 support for ONTAP in 98. So, so exactly a year. And, and over that course of the year. We, you know, kind of understood where are we really good in S3? Like, we were a customer, use ONTAP, S3 versus a storage group, in particular ONTAP, S3 fast. It’s real fast again. ONTAP, does it really does suck, you know, we’re not storing things as files or objects or lungs, right? We’re restoring with unlawful. And so the front and protocol doesn’t really matter a whole lot to us as far as performance is concerned dead.
20:45 – 21:45
0
en-us
0.90
So our object protocol is very similar performance-wise to what we do on on file systems, which which is normal for a file. Server is pretty blindingly fast for four objects a bridge. So that’s really been ONTAP. Niche is you know, as more applications want to use S3 as their primary, you know mechanism. There’s an increased demand for a high-performance S3 and that’s ONTAP shining, but that doesn’t preclude the need for protecting data. And so, so one of the wrinkles we had prior to 9:10 one is, is it was the date of a sort of trapped on that particular month, the ONTAP cluster. It just lived on that one. So you want to have multiple copies of your datas anything. So we, you know, we introduced, you know, S3 SnapMirror here. Now, what’s, what’s different is a rather than kind of having in that exclusive ONTAP to ONTAP replication, which ONTAP to ONTAP replication, is, you know, first off the most flexible replication out there because we do have ONTAP everywhere, you have a dog.
21:45 – 22:13
0
en-us
0.88
Wild. We have it on premium, all kinds of devices and software addition and VMware. But there’s also a desire to replicate natively to non NetApp. And so with S3 is not a S3. We can replicate natively to S3 targets things like storage Grid or even S3, compliant object stores like AWS S3 and have those off as a direct replication .44 ONTAP to give you that secondary copy.
22:15 – 22:52
0
en-us
0.90
Yeah, and that’s another one of those features that I think was one of those building blocks, right? We used to have the, or we started out with the SnapMirror to solid fire or SnapMirror to the, you know, now-defunct Alta V. But the ability to SnapMirror to other devices became interesting for customers and, and the field. So it got extended out to other things and now S3 because S3 is gaining traction for application developers and other in other use cases. So, you need to be able to have the ability to move to an object storage model. If you if you want to and it’s, you know, it can be used for data protection or it can be used for silly. Migrating to that model.
22:54 – 23:53
0
en-us
0.90
Not. That’s the that’s the trick there. It’s not just like native ONTAP. Snap mirror where you would leave it in a form that only ONTAP could understand. This this SnapMirror, but actually we can have we D waffle eyes it. So test script, the waffle lay-off, and we put it in there. And so when you replicate to those other object, tears, those objects are still usable just as they as they were. So that’s they really kind of super cool part of it is that we’re not kind of putting it in their proprietary format. We’re actually replicating it to these objects doors, but leaving the objects as is so. So if it’s off you put in, you know, ONTAP is an object. You can actually read from those alternate datastores. Yeah, and it’s one of the things you’d hear from customers. When you tell them something about something like fabric pools, like, oh, we can tear to S3 and they all can access those objects afterwards. I know you can’t, they’re still tied to ONTAP, but that basically was reading between the lines saying, oh, customers actually want the ability to turn these things into objects, so they can use it later.
23:55 – 24:47
0
en-us
0.88
Yeah, and you know, there are some caveats, right? So, so that’s, you know, let’s make sure we, we sure that is things. Like, when, if you do go ONTAP to ONTAP kick, the bucket name, preserve, right? So you can do a failover. If you lost the primary ONTAP cluster, you could, you know, do a failover and the bucket name reappears. And so your application would just suddenly find a storage again. We when we do replication to two other object platforms, obviously those ones have to be rewritable and so we can’t have the same bucket name presented out in the in the ecosystem twice. And so the failover isn’t quite simple. The objects are fully usable in the data still useable. But it just, you know, the kind of caveat. If you’re not going ONTAP to ONTAP you wouldn’t need to redirect your application or users to it to a different bucket name, but after that, then all the data is totally usable for proof, whatever you want to do with it.
24:48 – 24:56
0
en-us
0.88
So with this replication, is it strictly bucket? S3, bucket, S3, bucket, or can we do like, an ass S3 SnapMirror?
24:58 – 25:57
0
en-us
0.89
Today eight way to break the brain today. It’s S3 to S3. But it does. It does think things will start to get interesting when off again without saying too much. What if you start blurring the lines between those two protocols? What if you did it? Yes, what if what if users here without getting a book myself in a future call to discuss? Just yes, I talk about potential feature additions and in future podcasts here about ONTAP. So you’re moving on to the next thing that we want to talk about here, you know. Next Colts game of migration and replication. So SVM. Migrate was also added and one interesting about that is, you know, it’s not necessarily a fully-fledged feature. I think it’s like when preview mode or something but it’s also got some specific use cases that people might not be really familiar with. So talk to us about SVM, migrate is what the status of it is and what it can do and what it can do.
25:58 – 26:57
0
en-us
0.89
Sure. Sure. So you must be to go up the stack, right? We just what SnapMirror and replication and moving data from A to B is super valuable. But we’ve we’ve had this concept of a storage virtual machine. That is there’s so much more when you’re thinking about Naz and object, protocols in particular and so SVM, Dr. Was red picked up the entire storage virtual machine that had the name space and security settings and, and, and permissions, all of that was in there. So, when I poured it over, I could fail over to another cluster somewhere and, and all of that would be recovered, but it was an outage, right? I had to had to walk, you know, download the SVM on the primary site and bring it up on the s w. A cut over and users would disconnect and reconnect. But what we’ve done with SVM, my grade or data Mobility were referring to us is really the ability to transparently pick up a storage virtual machine natural language. We kind of make sense and move it to a different cluster within the same ruthless in the same data center wage.
26:58 – 27:58
0
en-us
0.85
Now initially, can I go? Well, that’s kind of cool. I guess it’s like a vmotion but an entire storage Works machine lever level. Right might have hundreds of volumes sitting under a big-name space to say it, but move it on disruptively. But why would I want to do that? Well, a couple of major use cases is one is, you know, we have a lot of service. Provider customers, and customers that operate private clouds. You don’t always know how your life is going to grow. And now, obviously ONTAP can grow, horizontally. I can add additional notes. The Clusters, if I have additional capacity requirements. I can certainly grow grow my cluster, but there may be, we just providers to have a, you know, a real pre-built recipe around how they deploy things. When they, you know, they deploy 8note, clusters or deployed twelve know clusters and that’s our that’s our building block. And when we need more capacity, we took another 12 note cluster. And so this a way for them as as they have these different clusters in the environment. If one of them starts to get hot, they have maybe have a tenant that comes in and grows faster than they expected. They can reload bath.
27:58 – 28:58
0
en-us
0.75
These these multiple clusters, the other use case, that that’s come up. In particular, is, is people from an availability standpoint. So when they, when they do grow a cluster, or when they do a hardware refresh, we can do those on disruptively, so we can do that without any sort of interruption. But there is risk in there are sending to get kind of in the behind Iraq and start, you know, pulling cables and swapping controllers, should be really careful about how you do that. To make sure you do things in the right order and you don’t end up, you know, causing an outage, which is sort of the last thing you want to do. So there’s, there’s an appetite to say, I want to build a completely brand new cluster and, and burn that Hardware in, you know, get it in there and performance test it and really, really stress it out. And then when I’m ready non-disruptive, we moved my workloads selectively over to it, you know, one at a time, rather than a big bang move. And so, again, this month, I am, I great, you know, lets us do that. Looking out a little bit, you know, there’s also appetite for maybe moving it from all your own all flash cluster to a hybrid cluster or a from a pre prod cluster into wage.
28:58 – 29:28
0
en-us
0.81
But and then, you know, the, the big super interesting areas is Justin when we get into Cloud, right? It’s not there yet. But shortly we’ll have that same functionality in to see video and that should be interesting. Okay, so it kind of, you know, tells us a little bit more about what the SVM migrate feature functionality would do for us. It also you know, is another one of those features where we’re going to add more to it and it’s become more readily available and people have more use cases for it. It my understanding is it uses that SnapMirror synchronous back and forth at the right?
29:30 – 30:16
0
en-us
0.89
Actually, I don’t I assume so it does become a guest. Give that synchronous replication. And that does do you become an automation trick to you know, seamlessly took it down or basically down the SVM on one cluster and bring it up on the other in such a matter that NFS clients don’t don’t disconnect and you know, they just seamlessly begin. Using I owe on the Samsung 9 plus turd. Yeah. Well that part of, that is the synchronous part of it is the file handles being transferred similar to what to do with SVM Dr. So it’s interesting that we take a lot of these features we develop and then we, we repurpose them. We received them right? Don’t waste, not want not, you’ve got this feature. Don’t don’t create new stuff. If you if it already Works, use the stuff you have that’s proven and works and then repurpose it for something else.
30:17 – 30:52
0
en-us
0.89
And Dearborn. Yeah, we started off initially. We’re doing, you know, AFF to AFF in the same data center, but that’s you said that’s the starting point, you know that used case and support Matrix is going to grow rapidly version version. Yeah, there’s really not a lot of technical reason for most of this stuff not working. A lot of it has to do with qualification has to do with limiting the footprint, right? You don’t want to create a bunch of problems that you might not have seen. You have very specific test case that you’ve done. Let’s start small and then just start building on to it later. But ultimately that near the direction that you’re going, is have it readily available for most everybody.
30:53 – 31:52
0
en-us
0.90
That’s it. Exactly. And if we’re promising you that it’s going to be a non-destructive move. We better off tested it, you know every which way to Sunday to make sure that is not disruptive. But previously, this linking all the facts together. I just talked about how things like SnapLock we can lock data for 400 years that sort of like to go. How am I going to? How am I going to maintain a, you know, a system for a hundred years, it’ll age out for sure. But this that. This that mechanism ready. You can, you can move these volumes from, from ONTAP, clustered ONTAP cluster to kind of cascade it for that long. Wait engine. I’m looking forward to the, you know, time and we have SnapMirror. It’s in our heads. Holy month ago FlexCache Netflix cash to the head and a hundred years. The Nano. The Nanobots in our head are going to have the datand we can just access it instantly. So, you know, that that would require a lot of speed. Of course, in to do that. I mean, we’ve added some things to increase the speed for Access of data. One of those things is nvme over TCP.
31:53 – 32:53
0
en-us
0.90
So talk to us a little bit about that. We’ve done a podcast on an in-depth, but just kind of give us the high-level overview with that is. Yeah, I level is you were the first to Market with this capability nvme TCP wage. Um, if you’ve already done the podcast, your nvme is the replacement for scuzzy which, you know, sketchy was kind of developed in the sixties has tons of limitations to it. When you think about what modern Christian Media can do from an IO standpoint and destruction set. It’s incredibly hampered by by Skizzy commands, right? Man? Is just, it was designed for spinning platter. And so even today the, you know, the most cutting-edge all flash, you know, you know, built for it for all Flash and tuned for all flash. If you’re doing sketchy commands, you’re still essentially, you know, bringing your IO commands back to a place in sector, right? You’re sure you’re still, it’s still thinking about it in terms of spinning media, which, which is not optimal, and it does things like cueing and, you know, as if you had spinning media, so when you, when you cut over to nvme E,
32:53 – 33:52
0
en-us
0.90
You all that goes away. Right? It’s nvme is really a new, a new Rio protocol design, you know, to support the next generation of media, right? Solid solid state drives, the wrinkles always been a. Is it, typically required all new networking, right. New cars and switches. And so the beauty of nvme TCP IP is the fact that you get the advantage of the improved protocol, but you can do it across your existing TCP networking. There’s no, there’s no requirement. It’s also a means to do high performance. I owe in Cloud environments that uses standard TCP networking. So super attractive for workloads that are you moving between on Prime and the cloud and also, you know, a very viable and popular alternative to iSCSI for for high-performance workloads. Yeah. It’s basically, you know, way to be more flexible in a way to get this type of thing into the cloud cuz of course the cloud operates on TCP, you’re not going to have a lot of access to block storage in the back end. So have having a way to speed up those things over TCP.
33:53 – 33:58
0
en-us
0.91
Important for the flexibility, as well as the interaction with things that maybe you don’t have control over.
34:00 – 34:15
0
en-us
0.81
Exactly exactly. Easy way of getting into that particular space and yeah, improving your performance. So we’re kind of leading the curve a bit on there. It’s just a matter of Now application support wage, big one. Being vmware’s, is it can’t right around the corner. So we’re really excited for that.
34:17 – 35:17
0
en-us
0.89
So another performance-enhancing feature that we added, I think is also. I don’t know if it’s preview mode or not. This NFS / RDMA. So talk to me a little bit about that and where that is and why we decided that yeah, that one isn’t in in Tech preview. The big that was fun cuz it’s super high demand in certain use cases and the biggest use case at the moment is doing analytics using gpus and so for for customers that are doing that heavy-hitting workload. They need a way of getting very low latency, a high amounts of data package off of the off of the system into the Jeep, you trying to feed these arrays of gpus. I was talking to a customer yesterday, that that has very large ONTAP clusters and they say, they typically have between eight thousand, twelve thousand GPS. That, that kind of drive it, which, you know, very impressive stuff. Anyway, the trick is to get the data there to keep those gpus wage.
35:16 – 36:11
0
en-us
0.90
Fed, and so and that typical AI automation is all done based on NFS. So it has to know, it can’t be done on a low latency block. It has to be done overnight. So in FSA, / RDMA is really a way to provide that, you know, high-bandwidth low-latency. I owe from the controllers to the gpus to, you know, the keep them fed during the analytics page is it’s in tech review because that’s a pretty narrow use case at the moment. But again, as that demand increases and as we do more testing, watch to have that sort of expand as we go. So, I know that with high-performance things, you want to have a pulse on, what’s going on in your system. And part of that is something we added in ONTAP, 9. 10:00. One with the filesystems analytics. So we had now have the ability to do more with that. So talk to me a little bit about what we’ve added an FSA, as well as the new stuff and System Manager.
36:12 – 37:11
0
en-us
0.90
For sure. So we first introduced FSAnd 98 just to sort of have a dashboard and what’s actually going in and on, in my file shares and you seem to be sort of a ONTAP whisper if you try to understand, you know, why is it particular to file share, you know, being problematic from Ayo or workloads to end point. But a nine ten one, it boils It Down super easy with in service manager giving you a dashboard to see On Any Given file share. You know, who am I? Who am I hottest files? Which files have the most, I owe to them, which have the most throughput which directory is driving that truck and that’s a great way of identifying like what’s this guy doing? You know, what, why is, why is this particular file, you know, doing so much? I owe. And we’ve also linked it top users. And so maybe I’ve kind of like a bad song. The network, who is just doing, you know, some sort of crazy processing on his final share. They shouldn’t be that’s causing, you know, causing this bioload happening on the system. So it’s a great way of identifying rather than just kind of throwing your hands up and going dead.
37:12 – 38:11
0
en-us
0.78
And this volume is hot. I have no idea why this fighting is hot. Now at a glance we can go. This volume is hot because it’s this directory and this file and this user that that’s hitting that system, really, really hard. And maybe some problem or maybe it’s not maybe you hit him with a qos to slow them down or maybe you kind of, you know, discussion with them and say hey that’s not the purpose of network storage. Right man. You want to be doing some of that on the local ask. Anyway, it’s a huge enhancement and it’s also part of the data that we pass up to back to our first discussion, which has Cloud secure. And how Cloud secure does love the user analytics. Yeah. We’ve also added some of the functionality back in the system manager that people are asking for. I know that we had a things like local, Unix, user and group creation. Being able to manage Aggregates like you were used to it made them before. I think that we went from went to tears at some point. Now, we’re back to more aggregate level stuff. I don’t know the exact specifics of how they do that. But are there any other, you know, top-of-mind things that they brought?
38:12 – 39:12
0
en-us
0.90
Fact that people were really asking for, for sure. U re we, we are first time that we probably over rotated when we, when we rewrote System Manager to use reg apis. And again, that was that was the reason for the rewrite and, and the idea was, you know, the, the simplification and obviously, over rotated heavily on that. So, we’re bringing a lot of that functionality back from being able to choose your name off of things cuz a big one. Right. Is how like I want my list named? A particular thing. I want my I want my my share are my my my volumes named a particular way. So being able to name things, as you create them, was there was a black one. You mentioned. You mentioned Aggregates our local tears that actually isn’t quite you kind of pre disclosed that a little but there’s Justin sneak peek. But yes, we absolutely are bringing back the ability to configure. When I add capacity when I add drives the system. I can configure that how I want to, but that’s not 9:10. That’s a future version of ONTAP but not too distant future for that, but we are getting. Yep.
39:12 – 40:11
0
en-us
0.89
Close to, to, you know, we definitely hurt all of the customer requests around. Hey, you guys took out something that I, I used heavily were expecting and, you know, kind of 20 20, 22 years ago, with the year two-thousand. Functionality back or all that functionality back plus which is, which is exciting. Excellent. So, you know, last but not least we have new functionality for FlexCache right off. Yeah. Yeah. So FlexCache years is, is, you know, something that’s been gaining popularity and a part of that is back to the cloud story, right? The ability to have your primary data live, maybe in a Data Center and then have all these cash versions out around the world using FlexCache and some of our customers in particular, our mediand entertainment. And there there’s scenarios where I actually, I don’t want remote users accessing a file when I’m in the process of changing it. So, so good. Being a good example is you maybe I have a media file. A video file that I want to do a rendering pass.
40:11 – 41:11
0
en-us
0.89
On. Well, I don’t want users using the old copy of that until I’m done, the rendering pass. So I actually have the ability to do a 9:10 one file locking where I can. I can lock all copies of a file off while I’m you updating the source and then and then and then re allow that. So users. Don’t you try to fetch the file? Can I when it’s mid rendering pass right when it’s when it’s what they’re going to get is not operational so that file locking, you know, although a bit of a specific use case, but for customers that have caches around the world and users working around the world on that file system is it’s a important capability to have to, you know, do that Global file lock. And then release it when you’re, when you’re done modifying that file. Yeah, this really four use cases where, you know, you need that read lock. It’s not, it’s not going to pay for everybody’s use case, but it definitely was something that was being asked for. Absolutely exactly the ones that needed. It really needed it, but it’s, it is like, he’s had a bit of an edge use case. Yeah, absolutely. All right, so, you know, I think that off.
41:11 – 42:11
0
en-us
0.87
Covers that the main major features there are some like hidden gems in there that are either in dying priv or things that are currently in the works. Right? But yeah. Sookie, anything else? You can think of that? We might have missed. This actually popular live, which is really, which is cool. You tried. But System Manager, the other thing we’ve brought in the system manager, is autonomous firmware updates. So, assistant manager, NetApp will actually, when it does, it’s call home for calling into active IQ. If there’s a new version of your disk shelf, or management process, your firmware will automatically pull that down and ask you how you, how you want to deal with that. Do you want to update it now, or do you want to schedule a time to update that firmware? And what do you want to do that in the future? You want to meet all? Always use that time slot, or August, just automatically upgrade disk and she’ll from Wells. One of those things that’s super irritating to do but it’s really critical for the stability and security of the system. So I think that’s a big win. The fact that all automatically pulls it down basically out of clicked.
42:11 – 42:55
0
en-us
0.89
And schedule it to upgrade in a change window and then just you know log back in to get notifications and it automatically applies it. So look looking to do that to a lot more things and that that’s really going to make, you know, kind of ownership experience of an ONTAP system better. But yeah, that was the other one. In nineteen one. I should I should have talked about that as well. But that’s right. Yeah, we did. Yeah, that that’s one of those things where you’re probably not even get a notice it because it’s not assertive exactly until it pops up and goes. Hey, it’s all right. Hosting ready to do it. And if you’re not comfortable with that, I think there’s a way to turn that off. Right? You certainly can. Yeah, if you want to just simply do it on your schedule and you want mayonnaise, you can certainly turn off disable it. Yeah, it’s not mandatory, but I think it’s a real nice to have it just skips that whole hassle of having to download something. Yep. Absolutely.
42:56 – 43:56
0
en-us
0.91
All right. So, you know, that’s pretty much it for 9:10. One. Now we are going to have another one soon for the next release and spring so, you know, stay tuned for that and the only, the only information I can give you about that release, is that it’s going to be 9 to 11. 1 and that is actually very close. In fact, we have some beta copies in Early Access program. So if you’re a if you’re a customer or partner that is interested in doing something that early testing and again it gets a little Beyond just taking the tires. There’s an actual test plan to run through but if you want to be one of those early early validators reach out, we can we can you know, see if you’re fit for the program and tell you what’s involved with it. But yeah, we’re always looking for for new customers and partners to be part of that Early Access program. All right, excellent song. I’m Keith. If we wanted to reach you, how do we do that? Keith and NetApp. Com? All right. Excellent. Thanks so much for joining us and talking to us all about the 9. 10.1 release. And again, we’ll have you back.
43:56 – 43:59
0
en-us
0.84
About the next release. Sounds good
Why Is The Internet Broken? 
Pingback: A Year in Review: 2022 Highlights | Why Is The Internet Broken?