00:11
Perfect. So we're seeing the first the attendees coming in. Uh, good morning, everybody who can already hear us. Um, we will start the webinar shortly. We're just waiting for people to get in to solve all the zoom and audio connection problems you might, uh, encounter during these kind of,
00:33
uh, web meetings. Now we're getting started. It was still early in the process. So, um, Ron Eugenio, um, how are you today? Very well, Marcus. Yes. Looking forward to this.
01:12
It's gonna be a fun morning. Yeah, very well. Oh, that. That's great to hear. Um, as you can see, um, maybe even spot by the names we have, Uh, a good portion of, uh, I represented and, uh, me as a guy from Germany run
01:30
from the UK and, uh, Eugenia, uh, from Italy. So it will be fun. Also, um, to see about the topic. Well, we have potential, uh, differences in our audience. And maybe, um, what we see in the field, um, with customers, what the trends are and, um, have good discussions on that.
01:54
Yeah, we'd love to see some, uh, some chats in the chat box if you've got any questions. Uh, please, uh, engage with us. Let's make it interactive today. Yeah, So, as Ron said, um, a a few things about logistics. And we have a Q and a panel, which you always can use to, uh, get your questions in. And the people not talking at the moment will,
02:18
uh, take a look at this and also answer the questions. We will do this either in life or, um, written, um, in the Q and a panel. Um, don't be shy. You can also click on the anonymous box check box so your name will not get recorded with that and not show up for other participants. If you don't like that, um, So, um, feel free. And as Ron said,
02:44
keep this interactive and yeah, with that, I think, uh, we should start, um, the coffee break event right now. Um, so the the theme of today is all or nothing. And is this the right approach for especially data in the cloud? Um, so I would like to welcome everybody to this, um,
03:10
a coffee break. As, um, some participants know this is a serious we are doing. So, um, we're having these kind of events from time to time tackling different, uh, projects, having discussions. And, uh, what we usually do is to get some experts for that topics in. So, um, you can get different opinions as well, And, um,
03:35
well, share the expertise across. Um, my name is Marcus Gall. I'm a principal technologist in the region here at pure storage. And I am your host today. Um, first of all, well, you, for those of you who have registered and, um, are not excluded from that, um, you have you should have,
04:01
um you should receive a gift voucher in your email inbox. Um, it's a €10 coffee gift card because it's a coffee break. Um, you can use it for coffee, but the voucher is also better for other things. So if you don't like coffee, um, feel free to get something else. Look out for that email address. Um, And if it,
04:21
um, goes into your spam folder? Well, um, just take a look at this folder, um, during the next week as well. So, with me, my experts today are Ron Eakins. Um, Ron, would you like to do a quick introduction of yourself, please? Thank you. Marcus. Yes, and thank you for having us today.
04:39
Yeah. So, uh, Ron Ekins, uh, I'm the ME a practise leader for databases. Uh, I specialise in Oracle, Uh, been using oracle technology for for more years than us, and I would like to think about probably since the early nineties. Uh, I am a, you know, thrilled to share that. Uh, I've been given the award
04:59
of director again for 23 and 24. Uh, so I'm one of only 70 ace directors globally, uh, of a, uh, an A community of about 600 whenever I present or, uh you know, or do a blog. I. I like to share my code at my github repository. And I have, uh, one of the the top oracle blogs, Uh, globally.
05:22
So it's It's always in the top 50 at dot com. So if you've got any, uh, Oracle questions, you feel free to, uh, DM me with Twitter or X, uh, and check out my blog. Uh, there's my QR code there. Or my my profile as well. If you want to learn a bit more about me, thank you. Thanks. But, um, I think you're doing a bit more than
05:42
just the database piece, right? Yeah. So I Yeah, I mean, I, I My whole career has been about data management. Uh, so, uh, you know, I, I My background is all close to governance, data security, uh, working in highly regulated environments, but, uh, yeah, that's, uh So next week, I'm at a cloud talking about that same sort of things.
06:03
Uh, interesting, because, well, we have the theme of today. How shall I put all my data into the cloud, or shall I keep it on Prem or shall well, or is it better to do something in between? Um and that? That's, I think, a challenge. Everybody's Yeah. For me.
06:19
It's always a data discussion. Thanks, Ron. And, uh, Eugenia, uh, welcome to, uh, this coffee break. I see a nice picture of you. It's probably You went snowboarding that day. Yeah. Thanks, Marcus.
06:34
Yeah, well, I'm a fan of a split boarding. Not not actually. Uh, snowboarding, so sort of a much more hardest way to do that, But, uh, anyhow, I'm a field solution architect with pure, uh, since seven years. So far, so roughly at the same time as Ron and, um I visited out of Italy, and before joining PR, I was working in different,
06:58
uh uh, areas, uh, and in in mainly in the telcos and in the PR M, um, sector as well. Here at pure I'm covering all of the, uh what pertains to cloud and virtualization. And, um uh, in addition to that, I'm, uh, uh one of the maintainers of our code for automation. Mainly the, uh, answerable playbooks.
07:25
And, uh uh, the the AL Also, the open stack integration that we have with the and, uh, um drivers. Well, great. That's a Well, actually, that's what you're covering. Um, but it's interesting to see, um, that evolving from virtualization platforms to cloud to automation,
07:48
Uh, and obviously things like, um, you you listed here? It's an open metrics exporter, because, um, also, the monitoring and observ ability, um, is becoming more and more important, especially across these environments, right? Yeah. Indeed. This is a thing that, uh, in this case, we we cover broadly with with the with the with the,
08:10
uh, our equipment. What the, uh the the the physical and the appliances. They are fully covered by our, uh, um, observ ability. Uh, uh, Matrix Exporter. So, uh, we integrate with that, uh, with, uh, both open telemetry and, uh, open matrix. So we are offering, uh,
08:35
process the the the the the the greatest, uh, chances for including, uh, storage arrays both physical and virtual, into a fully, uh uh, observably mode modern observ model. So we we will hear more about, uh, virtual storage arrays in in a in a few minutes. Uh, probably because, um if we're talking virtual arrays and that means probably public
09:04
cloud, because in the public cloud, um, well, there is no notion of physical arrays. Typically, Um, so the the theme is of all, um of all in into the cloud and what we'd like from our audience is to know about. Well, does your company or do you have an all in cloud strategy And that's really defined by, um, there is a directive of the company to say,
09:37
by date X, we want to get rid of our data centres and move everything into the cloud. Because, quite frankly, sometimes when when I'm at events and, um, have talks to especially people from the the digital office or from a, um from for CD OS.
09:59
Often the question comes up. OK, um why is there still a need for data centres? Isn't everybody moving into the cloud? On the other hand, the whole storage market for on Prem storage is about 60 billion of size, so there must be something in between. So, look, the question is do we have a fall in class strategy?
10:25
Um, and the way we've picked some answers, some possible answers here say, Hey, we're already there, Um or we have plans to do that, or um, no, we don't like the card. We try to avoid it, uh, where it's possible or well, the the mid the the myth of this is basically well, we will choose the services which will run on Prem and we choose.
10:50
We choose between the services which are better suited for a car. So, Ron Jenny, what do you think will be the majority, you know? Do you want to go first? Yeah. Yeah, well, the, um, strategy here could be probably much more. Uh uh uh, in the direction of having a mix,
11:16
Uh, not in, like the hybrid cloud, but the the the M, the multi cloud and, uh, the the The strategy that in general I've seen has been followed by the vast majority of our customer is instead probably going, uh, fast for one direction and that, uh uh, move back to the high, the high, or if you will, uh, the intermediate multi cloud model afterward.
11:44
What I'm seeing right now, what I'm registering right now is indeed this is second stage in which, uh, whenever things have been migrating probably too fast to the to, uh, trying to, uh, steer from the all on Prem into the, uh, uh, fully public cloud approach. Uh, I've see, I'm seeing that the the the the, uh, most,
12:06
uh, many other factors came in into into the into the, um into the pitch and and the, uh, things like, for instance, that were not, uh, uh taken in account at the beginning, Like the costs, the hidden costs and the even the hidden TCOTCO cost that arise at the, uh, during the move.
12:27
Uh, they are, uh, forcing, reconsidering the strategy and moving back to a much more better, uh, multi cloud approach in which things that runs better in, uh, on prem are are are still kept there. And, uh uh, possibly the cloud is then used it as an expansion of the internal, Uh, the the internal infrastructure.
12:52
That's interesting is, you know, so yeah, I. I think what we'll see is for companies which are being born in the cloud. They're already there. But for the vast majority of our customers, it's a bit of everything. It is, uh, using, uh, a where it's appropriate. It's using platform as a service where it's appropriate, but still, uh, the majority of their data,
13:16
it's hiding on premises. Yeah, we can see that, uh, in the results of our audience as well, Pretty, pretty well reflected that, um, many of you have chosen. OK, um, we might have a all in cloud strategy or 100% on Prem strategy, but, uh, since the world evolved and the also the offerings in the cloud evolved,
13:40
um, we now have a more of a strategy to put to maybe start services in the public cloud and then, uh, repatriate or, um, just pick very cautious where to put which service because, um, if you do it wrong, it's basically a huge cost burden. Um, to change that right if you, for example. Um, if you already have 100 petabytes of data storage in the cloud, it may become cost prohibitive to extract that 100
14:12
petabytes back to an on prem, uh, deployment as an example. It's an extreme example. But, um, not everybody is dropbox, uh, and and can can do that because this is exactly what they did, right? They moved from a public cloud to on prem IT because it's better for them. It's not.
14:32
I'm not saying cheaper in general, but, um, what they said is it's better from a reliability perspective from an operational efficiency. And, of course, it's cost in the end. So, um well, first is that that's a interesting, uh, thesis run. I think, Uh, you you did that slide,
14:53
but yes, but also, um, if you if you read through analysts, um, I think it hits the nail. Yeah. I mean, the the message here is for many organisation, you know, it was everyone is doing AWS everyone's doing as your you know, we need to do that. Uh, so we want a cloud first strategy,
15:18
and we want to be there by a certain date. I think What we're seeing now is more of a measured approach. Uh, you know, and that that's why the term cloud smart comes in where you look at your data and it's all about the data, uh, a and associated services that you use to manipulate that data. So we want to put the right workload in the
15:38
right cloud with the right data all at the right time. And there were different ways of delivering that. Whether that be on premises cloud adjacent public hybrid, multi cloud, lots of versions. And, uh, you know, most people, even if they believe they are all on premises,
15:55
they probably have still some S services. So we do see a a very much a a mixed blend of all of those. Yeah, so you know, and that's the the comments at the bottom. You know, Cloud Smart involves making the best use of cloud concepts. So what we want to do is take the best of the cloud ways of working,
16:13
whether that be automation, uh, monitoring and observ. Maybe even the, uh, the the the models for for payment and and billing and try and adopt those on premises. So we have a constant way of consistent way of working from a technology but also from a business process. Point of view. So that's that's what Cloud smart is.
16:35
And it sounds a bit like, um, uh, the Golden Nugget here is to, um, have a an operation or operational and financial model, which you can span across all of the cloud. Whether it's public, whether it's hybrid, whether it's SAS even right, yeah, so very much a consumption model.
16:58
Yeah, indeed. It's exactly that it is. What the next step consists of is just to have an intermediate layer that hides all the complexity of running the infrastructure like and like, they were actually in the cloud. So it's not just the, uh, uh, IT approach. It is the, uh, the the whole If you if you will,
17:22
uh, management layer, which is, uh uh then which has to be, uh, used in a cloudlike access model. I mean, we've tried for years or companies tried for years to build their own automation models or whatsoever and and call it cloud. Right. Um, but so far, we I think nobody really, um,
17:49
took off in in the end, right? Because it's a very it's It's a very complex thing to solve. And, um, my thesis is there will be not a product solving the problem. It's more, um, the model you apply, how you're operating these kind of infrastructures, which, uh, will make it in the end.
18:12
Yeah, I think it's it's cultural as well as, uh, technical. Mm. So it's it's more of a mindset than you say than a product. Yeah. And you said data is is is key where to put Put it on, whether on Prem or, uh in the public cloud.
18:32
So why is it so important, uh, to take a look at data because everybody's well, the first thought is, OK, I have these 180 services available where I can do fancy things with in in the public. And why is it so important to make a conscious choice? Where to put the data on? Because nowadays we're talking about server.
18:59
This we're talking about, um, stateless and all that, uh, fancy things, but somewhere there is data. Why is it so important to make a conscious choice across, uh, Europe? Uh, you know, we we very much have to consider data solvency. Uh, data ownership, data,
19:18
responsibility, data governance. All of those rules will apply. Different countries. Different states have different rules. Different requirements. They have to be considered, Uh, so the cloud vendors are seeing this, and a lot of those are developing government clouds.
19:34
Uh, an EU cloud, for example. But you, So you have to consider, you know, is this cloud service the right service? The the model. I like to, uh, to think about with the large data sets I typically see is you. You're better to move the application than you are of the data. Your data has gravity.
19:52
Uh, the the larger the data set, the harder it is to manage. So application mobility is a lot simpler. Uh, so we contain organisation, you know, we can move an application moving hundreds of terabytes of data, you know, has time and also has a financial impact. You know, maybe you can have a cost one way a network cost,
20:13
but actually, there's a cloud cost for coming back. So we're talking about and risks costs. So you want to consider what your long term strategy is? Where you place your data? Who owns the data? Who has access to the data from a gate to governance. Point of view when you have the data in your
20:27
own data centre or in a managed, uh, environment that you very much control. Who has access to that? You're part of that recruitment. You're part of that training process. You you you can have your control of the physical security, the network security.
20:42
When you take a cloud service, you have to give some of that away, and you have a a joint ownership of the the the cloud provider managing the platforms. But you managing the access to those that data. And, uh, you know, nobody wants to put all that data in AAA wide public, uh, S3 bucket that everyone's accessing. So there is ownership on the customer to be
21:09
cloud wise, so they actually understand the exposure they have in this new way of working. Yeah, I think that's, uh, often described as the shared responsibility model. Um, across, uh, these kind of things right. Um and so since there are lots of cloud providers out there, Well, we wanted to know where Which cloud provider are you using today?
21:31
The most? Um, you can have multiple choices. Um, because, well, there is also this notion of this hybrid, multi cloud or multi cloud thing. Um, we see often people tending to start in one cloud because there is a particular service which, uh, this provider is offering.
21:52
We're seeing also, um, customers being more or less, um, forced into one particular cloud because a particular service might only be available in the public cloud Or if you, uh, see the last announcement of, uh, for example, SAP where they said, Well, if you're not going all in, um, well, there might be a difference in features between the version.
22:19
So, um, the customers are more or less forced to do to make a choice here, and it's not a free choice. Um, and what we can see Um, here is, uh, it's a good I think it's a good mixture of, uh, the big cloud platforms So far. So, uh, obviously AWS, uh, Microsoft azure are leading, but we also see customers going
22:47
into the Google Cloud platform IBM Cloud and also, um, OC. I So the oracle one and, um, just to share the results? Uh, real quick. I think it's it's not. Not not a big surprise. Um, it seems like our audience here is not,
23:06
uh does not operate much in the Asian region or in China, because, uh, then we would see Alibaba Cloud as an example, because there is no AWS. Um, But what's interesting is to, um, to go into that multi cloud deployment is how do you leverage that services?
23:30
Because you might have a contract with some, um, commitments in Microsoft Azure. But then you have a Oracle database you want to leverage. And we all know it's probably a big effort to have to maintain contracts and billing with more than just one cloud provider. So how can this be simplified? That's a That's, uh, one of the questions. So, um,
24:04
how could you, for example, uh, use things like a or a computer base without knowing that you're on OC I OK, so what we are seeing, uh, Marcus in in this slide is the large cloud vendors working together rather than being in silos, they appreciate that, uh, other cloud has services that they can't offer.
24:32
For example, you mentioned the or database there, So if you wanted an or call rack database to be fully supported. You need to run out of the cloud or on to run an extra data service. Uh, so all Microsoft, uh, have worked together, uh, to have this or fast connect your express route and you can be in as your user a new confusion.
24:55
A database in OC I But it's not about you having to do that work. They've stitched it together. They've got the the federation in place. So the billing goes through Azure, the the Federated single sign on so they don't look like two solutions. So the azure administrator works in his azure console. Likewise, an OC I cloud architect could be
25:21
working on the OC I side and do everything from that side and consume services from azure. So if you if you had an application, maybe the application could be deli in Azure and the reporting services, but the date back end database could be living in OC I. So you can see how we can start having applications which spann multiple clouds now. So and and what is What do I see as a as a consumer in azure?
25:50
Do I have then to get an OC I contract? Then go to the OC I management portal Or can I do it? Basically, just with a There is an OC I service in a which I Yeah, there's a There's a service in, uh, in the Microsoft Azure platform, Uh, and you configure the the network connectivity. There are some steps you have to step over into the OC I to set it up,
26:16
but once it's in place, it's all managed from the azure side and build it, uh, against and and build from that side. Yeah, And when you you use tools like, uh, to form telecom spans are on premise and cloud solutions. So you have one language here which you could use to provision infrastructure and services across both platforms. Hm.
26:41
I think that's, uh, one step, uh, to make it a bit more simple in the multi cloud world. Because, as I said, um, you may not have the experience on all the cloud, uh, platforms to manage it, uh, especially efficiently and, uh, secure. Uh, and this is also something which we we can talk about a little bit later on is the
27:05
integration and from a security and performance perspective, when you connect through multi cloud and hydrocarbon models because, um, well, data, as as we said, there is the data security aspect in terms of keeping it safe. And, um, fence it a, uh, fence it, uh, against access from, uh, from parties. You don't want to have your data access by,
27:33
but there is also some performance considerations you have to take, especially when you're doing, um, a hybrid or a multi cloud, right? Yeah. So I mean, I. I put that that. You know, when you consider the multi cloud the as the consumer, you decide which is the best, best best service to consume.
27:53
You could have a database service in azure or in, or but they have different characteristics. Different scalability, different cost profiles. So you would choose the best option for you. Um, and I think this one, ties in. He ties in very well here because, um, again, it's something you don't have the database per se. Right. Um, because you may may operate many of them
28:17
with different, uh, different bottle characteristics. Yeah. So the other slide was talk about cloud to cloud. This one is on premise to the cloud. So, uh, here we are showing a AM SQL application. Uh, transactional OLTP application.
28:36
Very traditional application, uh, existing in an on premise solution. Uh, you want to have the machine learning and intelligence of heatwave, which is a cloud only service. So, you know, how would you join us together? And this is an example. This is using my square replication, uh, to populate the Myers Square database in OC.
28:57
I and then the heat wave engine can, uh, use that data to provide the intelligence, uh, the out of that. But at the same time, you know, only a subset of the data gets transferred. So, uh, the my scale database in the cloud, it's more of a warehouse. And whereas the one on premises is more transactional so we have cloud services and on
29:21
premises working hand in hand here. And, you know, we're starting to see this more and more. Is this just my sequel specific, or does this also apply to to other data services? They have other data services. So, uh, you have cloud services which are predominantly in the cloud like snowflake, and that could actually consume data that's on
29:40
premises or on the the edge uh, so we want to move the data the right data to, uh, to the right service. So if the service is only available in the cloud, you know, that's what you have to move. I think that's the example of leveraging well, functions like the capabilities of databases to add a external data repository. Let's call it that way.
30:03
Um, which is not necessarily, um the same they are operating on, uh, by D. Um, And what? What we can see is snowflake as an example, um, they have the capability of, uh, external tables so you can connect to a existing S3 repository, which might be not in the public cloud as an example of snowflake,
30:26
because snowflake is usually done 100% in the cloud. The other example is, uh, my my, um, Microsoft sequel. They also can just suck in, uh, external files, which are basically also represented in a huge object store or data lake. You might have on Prem. And just to create your your whole data set, right?
30:51
Yeah. This this whole idea of databases accessing S3 external data. Uh, you know, it's a very exciting initiative that, uh, all the database vendors are doing. And what this means is we no longer have to have lengthy extracts and transformation and load processes.
31:08
We're not moving data into our database. We can leave the data where it is, and the data base can join the, uh, the transactional data maybe with sensor and IOT data, which is very large, but you don't need it all in the database, so you can leave that in an X three bucket and then use that when you need very powerful Honestly, um, since we got this question about, uh, especially performance when you're
31:37
doing these kind of things, like connecting snowflakes to an on Prem flash blade as an example, um, you have to consider that because there is obviously latency in between. And, um, there are cos like where you can mitigate that. But what I think is really appealing here is that database vendors can now access data on S3 is that you can apply the same
32:05
management model in the cloud when your application and data resides in the public cloud as well as on prime. So you don't have to make changes from the database management in the public cloud. Uh, compared to how you're doing it on Prem. And as you said, the lengthy extractions and so on are are then gone.
32:27
Yeah, We've got a question in the, uh, in the Q and a, uh, panel. Have you seen that, uh, come up there, Marcus? Yeah, that That's basically the one, about considering performance and and security for for the network in In between the clouds or, uh, from on Prem to the public. But yeah.
32:47
So, uh, you know, security is a is a major concern when you're looking at data. Uh, yeah. S3 does have some, uh, some really good controls. Uh, but you have to be aware of those and make sure that your bucket is secured. Uh, that it's you have, uh, the process in place. You understand?
33:09
Uh, data masking, if that's required a data encryption. Odd. All those components that you would use to secure a database on premises have to be considered when you're pushing stuff into the cloud as well. Uh, one thing which, uh, you know or is doing, and it's going to be discussed next week at Oracle Cloud World. We've all called 23 C.
33:29
They have a database firewall. So we talk about network firewalls, but now we're having firewalls at the application at the database level. So the idea that you block sort of sequel injection, uh, at and, uh, detect for queries which you don't expect to see, So we get that level of protection now,
33:49
So, uh, uh, it's, uh there's lots of, uh, lots of initiatives there to, uh, help us protect our data going forward. I mean, um, obviously, um, we we've talked about this, I think, this week already about, um, where to apply, for example. Encryption. What do you protect against?
34:10
And you you've mentioned S3. Um, when you have an exchange of S3 Well, probably you might have a storage system already, which is doing data address encryption. So it's against physical theft of the media itself. That's the protection against, um, you might. When using S3. Often you apply TLS encryption.
34:34
So that's data in transit, which is encrypted, or send it over on BPN, which you have control of because again, then the data transfer is encrypted. However, if you store the data then on a publicly accessible as free bucket, well, the data is in clear text there, and and this is something, um, you should be aware of that.
35:02
Um, if you have to mask data because of requirements or if you want to encrypt data when it's stored in the cloud and there is a third scheme, you have to apply. And that's basically data encryption itself, which is, I think, the hardest thing to do, usually because it involves the most, uh, work. And it breaks most.
35:26
Uh uh, Or this is the way you're breaking a lot of things in the end because, um, if the data is encrypted, it's not visible by users anymore. Probably which, um, don't have, um, enough permissions. It breaks things like efficiency features and storage site like data reduction. Because encrypted data does not compress very well and does also de Dulic.
35:51
Oh, it doesn't delicate. Um, if it's not exactly the same data written twice. So, um, lots of things to consider, um, and the performance aspect? I don't know, Ron. Um, it's it's more like, well, a matter of the pipe. You have to your public cloud provider
36:11
potentially, Um, and this is something where you can solve with money, basically like, but on the latency side. I think we would go that into into that a bit later on. Um where are where Can we mitigate that? Because, obviously, if you have to reach a service which is running in an azure data centre or AWS data centre and the latency will be more compared to you have to serve the
36:42
running on your, uh, in your data centre, accessing data locally. Um, and there are ways to mitigate that. Um, we can't get a way. Get rid of that because, well, that's physics. Uh, there are ways to mitigate that. Yeah. And, uh, modern databases. Uh, most of them support shing uh,
37:01
so distributed database where we have data, uh, distributed in different data centres or different geographies. Uh, so that you don't have your whole database in one location. There is a sub of it in a yeah, in local regions. But that that's that, uh, implies more cost, more cost.
37:24
Yeah, but redundancy as well. So yeah, it's you. You design your solutions to match the business requirement. That's that's something very interesting. When when you read some articles or news about hey, uh, company X lost that service or the service is down. You you think of. Well, why is that service not highly available?
37:48
And this is something for discussion, OK? Why people want to go into the cloud at the first glance, it looks cheaper. Or if you apply the same primitives in terms of, uh, reliability, high availability and so on, Well, that equation might change, right? It can become very challenging very quickly.
38:13
Um, there's something else we've seen in, uh, basically four years. Is, um, customers looking into OK, how can I really shift my everything to the cloud? Which means not just a particular service, but a I want to get rid of a whole data centre. And most customers are standardising on virtualization with VM ware.
38:39
Uh, underneath. Right. Um, so a few years ago, VM Ware launched something like VMware in the cloud where you can basically get VM ware services in the public cloud. Right. And the question for our audience here is Hey, um, are you considering running? We work codes in the cloud.
39:03
So we're not talking about, um, using the native, um, operating virtualization capabilities provided by Microsoft Azure or um, AWS is really OK. Would you like to run V MS VM ware V MS in the VM ware format with the V MA management tools in the public cloud to build a probably a hybrid VM ware
39:30
cloud or, um, well, move into 100%. A cloud based VM ware environment will be interesting to see that, because most most customers I've seen are going into a more of a hybrid model. So have you have seen that recently, Eugenia? That, uh, people are going more and more into,
39:55
um, the cloud for VM ware work clothes? Yeah. Yeah. This is mainly, uh, for, uh, of course, VM ware. Uh, customers are willing to keep the same uh uh, look and feel and, uh, operational availability they have in their own premises. VM Uh, a plat platform,
40:21
which is already running on this. The most natural way of doing, uh, uh, of moving into a hybrid or multi cloud solution is to leverage what this kind of service that has been launched by VM Ware. Uh, and, uh, uh have the same approach and the same, uh, management, uh uh, capabilities that mostly the same management capabilities they have,
40:45
uh, into their VM ware environment. The everything uh uh appears to be the same. The interface, of course, is is the is the the the the There will be centre spanning across the the the various uh uh, part of the infrastructure. And they can, uh, have the same additional, uh, components they normally want,
41:09
uh, in their environment for backup, for instance, for backup and restore, Uh uh, you know, restoring operations and, uh, also the migration, of course, between the the two sides or the three sides of the this, uh, cloudlike environment, uh uh, is totally, uh, possible. And the the final result is that, uh, the, uh, in in the environment.
41:32
Uh, it looks like like it's look like a single BP, Uh uh, platform that spans across different uh uh, cloud public cloud Providers. So So to me that, um, moving a workload, which is today on premise in a VM on VM ware to the public cloud seems really straightforward because I I don't know, um probably maybe our audience had made
42:06
some have some experience with that, but changing from a VM ware VM to a native Amazon or azure VM. Um, on the first at the first glance. It sounds simple because you just convert the image and it will boot up. But the things like what? What you just said is just OK.
42:27
What What are the things around that the backup, the, um the networking, the security policies, how to switch that over to a native cloud provider model compared to well, stick with VM or what you have and just move it so natural to me. Yeah, it definitely can accelerate that programme. Uh, if you're if you're not having to
42:51
reengineer you, can you can do that? Um, my simple migration, you know, that's gonna derisk the project and also derisk it from a skills point of view. If you've already got those skills in house, you might as well leverage them and moreover, it in it it's a It's an approach that could be could be, uh, used it if you entire I migration at all.
43:15
So a one, a single migration that you use for moving, uh, everything into the cloud That makes a sense in that way. But if you want to keep a mixed environment, then that means you have to afford an extra cost because you have to learn you have the learning curve. Uh, of all the, uh the the services that you that you buy from the cloud provider and mix
43:39
match that with what you have already in place in your premises. From the application standpoint, it doesn't make any difference, of course. But from the IT administration makes a lot. So for a for, uh, uh, organisation that they have the they got habit. They are using it to use VM ware for, uh, VM ware tools for for managing their
44:03
infrastructure. Uh, having that, uh, mixed approach that seems at the beginning, very easy. Because we have a migration tools. It becomes a nightmare whenever you have to. We have to address issues like rebalancing your workloads because of maybe for costs or for uh, specific usage scenarios that have changed in the meanwhile. So creating a mess again.
44:28
You also de risk the project from a support point of view. If you have a versions of applications databases which are not current, maybe they're not a database service, uh, VR DS or the the database service as you. So you want to carry on running your version of or that you run today your version of server. You know, maybe the VM is the easiest way of doing that.
44:51
Because you can continue to run that version without being forced to do an upgrade. Yeah, I think so, too. Because a lot of people are just virtualizing, uh, proprietary or old software. If you wanna call it, um, into VM. Because it gives them all the benefits of portability backup and so on, because it's at the VM level,
45:11
right? And there are lots of integrations in our audience. Uh, the the responses are pretty interesting. I think, um, that already a, uh, more than 25% are running VMware in the cloud already. Um, about half of them are Haven't decided yet, so it's under consideration. And I would say I would call it. Only 30% will stay on Prem with their
45:35
environment. And, um, there are reasons for for doing one or the other, um, especially VMware in the cloud. Um, so so far as as we know, it's based on VMware's visa and storage infrastructure, and that was Well, I would say it was OK, I because, uh, it it works.
45:59
But if you have very storage heavy workloads. It could become a burden, as you said, for cost perspective, uh, pretty quickly, because you have to scale, um, the the nodes, um, accordingly, even if you don't use the CPU and memory power. But you have to do it to get the storage. And this is where we had to share some exciting news.
46:23
Uh, I think two weeks ago about an integration with the azure VM service and our virtual platform club Los store in general. Um, you you have I think you can explain that picture, uh, better than anyone else here. Yes. Although it's pretty, it's pretty simple. Uh, essentially with the cloud blow to what we
46:50
do, we provide the, uh, uh the the the the cloud, uh, side of a of a BS with block storage, which is a valid alternative to the native VI. And, uh, and this provides actually block, uh, devices to, uh, the, uh, infrastructure in, um, in the cloud with the same
47:19
approach as any other, uh, block device that you could add in your, uh, vsphere environmental prem. So same capabilities. Uh, and if we compare this to the, uh, our for instance, our flash array, Uh, on Prem the in this case, we have exactly the same functionalities. Whatever is, uh, is available in, uh uh in in in an in an on prem environment is also
47:49
available on for, uh, a VP or a platform, uh, in, uh, in the, uh, uh, AD model And, uh, for instance, things like, uh uh, the the additional, uh, resiliency that we had, uh, uh, to to, um uh the the data. It looks like we lost Eugenia.
48:26
Or is it me, Ron? No, I think he's a maybe stuck up the side of a mountain. Uh, might give me a few seconds. So, uh, looking at this picture, Marcus, the cloud block store being a a software device solution, I pure. I guess this means that we can replicate data
48:47
simply from our on premise. Uh, flash away. Um, yeah, that That's one of the benefits, because, um, what Cloud block store is, essentially is it's a virtual version of the flash array, which means we we built up. We built the purity operating system above the infrastructure of azure, so leveraging azure compute as the storage
49:14
controllers leveraging azure storage in the back end for the capacity piece and then offering just everything a flash array on Prem offers in the public cloud. So it's, I think it ties very, uh, well into things like the management aspect, because it's managed as a data store from the vsphere side, as you do it. Uh, on premise, it's managed as a storage array
49:42
where you can exactly do what you described replication like you've used to with flash array, and they can spot on the left hand side. There are some, uh, rectangles, which says management supplement I replica and replication basically means well. You can replicate within azure, meaning that you can have things like an active cluster, which is synchronous replication with
50:12
transparent CRE feedback in azure for your VM ware workloads. But you could also do replication in in an asynchronous way between clouds or between on premise and azure, which is, um, maybe a better solution, even if you want to ingest large sets of data to your a BS solution because the replication is space efficient.
50:40
Um, so it might save you a few, uh, costs here and there, and it can go over a smaller pipe. And with all the data reduction benefits of the flash away and cloud block store exactly so Oh, Eugene, Back. Uh, Welcome back, Eugenia. Yeah, and probably one point that, uh uh you you didn't mention here is the possibility to
51:03
leverage also the evils. Exactly like they are in, um, in, uh, in the UN on Prem, uh, infrastructure. So all the advantages and soon that will be will be totally integrated, uh, for and for replicating the data between on Prem and the cloud infrastructure as well.
51:29
So and following on to that cloud model way of working with the same API S and automation scripts that you develop premises will work in the cloud. Yeah, so, um Eugenia from a AVS admin perspective. So I'm the, uh, VM Ware administrator, and I'm responsible for the as a VM ware solution as well as I might be responsible for on Prem as well.
51:54
What do I have to, uh, do with this cloud block store thing? Because so far I'm used to, uh well, if I need storage, well, I just expand it in my azure console. And then then I'm getting more vis a nodes or whatsoever Programme. Correct. And how do I do this with this, uh, pure CBS thing?
52:16
Because cloud block store so far has been available in the marketplace, you can deploy it using tele form or whatsoever. But, um, it was not it wasn't possible to connect it to the V field level in a BS because it was a closed system. How does that work now? Yeah, that that is pretty easy. It requires, of course,
52:37
the cooperation from, uh uh, uh a W for as for getting the connection created, But once the, uh, underneath the network infrastructure is set by set up by the by the azure, um, operational layer, then the cloud blow store, uh, is completely exposed as an IC, uh, device or set of devices, depending on the amount of, uh, instances that there are for cloud blow store.
53:06
And they are They are exposed on the vs layer, and they can be consumed as usual block devices exactly like they were from, uh uh in the, uh, on prem, uh, sand. And they appear like they are a part of a of the this. The ice sand that is now possible to create within each ES six i node provided by,
53:34
uh, a BS. So the only interaction that has, uh, that that the additional interaction that has to be put in place is, uh, by requiring the, uh the the connectivity between the two lay the the the two, the two layers. And that's it. This is this is something that, um as takes care on behalf of the,
53:57
uh of the, uh, administrator. And once that is done, the administrator, uh, the V administrator, uh, if it is the the same person that is in charge, also called can get the the management interface exposed from both the environments. If it is not, uh, it would be exactly the same cooperation model that could be in place in the in, uh,
54:20
for the on Prem um vhe plus, uh, pure administration. Uh uh, um, competencies. So so net. Nap to. To summarise this. A club block store is almost a native service to to azure because all
54:39
the manual work. But the thing is that we are partnering together with Microsoft and the Microsoft exactly to connect to the ESX level in azure, which is usually not allowed and not exposed to um, to the administrators because, well, it's a service, but you benefit from all the the things that a flash array in
55:05
this case a a virtual one called Cloud Lock Store gives you in a V centre environment, which means obviously in line data reduction. It gives you the, um, high availability and also, um, resiliency using, for example, H a using also replication capabilities with active cluster or, uh, a sync replication or active DR.
55:28
And since it's block storage, I guess we're using also the things, uh, which brings VA. I like the efficient cloning and X copy within that solution, right? Yeah. It is exactly like the the flesh array. How the flesh array works on Prem. This is the equivalent, and all the future are are preserved. But And as you said,
55:51
it will be soon possible to, uh, leverage, uh, high high availability. And the metro stretched the metros cluster also in the cloud. So had the much more resiliency that what you can get from the native visa and, uh, with the, uh, uh, significant. Uh uh uh, cost reduction. With respect to that, uh uh uh.
56:14
Additional, uh, implementation. So talking about cost reduction. I think, uh, sometimes we're seeing a, uh, five or 10 X cost benefit to the native solutions. Because if you wanna scale storage, you don't have to scale your compute notes on the, um, a BS side, which makes the whole solution more affordable and more flexible in the end.
56:38
And we get Ron said it it it I think this is a good vehicle of accelerating. Um, the the journey to the cloud for these kind of workloads and also, um, taking of risk doing that because it's just the native VM ware stuff. It's the native, uh, pure storage stuff you're used to with best in class or efficiency and availability. Whoops.
57:07
Circle that that was so good this year. Um, so let me take a look at the circle that I clicked again on this picture, um, to summarise what is club block story? Club block store is essentially the operating system of the flash race. It's called pure T FA providing all the features we have.
57:33
And 11 thing we didn't mention by the way to derisk is the ransom and remediation. So when you're running a BS as a V MA solution. Together with club lobster, you also get into the joy of having safe mode. Um, for your B MS in the public cloud. So again, it's just everything you're used to with purity you're now getting for your as a VM
58:01
ware solution as well. But they are not just V MS there is containerization. And, um, every everybody talks about stateless applications. But as soon as you want to keep your data, you connect your stateless applications to a database, right? We've had that.
58:24
So what does port works now? Um, or what is this human thing causing problems or, um, obstacles When doing that, I mean, is made for cloud independence, infrastructure, independence, application mobility. But if I tie my application to a database service of one particular cloud member,
58:51
it doesn't make much sense to me, right? Are you generally a mute? Eugenia, you are on. He's on mute. OK, Yeah. In this specific scenario, if you are going to use, uh, a database service from, uh, a cloud provider, it may be it might be a
59:22
solution, but in general, you have also to, uh, to to support existing solutions. Uh, existing application from, uh what you you already have in your environment into the model and, uh, more over Uh uh, the vast majority of the, um, solutions that are based on micro services. They run on top of their own smaller database
59:50
And, uh, and and that since you want to preserve that approach means, of course, that you have to if you have to have a state full applications that runs in internally to to and you perhaps may want to have the same, uh, uh, behaviour that you get, Uh, uh from, uh, uh, uh an enterprise standpoint. And you have a You have to have, uh the the data underneath your,
01:00:17
uh, database is somewhat protected in a high available and, uh, re resilient approach. Uh, which means that you have some water to have a layer A storage, uh, provider layer within your environment that, uh, solves that problem. And that's something you don't get in the cloud you don't get probably even on Prem.
01:00:40
And, uh uh, this gets over complicated when you start also to think at things like how could be possible to take a backup and restore uh uh, your state full application in this is, for instance, what we solve with ports, which currently is the only possible solution that, uh, this, uh, possibility of providing the, uh, high availability or your data across multiple, Uh um,
01:01:12
internally and across multiple, uh, K instances. So And this cover, of course, being a, uh, across, uh, a cross application across the system. Uh, solution, uh, covers a multitude of different cobern flavour, whether it is a cloud based solution like it could be a KSEKS or similar. Uh uh, uh, services.
01:01:42
Uh, it it works the same also for, uh uh, for instance, for openshift or for any, uh, other flavour. So with the with the port works, we have a layer that abstracts the storage and provides all the, uh, protection uh, uh, mechanism that you want to put in place to get, uh, your, uh, state full application ported raised to a to the enterprise level also for
01:02:09
tubers. So in the end, I'm getting, uh, an enterprise data service anywhere without the, um, requirement to connect to a specific, um, provider. Like a hyper skater. So I don't have this dependency anymore, and it gives me the ability to have true application and data mobility across all these clouds.
01:02:36
Did I get that right? Yeah, correct. So the the since it is a layer that that provides the full Astra with respect the storage and you can leverage it. And it's perfectly compatible with any other storage system that you may want to use on your environment as well. Ok.
01:02:56
Oh, sounds cool. Um uh, since we're running a bit, uh, over time, uh, one last question about the containerizing databases and running them anywhere. Um, is that possible for more or less every database? Or is this something very specific to a database vendor or something like that?
01:03:16
Nearly all, uh, database vendors have a operators. Uh, cubing is operators that you can use to, uh, deploy the database. Uh, when you look at databases like Myers and Post, you'll see that they are. They are forked, and there will be features that are available in some clouds that aren't others there. Or so just because you're running post doesn't
01:03:39
mean post post post. So, uh, you really when you look at the operators or you look at the cloud services, you know, make sure that it has the features, uh, that you require. And, uh, you know, the reason that they, uh, they are slightly different is for vendor locking so that you can't easily migrate a post database from
01:04:04
as your to AWS and vice versa. So they are slightly different when you use something like port works and the the generic port, uh, post. Then it becomes cloud native. But nearly all databases, including Microsoft, and or are available as containers. That's good to hear. So but But do you have then to use the vanilla
01:04:24
version and, um, not tie up to the the hyper SCA suits? OK, Cool. Um, so, um, we're at the end. Um, I want to thank you both, uh, Ron and Eugenia for the I think, uh, amazing insights you've shared. And, uh, if you want to know more, we have several, um,
01:04:45
accelerate events running in across Europe. Um, accelerate is our, um, flagship event. Uh, once a year in this this year, it was in Las Vegas, where we unveil new stuff, have breakout sessions, uh, the ability to talk to the people, um, presenting here as an example, And we will go across Europe, um, over the next month and,
01:05:12
uh, hope to see you there. Um, there's also a, um if you're interested in storage as a service, um, there is a webinar upcoming in September 27th where you can register and has run. Uh, outlined. It was a shameless plot. By the way, Ron, um, he will be at, uh, Oracle Open World.
01:05:33
We can, uh, meet him in person and ask him questions and have discussions about your future data architecture. Thank you for that plug. Have a great day. And, uh, see you soon at a coffee break or at accelerate. Thank you. Thank you.
01:05:56
Thank you.
