Axians Network S1 E5
00:00:24] In the beginning, there was Cloud, then we had public and private Cloud, then we had hybrid cloud and multiple clouds and now we have multi Cloud, so multi Cloud just another trendy tech deleon’s for the network fashionistas to fuss over.
[00:00:39] Or is this a real move and evolution of cloud services?
[00:00:42] Here with me to discuss this, Elliot Townsend not showing it any more earlier that senior systems engineer Serial Dandi and also All-Round Multiclass guru Juniper Networks. Welcome, Elliot. Thank you.
[00:00:57] And Chris Gilmore, Axians CTO and my co-host with our live shows.
[00:01:04] And I know around funds of network knowledge. So welcome to you both.
[00:01:09] And so, so, so openly questioned and multi Cloud just another Cloud trend, or are we seeing a real shift in evolution?
[00:01:16] I think we’re seeing a real shift now. So if we think about a few years ago, we saw for moving up into the public cloud to try and save money, we’re now starting to see a shift backwards because that’s not actually the case. You know, there’s different Cloud aspired to be different things now and each other and different benefits. So now it’s all about people trying to kind of utilise those clouds for the right workloads.
[00:01:40] It is very much a people cloud in. You know, if you you’re told by your CIO, VMware Cloud first was literally a wholesale movement of as much into a particular cloud of choice at that time. So it was everything into us, everything into Azure. Now, whether that actually makes sense in terms of where their strengths and weaknesses are, I mean, that that’s half the problem we see now is that they’re sort of evolving into their own specialist kind of areas. And one size does not fit all a bit daft.
[00:02:10] But there’s also a confusion sometimes between multiple Cloud and multi Cloud.
[00:02:14] So definition of both must be quite simple.
[00:02:20] I think multiple Cloud is where you have a distributed workload across a number of of of Cloud vendors and you might have some services from Azure or the best. You might have an Office 365. You’re taking as a service from from from those platforms. You might have some workload inside eight of us, something an Oracle Cloud and maybe, you know, Google as well for good measure, but. It’s very much a sort of distributed architecture and multi Cloud, as far as I can see, is trying to to to bring consistency to that and allow a kind of more controlled exploration of Cloud from from an enterprise perspective.
[00:02:58] Yes, from from from my perspective, it is using, uh, the abilities of all these different Cloud, whether it be public or private, and then seeing them as one and having all the benefits that come with that, because it’s been a sustained growth in Cloud adoption over the over the past year or so.
[00:03:15] But that’s not always been for the right reason, has it? I’m thinking this example here as being Cloud Spruill.
[00:03:21] So I think, you know, if you look at it, it’s it’s very easy to go off and go and create your own Cloud. Um, you know, we’re seeing a lot of, uh, developers going off and going starting up their own Cloud with that, you kind of lose a lot of control SDSN this kind of organic sprawl across the public garden. With that comes lack of control. And with that lack of control, you lose compliance. Uh, so heavy security. So where whereas you get this this great agility for your business because you can spin up workloads fast, what you’re missing out on is that kind of that real security feature, um, which could be a big risk.
[00:03:53] The business it’s driving up through compliance is never an exciting or fun thing to try and mandate in a business. But I think, you know, we had service Bloor then we went to a virtualise the environment and data centre and we had virtualise service for all. Always happened now is we’ve moved into the cloud and in effect, the same model happen again, the VM based version where you could spin up your own ESX and spin up your own VM and then go, I’ve done that, and you’d walk away and not spin it down again, not disable it or remove it. It’s just sitting there taking a resource that’s just been translated fully into the cloud. And we’re seeing large enterprises. Whoo hoo hoo! If you were to ask them hand on heart what the extent of their service, their estate would be in the cloud, I would find quite a lot of real struggle to tell you exactly where their footprint is. And as you mentioned, from a compliance perspective, that’s a matter.
[00:04:45] And they’re not following best practises or haven’t done at all.
[00:04:49] It is very difficult because this agility, you know, if a develop and go off and do their own thing. Yeah. And they’re not going to follow the best practises, the security in the networks will force them down. So, you know, with this, you’re kind of seeing a real cost added to it. So whereas before you had the virtual service problem and the service provider was a cost, but it was kind of hidden. Now you’re saying, yeah, now now you’re seeing a real cost to that and that’s kind of spiralling out of control.
[00:05:12] Yeah, control is a word you hear a lot in Cloud Cloud conversations. Yeah. Is there a sort of context or to define the word control in multi Cloud?
[00:05:22] Yes, I think so. For part of our vision is around the control of hovering all of these different multiple, multiple different clouds altogether. So you’re seeing a single policy set across workloads across any cloud. So if you have a workload on your public, uh, within your public cloud, sorry, and in your product cloud, you can you can move those workloads around without any any concerns about any policy rewrite. And you get the micro segmentation piece as well. And that’s massive consistency.
[00:05:53] I mean, it is it’s being able to apply a consistent security policy irrespective of whether workloads located. So rather than having to build bespoke environments in Egeria and Google or whatever and then try and work out how to apply policy, you can you can apply a consistent security policy irrespective of whether the workload says, yeah.
[00:06:12] So if you look at the public cloud context, you’re seeing that as a currently that’s kind of like a real open field. And that field is goes on way in the distance. And what we’re trying to do is kind of build a walled garden. So there is an element of control and consistency. Um, so with that or the, uh, the disadvantages that we were talking about earlier around, um, organic organic growth and organic sprawl is going to bring that all back in and under one umbrella.
[00:06:38] So I question and also then allows you to have one of the main reasons you go to the cloud, which is cost control. Yeah. Yes. You know, if you have a walled garden environment, you still have the ability for the devil to spin up their own workload. But it’s under a certain set of processes and controls. It means that from a business perspective, you can manage that cost a lot better because you have ultimate control over how many times.
[00:07:02] But you’re also making sure that the the developers behave properly so that the nicest and so, you know, before them to do that. Well, I think it’s very easy for them to kind of architects because creativity.
[00:07:17] Well, you may be, but, you know, what you’re seeing is kind of, uh, the ability for these guys to go off and go and spin up their own services and go and develop their application quickly for the Cloud. That’s not always the right thing to do.
[00:07:30] So doing it this way means they have to accept the application that’s suitable for either the public or the private cloud, uh, meaning that the business has the ability to move that workload wherever they see fit.
[00:07:40] For Sârbu, it’s also. A problem and under the compliance and Cloud myself. So if we roll it back a few hours, a few years is Cloud Cloud or was it originally to be for large enterprise organisations to replicate data all over the place? So that didn’t have to hold it in one. And so what what is either change or what is multiclass change or what is the reality that it’s delivered, if that’s a fair question?
[00:08:09] I think I think the use case for Cloud has evolved quite considerably. I think initially it was I want to get rid of my data centre. All things are there. I’ll push up into the cloud. But they’re still in effect run by that business. Yeah, Azure should move in the resource to a different location. So you’ve in effect, outsourced the box it sits on. And I think there’s now a much more, uh, wide ranging approach by many enterprises. And they’ll take a they’ll take a combination of such services where someone else will host an application and they literally consume it. And that that that consumption model is very attractive for enterprises because it means that they can they can pay for what they’re using and not anymore. Um, so I think that there’s a definite change in how the Cloud is being consumed. Um, a lot of third party vendors are hosting their own platforms as a service, whereas in the networking space it’s a lot of tools and analytics platforms and and third party products that you would ordinarily you would buy a new implement into your network.
[00:09:13] And now even those are starting to be delivered as a software. I mean, it’s it’s the whole model is being moved to that sort of subscription base, but it’s very much a best in practise and moving to the model you need to work to for your business.
[00:09:27] So if you look at it, so you have, uh, you have your public and private cloud and you have this kind of complexity around how applications are architected and how the operations team actually manage that. The the solution you’re essentially abstracting away that complexity, uh, making it easier for your operations team to manage all these different clouds as one, um, but also not kind of taking away that agility that businesses are looking for when they look to the public cloud.
[00:09:56] Um, so that that’s that’s why I see it playing a part.
[00:09:59] So you’ve sat in front of many customers, I’m sure, talking to the end of degree about M.V.E, Cloud. What are the most common requirements then that customers bring up? And so that leads to a multi Cloud conversation.
[00:10:13] I think the biggest the biggest ones are typically around micro segmentation, um, the ability to work on, uh, essentially any modern, uh, networking equipment. So it doesn’t have to be juniper. Um, those are kind of quite key pieces, as well as the walled garden effect that I spoke about earlier, that that definitely rings true to a lot of people because they’re losing that control. Um, you know, I’ve had a lot of businesses that have run reports, uh, across the public clouds, and they thought it was AIX percent. And actually it was a lot greater that they were using. Um, and with that, you have that, uh, that non-compliance issue and the lack of security which essentially could affect their brand.
[00:10:54] So security, the U.S. is I’d say that’s one of the biggest, biggest, biggest areas.
[00:10:59] Definitely consistency and consistency of approach and applying a policy, irrespective of where that data is going to sit on your private cloud or in one of the hyperscale is applying a set of criteria that then consistently applied.
[00:11:14] Yeah, and it’s the model that we’ve taken is more of an implicit allow an implicit denial which changes the whole security model on this on its head, essentially. So we kind of adapt to the zero trust model. Um, and that means that that security that Chris Christie has mentioned is suddenly completely turned over. Um, they have to really think about security and consistency, then breach that security architecture.
[00:11:40] I mean, it’s not you can’t just translate your business into the cloud and and equally, you can’t just translate your business into multi Cloud. It is a step to approach. And you have to look at how your applications work and the kind of resource requirements. But as you mentioned, the security’s a complete 180. You have to look at it from a completely different viewpoint and a bit boring. It’s just not a case of a simple, um, easy task to do. And it’s not insignificant, but it is something that will bring a lot of sort of benefits to businesses once you get there to management orchestration, multi-cap.
[00:12:16] In a beautiful called.
[00:12:20] What does it consist of? So it consists of our control product range, as well as using our art form AIX for the analytic side of things. So that’s a really key piece because, you know, looking at these different Cloud, AIX was an acquisition last year, two years to two years ago now. So, you know, it’s really important because you can use that form AIX to actually analyse to see where Workload’s actually placed any one particular time and then drill right down, even right down to the underlaid to actually see where your pain points actually are.
[00:12:49] And at the moment, that’s a very difficult thing to do. Um, so for us it’s it’s control the control products. We as well as, uh, fornix as the monitoring section I think.
[00:12:58] What are the important things? And one of the things that I quite like about the solution is the fact that it’s not it’s not an end to end locking up. So Juniper have always been quite open about working with others. They realise they’re never going to be the greenfield sites out there. Is it shrinking on a day by day basis? So they’ve had to play nicely with others from day one. Yes, they’ve always adopted an open standard model from their networking background. But I mean, in terms of the multi Cloud, there’s. There’s certain key pieces of software that make up the multi Cloud solution, but in terms of the hardware in your private Cloud, that’s not a mandate you must use Juniper. Is this a case of as long as there’s certain capabilities that I think are pretty much available in most modern switches? Yeah. It can be harnessed to kind of to build the bigger picture, and I think that approach gives them a much more relevant price in enterprise because, you know, Cisco is everywhere, but not everywhere. Yeah. And it’s one of these things where you irrespective of what is there, there’s an argument for work.
[00:14:01] And so that takes up the nice question. The next question then. So if I can put it to you, what is the differentiation then for GDPR in terms of your approach to McLEAR Cloud as opposed to the others?
[00:14:11] I think it’s essentially any work Cloud and the Cloud we can control.
[00:14:15] Um, you know, I think you touched on any workload, any Cloud and you can control what we can control.
[00:14:22] So essentially what we’re saying is, is that we know we don’t have certain product features or product sets and there are others in the market that are better at doing certain things. No names mentioned, of course. And what with that, what we realise in this environment is that we need to be able to put on those and actually utilise those in a multi Cloud environment to benefit the customer. So we don’t want them to lock in such like like some other competitors because we don’t feel that’s the best, best method for the customer.
[00:14:52] I mean, the platform also has the ability to to extend into the branch, extend into the campus. But it doesn’t necessitate that, and I think that’s the difference, so you can go as much or as little as you almost want to, um, control this ginormous, uh, the amount I can do is incredible.
[00:15:11] Um, I could probably sit here talking to you guys for two days, talking about what control can actually do.
[00:15:17] But I think as Chris touched on, you can you can just use just certain elements until you feel comfortable and then build yourself up or you can go for something as high as you grow. Yeah.
[00:15:27] So essentially, I mean, just not a rip and replace for a lot of things. And that and that’s a very key message for a lot of enterprises.
[00:15:34] So yeah. Because like typically immigration plan, if it’s not a greenfield is it. Typically we build it in a certain segment within the data centre until they get more confident of it. They start looking at their business processes because they need to amend their business processes for this. And essentially what we do, it then starts to build out and and grow.
[00:15:52] Actually, that’s that’s a good point. In terms of migration plan, when you turn up customers typically, are they are they already on that sort of pathway or where do you do you find they are sort of you’ve come in because actually they’ve come to a point of thinking, where do we go now? Or are you being sort of invited in very early DDoS to do that, planning that process, planning, management or engineering piece?
[00:16:15] Yes, we as a university like a like a five step. Um, so it’s five different steps in terms of where we categorise customers. Yes. On that kind of on that journey. But we have seen people jump from step one to step five, step five being the all encompassing panacea panacea of multi Cloud. Or they just jump to stage three, which is essentially just the way segmentation placement in their own private data centre. So it’s it’s a real mixed bag, so to speak.
[00:16:40] And I think what’s great is we’re not we’re not, uh, constricted in terms of what the requirements are from the customer. We look at that and we can use the power of control to fit the use cases relevant to the customer at any one time. It’s not like it has to fit our exact use case. We’re seeing lots of different use cases coming out all over the place.
[00:16:59] I think there is also, you know, it’s when you look at the kind of five step plan, it really reinforces the fact that the systems integrator is an important an important part of the party. So, yes, having someone who can work with the customer from the beginning to the end through the different steps and the different phases and however small or large those chunks are, is a very, very important part.
[00:17:22] This is not a case of well, and it’s very rarely a case of his AIX million pounds.
[00:17:29] Give me multi Cloud. It’s a step the process of changing processes internally and how their users use and consume their resources.
[00:17:38] A company like Axians is an important, very important time opportunity to do. Um, I don’t know whether we’re talking out of turn here, but we’ve just completed as a lab, we’re building a multi Cloud lab in and around that job centre currently.
[00:17:54] Yes. So we’re we’re we’re going to be able to do the same levels of demonstrations that Juniper can. And then we’re hoping to plumb a number of, um, backend resources into that to kind of enhance that, um, Lapapo ability to fully demonstrate to customers to Tabac GDPR and what you’re doing that.
[00:18:14] Yeah, it gives us the ability to show the steps and show how we can actually. So bring them from where they are today at VMware, wherever they are in that kind of process where where the next steps are and then how to actually model that out.
[00:18:28] Yeah. So we’re aware that we are a piece of the puzzle. We are. We are not the puzzle. Yes.
[00:18:33] Um, so we look into the likes of which is quite enlightened in my time. Is that quite a lot in view of the NDA.
[00:18:41] So we’re very aware of that.
[00:18:44] And it would prolong on the likes of Axians to actually build this puzzle together, because we’re not just looking at the networking element, we’re also looking at the application side as well, as well as the compute and storage, as well as the business processes. So it’s it’s not just us. So we need to work together. No, and it’s not a single technology.
[00:19:01] Fix it. So there’s a whole breadth of things.
[00:19:04] You mentioned use cases earlier on. If you think of just a typical use case, one each of of my mind think it was that question or I was going to ask you, what are the five steps?
[00:19:18] So I’ll answer very quickly, um, for me, the conversations I’ve had with customers, one of the biggest use cases for multi Cloud.
[00:19:28] And this is. I sort of a more immediate thing than a lot of things is that compliance piece to stop shadow IoT and sprawl. I mean, I have known customers who have inadvertently got into Amazon’s top 10 customers in Europe through shadow right now. There’s a lot of zeros being. I had no idea. They had no idea until they literally put the bill through. And then if you’ve got things like GDPR, whether it’s fines of a significant money, if you are the compliance officer for that business, that is a living hell. So if you look at short, medium and long term benefits, I think one of the immediate things that would would would give value is being able to control that.
[00:20:14] So it might be a given, but so is GDPR a bit more focussed on on this or not?
[00:20:20] I would say for McLEAR much the use case, but it’s around the control agility. Um, so, you know, Cloud Cloud first gives you the agility cost savings and inverted Brackett’s. That’s not always the case. The cost effectiveness. Cost effectiveness. Yes. And what we’re seeing now is what the customer wants, the ability to essentially have a self-service portal for the application dev guys to be able to click on it and say, I need X, Y and Z.
[00:20:45] They click on it and it goes off and it goes and delivers that set on the rules that the networking, networking and the kind of the server team have built together. So, you know, if they say, I want a web front end, it essentially drives into control, control, then push that networking out into the public cloud because that’s where the company has deemed that the best place to actually place that workload and essentially the database that they feel it’s cheaper to have that on the on the private cloud. So essentially spin that up and tie that networking together. So they still have that agility. And essentially it’s all controlled. That control and agility is definitely the one of the biggest, followed by the micro segmentation and compliance piece of it.
[00:21:25] The nice thing is that, I mean, a consistent bugbear of many, many businesses is inconsistency in design and lack of security built into the design for applications from application developers.
[00:21:40] Um, and this gives you a way of building a set of criteria that dev ops and application developers need to work to, which means the end product that comes out is going to be more consistently secure. It’s going to be more consistent in how it can be handled by the state. And that and that as a sort of by product has a lot of value in itself, having worked and a lot of states and a lot of networks where. There is almost no control over how the applications are built and and then having to have the network handle applications that have been built in isolation and he is a hardened to hell and been able to have a level of control over criteria they have to work to before they can release.
[00:22:24] Application has has immense value in itself as well as the hidden decommissioning process, because I’ve worked for many companies thing was just a term that was never actually practised.
[00:22:37] This gives you the ability to actually make sure that you can fail fast, if you will, or rip rip down if you need to without having to worry about the AIX sitting there and costing you money.
[00:22:48] We’re almost out of time. So, um, I feel let me summarise already, but if we’re going to summarise, then we’re talking about security now. Control in. Well, how else would you summarise if you if you were summarising most Cloud to the elevator pitch, Eliot, for you, it’s controlled urgency with with security.
[00:23:08] I’d say EMC.
[00:23:10] I thought I’d say it’s a consistent approach to utilising Cloud to the best of their ability.
[00:23:19] Yeah, right. Cloud for the right workload. That’s almost a strapline that’s like that is the strapline I tried my trademarks from the very guy that’s been very, very informative.
[00:23:29] Chris, thank thank you very much.
[00:23:31] You can be found out by to Cloud on Axians DR UK or GDPR dot net. I’ve been David eighteen. This has been an Axians network production on Disruptive Live.
[00:23:40] Thank you for watching.
[00:23:46] Hi. OK then we’ve we’ve sat down and had a fairly good discussion around this. Let’s have a little look about what that actually means in practise, Zweli. If we look at multi Cloud, if the kind of the clouds parted, that single beam of light was to hit the kind of the ocean at that right moment and the ultimate embodiment of multi Cloud was available, what would it look like?
[00:24:06] OK, well, that’s quite an interesting discussion. So let’s start off then with our developer here.
[00:24:16] So lonely little app developer wants to make their last source of all evil, wants to make their life nice and easy, and they want to essentially what we’re going to do is they’re essentially going to interact with their O.S.S., all their best front end system, and essentially that’s going to be their portal to make all the changes that are necessary for this.
[00:24:37] So essentially, they’re going to log onto this tool and they’re going to essentially say, this is what I want. So in the background, what does that actually look like? So the process will be assessed for system. We’ll talk to Ansible.
[00:24:53] Now, is that Playbook’s or tower or that is Playbook’s, as we say here, we have a service.
[00:25:01] Template, repository. Yeah, so that’s essentially where.
[00:25:06] All of our playbooks are actually stored, so all the kind of multiple different types of scenarios are actually stored here. Now, from whenever one of those is invoked, that will then talk to our IBM or identity management solution. So that’s who’s doing what and where he’s doing what. Where Sarabi Access. You could say our directory services. So who is that person? Yeah. And then from here, we’re going to talk to, like, our project tracking tool. So, for example, that’s a Egeria. So that’s logging all the projects, all the issues is income.
[00:25:41] That means that multiple developers could book stuff for the same project to correct maps through. It’s automatically logged somewhere.
[00:25:48] Then we also have a collaboration tool.
[00:25:50] So that’s used as a conference, as an example.
[00:25:56] My great spending here now from here also we have our source control so that say this, I get up.
[00:26:08] And also. Our IoT binary repository, so that’s say artefact.
[00:26:18] So that’s what the application side of the story. Yeah, so kind of what they’re used to dealing with.
[00:26:23] But obviously what normally happens is that then goes off to the networking team to actually go in provision or the networking elements on all the security elements except for six months to develop from that. Exactly.
[00:26:34] OK, so what we’re talking about now is actually from here, it’s also going to talk to control networking. OK. And also control security.
[00:26:54] Now from here, Contran Networking is going to talk to the likes of, you know, IACP.
[00:27:01] It’s good to talk to yes, AIX I. It’s going to talk to. I was going to create a bare metal.
[00:27:10] As a service, this is where things like IPV and KVM and those types of you know, we just do that. So that goes on and on.
[00:27:20] So essentially what happens is also project dev or application developer logs onto the process of assessment and system says, I want this sort of application and drives it all, but essentially talks to all of these different pieces to actually bring that all together Tractable to really key parts and was a very key part of this.
[00:27:38] And the thing that strikes me straight away is that it’s not it’s not a waterboard juniper solution. It’s not what you are. You are a significant part of. But not all of it.
[00:27:46] Exactly. And I think what really presents Jennifer’s hands is that, you know, we can work with with pretty much anyone because before that open source kind of following it, it’s ingrained within us.
[00:27:58] It’s quite a complicated overall kind of picture. I can’t imagine you could jump straight to this. So is there some sort of phased approach that people could take to kind of get to this end point?
[00:28:08] And certainly so, you know, we’ve devised a five step programme such as the general five step programme that we can go into more detail to be good if could. Yeah.
[00:28:18] So, Elliot, can you go into a bit more detail on the five steps?
[00:28:21] Yeah, of course. So if we start here.
[00:28:31] Are five steps of such service, a step one step, two, three, step for step five, what we’re seeing genuinely is is kind of people around around this mark stability where we see most of our customers one and a half to two. So one being kind of like everything is very manual. You know, your network technicians doing everything manually. You know, they’re actually technicians, they’re not being technologists. They’re not doing anything fancy. That’s kind of where a lot of people actually are and they’re trying to aspire to be here. So this is kind of where you start automating workflows. You know, you start to introduce that that automation into into every type of aspect that you can. Step three is where you start to look at, you know, the analytics and how they then kind of event driven, you could say so looking at the analytics. So feedback loops and how you can extrapolate from your network, automating your testing as well and automating your workflows. Stepha’s around the continuous automation process. You know that full flow end to end requires very little manual intervention. Then step five is the secure, an automated multi Cloud that we discussed just a minute ago.
[00:29:43] OK, so I mean, it seems to me that step one is almost a data centre of old. Yes. The default the default data centre.
[00:29:52] So you could really look at this in sort of three main chunks. You could say sort of the first phases is taking your data centre and turning that into a private cloud. My son, and then all the benefits you get from that market segmentation, the ability to access resources and consume you in a completely different fashion and in the middle seems to me to be almost a kind of a taking that private Cloud and then implementing a lot more automation.
[00:30:25] And optimisation. Yep. And analytics, great. And I think the key word here is also control. So then the third sort of chunk is almost taking that and extending that with the control into hyperscale is correct and using the right Cloud for the right word Cloud AIX essentially what we’re looking at.
[00:30:52] So you can see how people can migrate and move through this process.
[00:30:56] And most certainly this is kind of, you know, I hate to say it.
[00:30:58] I use this phrase based essentially the journey that can go on the journey is good because where people like Axians can help our customers, certainly we help with kind of the moving from one to another. Yeah, that explains a lot more detail. Thank you. OK, so this just gives you that kind of that process through from Luddite Data Centre to Nirvana. If I had a big enough bag of cash, could I come to you and say I would like a full multi Cloud with the hyperscale control and by that now is achievable. Now, is this think it’s a kind of a roadmap or something is fut. that is achievable now?
[00:31:31] I think the most important thing to think of is that this isn’t going to fix your your business processes. You know, you need to get those in order in order to be here. There is a transformation.
[00:31:42] So technically and technologically, you can get to here and get to there for business process and business, organisational perspective. There’s probably a parallel set of steps to get to the end. Correct. OK, OK, so this is an abstract sort of process view of how you achieve this multi Cloud deployment. How does it actually look?
[00:32:04] Technically, the Kaizo that there’s this kind of draw this up in terms of a high level technical diagram? Tractable.
[00:32:12] OK, so can you show us a kind of high level view from a network perspective of what multi Cloud actually looks like?
[00:32:18] And I certainly let me grab my colours. So we start off with our first data centre in DC one, for example, and we’re going to have our gateway routers on the edge right there.
[00:32:34] Now, from there, you know, that side that connects into our spines. And then from there, that connexion to our IP fabric’s our leaves. Typically, like an close environment and IP close, exactly Pannovate clan. Whatever, whatever you see fit. So from here, you know, obviously that scales for however long you need your your decentred will and then from here is essentially. Well, all of your applications and services are currently set at the moment, and that can be functional and service that can be functional.
[00:33:17] So, you know, we could have, you know, our IBM solution that we spoke about earlier. That’s where control would set, for example, the actual application itself, you know, our storage, everything, our computers or SAP HANA.. This is your private cloud infrastructure in terms of the compute part of it. Yeah, OK. Because Axians have got a private cloud reference model. Yeah. That’s where we can actually add to the overall picture.
[00:33:45] Would be brilliant to bring that kind of peace, and ironically and quite helpfully, the pack is actually controlled ultimately by announceable so you can tie it into AIX and stuff.
[00:33:57] So that’s our that’s our initial data centre. And we have the same method again.
[00:34:06] So, again, our gateway route is. Again, that goes into our spines. That goes to our leaf’s.
[00:34:25] So we’re locked into DDoS. Well, OK, that Kaizo and then that connects down into our application stack again. OK. So, again, it’s not altogether so, you know, not going to happen, is that again mirrored here? OK, so you know that that all looks well and good. That will connect to.
[00:34:49] Either a private one or public one, I just call this public for now, Kaizo connecting to to each other.
[00:34:58] OK, now obviously, you know, if you’re using private IP addresses down here, typically are not DDoS so that they can then be accessed over here. But what we could do is essentially we could create this what could I do this and do this in a pink again.
[00:35:13] So there’s this IP set gateway. And again, here as well.
[00:35:20] So what we’re creating essentially is an IP set mesh between our two data centres being able to transfer easily between the two, you’re probably thinking, well, that’s all well and good, but what is the point in that? And the reason for that is if we look at an application.
[00:35:38] It’s a very bad application, so this is. Out to. And it’s not one.
[00:35:48] And obviously, they need to talk to each other and we provisioned it with in control, what Contri would do is actually stitch these together by creating overlies over the top of this, make it nice and easy, and applaud the security policy required and access control using the item in the various.
[00:36:03] So that zero zero trust model that we talked about earlier with control security, that will then ties into this as well. OK, so that’s all well and good from that point. And now we start to look at the full secured media. Multi Cloud. This is our hyperscale, OK, this is the hyperscale hooking.
[00:36:17] So from here, you know, we’re talking around oestriol and this is an example now to mirror what’s down here, best practise and essentially be pieces. And essentially we’d have like.
[00:36:34] Like you said, Gateway.
[00:36:36] Again, yeah, and again, this thing gets added into our OK, typically that can be done over violence as well as a new VPC comes online that will get sportin. And again, you know, bring in another.
[00:36:52] Public Cloud are the public clouds are available, for example, we use Azure again, same again and time that back in that and gives you physical private Cloud resilience from an application level. Yeah. And also resilience in the hyperscale themselves. Correct. And the ability to move your applications from one to two to another to the Cloud and any Cloud you want to.
[00:37:18] Correct. So what you can do is from but you can either pull in directly and talk to us directly or you could if you, if you wanted to, you can still control.
[00:37:29] And these two instances, for example.
[00:37:31] And essentially you have all the smarts that you would have in control that you would have in your private Cloud replicated in enterprise as well. So, you know, you kind of have that flexibility to kind of go where you see fit.
[00:37:41] Sure. I mean, the thing that strikes me is that if you’ve got IP set concentrators here and here and we ultimately you could use for your one connexion, you could actually build an SD-Wan environment, certainly, or in fact, leverage an existing SD-Wan environment, connect those data centres up to the Cloud using virtual functions.
[00:38:01] And then what I like to do is you then have the ability to have your headquarters. And a number of brunch sites.
[00:38:13] She’s not going directly into Cloud IoT correct, and you could, in effect, bring that whole Cloud environment on netz those those sites. Exactly. And, you know, you could also take it one step further and say, right.
[00:38:26] You know, I’m a parent company and I have different operating companies and essentially I want my operating companies to be able to consume all of this data.
[00:38:35] And essentially you can you can have that.
[00:38:37] You can have shared resources. You can have shared environments down here that are controlled by the security policies. Yeah, there’s no leakage issues. Yeah. It might help towards your GDPR compliancy. They might. They might and allows you to to to in effect, build the cost model into a kind of a larger picture you’re looking at from a kind of a group perspective rather than individually. Exactly. It’s all about control. It’s all about security and it’s all essentially about agility.
[00:39:09] Here we go. How can a boy one comes to Juniper?