Craig Lodzinski: Multi-cloud is going to happen to every organisation that consumes technology whether on a small or large scale. The myth that this is something that you can choose one day that suddenly you’re multi-cloud, it's just not reality, this is coming and it's about preparing for it.
Michael Bird: Hello and welcome to Explain IT, brought to you by Softcat.The show for IT professionals, by IT professionals that aims to simplify the complex and often overcomplicated bits of Enterprise IT without compromising on detail. I'm host Michael Bird and over the next 30 or so minutes I'll be challenging our panel of experts to take a different area of the IT ecosystem and of course explain it. In this episode we’re going to be talking about multi-cloud - what it is, why an organisation should care about it and what the future of it is. And with me to help is Dean Gardner who is Softcat’s Chief Technologist for cloud. Dean, this is your fourth time on series two of the show - did you bring another interesting fact with you?
Dean Gardner: I don't know if it's an interesting fact, it's a public service announcement. So in 2004 I was actually diagnosed with testicular cancer. And it’s one of those things you go through as a 29 year-old as I was at the time, where you don't think about getting yourself looked at and it's kind of a big thing. So my public announcement is that if you have something that's a lump that you need to get checked out, get it checked out, it's a thing you should do. So I did that and got caught very early and it means I'm still here to do these lovely podcasts.
Michael Bird: Well thank you for the public service announcement.
Dean Gardner: You're welcome.
Michael Bird: And we also have Craig Lodzinski who is Softcat’s Chief Technologist for emerging technologies. Craig, this is also your fourth time on the show, what is your interesting fact?
Craig Lodzinski: Darryl DMC McDaniels of Run DMC follows me on Twitter.
Michael Bird: That is an interesting fact. And to help us we’ve also got Jaspreet Singh who is the founder and CEO of Druva. Jaspreet welcome to the show, welcome to your first time on Explain IT. As with everyone else did you bring interesting fact?
Jaspreet Singh: I did I was debating between two - I've been to 26 countries so far on personal travel and I've never watched Game of Thrones!
Michael Bird: Of the 26 countries you've visited, what was your favourite?
Jaspreet Singh: My favourite probably was Switzerland. I did my first base jump in Switzerland, so it's pretty interesting.
Michael Bird: And Jaspreet as founder and CEO what does your job entail?
Jaspreet Singh: I think anything between cheerleader and a ringmaster. Just to set the track for the company, inspire people who can do awesome things and I can watch them grow and have some fun.
Michael Bird: So multi-cloud what is it? Is it just a fancier name for hybrid cloud?
Dean Gardner: So we think that hybrid cloud has been an evolution of the datacentre into one public cloud. So if you are within your environment DC you want to consume something from Azure, that in itself, in our world is a hybrid cloud. But what we’re seeing is that evolve as more and more Saas based applications have come along, we then move into multi. Or if you’re in public and in multiple versions, you can be in Google, you could be in Amazon, you could be in Azure, and we’re seeing that as well. That in itself is multi-cloud. And so there's an evolution I guess from where organisations were going, “well we're just going to go public and will plug it in or we'll go all cloud first,” then at that point you're in a hybrid state, but we’re seeing customers go to Office 365, we’re seeing G suite, we’re seeing Workday, HR application, we’re seeing Salesforce, we’re seeing all these different Saas based applications running cloud and obviously being able to bridge all those together with identity strategy, with security and being able to operate those you're naturally moving into a multi-cloud world and we are seeing that happen everywhere.
Jaspreet Singh: I think multi-cloud is a natural transition to what Dean said. Cloud by itself is buying a resource anywhere in the world at a fixed price point without having to worry about how do you build it from scratch. As enterprise last 10 years have tried public cloud and they've grown by and understood what the merits are, or demerits are of public cloud. The security parallel operating a public cloud. The natural transition is to look for a freedom of choice to understand best-of-breed applications or best-of-breed internet providers in the cloud to think about how they should distribute the workload within the best-of-breed cloud providers. So what the public cloud era is naturally going to be a multi-cloud era.
Michael Bird: So surely having all these applications and data and things like that in lots of different places with lots of different providers, presumably lots of different contracts as well, I mean that must be a huge pain to manage.
Craig Lodzinski: Yeah absolutely we've seen it in the traditional IT world as well where you might go to an outsourcer or single supply your entire stack but you're going to lose certain advantages, the flexibility, the ability to change, and inability as Jaspreet was saying, to go to the best of breed. And in cloud it's a lot more pronounced firstly because it’s a moving target. It's not the case of previously if you were tied to a platform you were tied to a hardware platform, even going back to the mainframe days, if it was a bad choice for your applications and your data, you're stuck with it because it's a big capital investment. Now with the desire for increase in agility, increase in performance requirements and this really moving target, it is very difficult to manage but it's about trying to find the best of both worlds, trying to find the best outcomes for your organisation.
Dean Gardner: And one thing we are seeing is that you're starting to get business units deciding what IT they want, so it's the case of enabling them to have applications that are specific to that particular business or the actual business units within that business and that means it could go in multiple locations, it's difficult to govern and control that, but you need to as an IT function. There’s a lot of foundational stuff that needs to be done from an operational standpoint to facilitate that way of working and I think that's where some companies have struggled in that multi-cloud world.
Jaspreet Singh: Let’s just for a moment try to deconstruct the cloud. Public cloud is a set of applications which enterprise deploys, following with some durable tools, and then comes some foundation infrastructure. If we draw parallel to a mobile world which happened 10 years ago there was a platform, either iOS or Android and then you had apps on top of it. The original use case for mobile was to consume a primary application, being emails or contacts or a basic productivity on a new mobile environment but then every single software in the world got mobile same trend as fall into cloud. You’ll have a few pieces of foundational be at Amazon, Google, Azure and then every single application in the world will have a cloud front end or a cloud consumption model or a Saas consumption model, so naturally the default of cloud will be a multi-cloud because you have a plethora of applications running which you have to consume and to Dean’s point you have to figure out interoperability or integration points through access through identity to map these applications together and then they'll be some convergence of thought just like you had app stores on mobile, you’ll have some way of integrating billing, some way of integrating identity, some way of integrating consumption of these apps. The foundation layer is tricky because foundation layer is how you build an application and how you consume an application, a custom build application, just like a mobile, and most enterprises in my mind will try to have optionality - one or two - just like in databases you had Oracle as a sequel not interportability but interoperability and optionality of one or two platforms of choice they would like to build applications on top of it. So in my mind multi-cloud will be a plethora of applications which you will stitch together which will - natural course of action - and a choice of one or two foundational pieces of architecture which will deploy in a company.
Michael Bird: So do we have any cool or interesting use cases for multi-cloud?
Dean Gardner: I think what we’re seeing over the last few years, development teams are going to the likes of AWS and we’ve got a lot of organisations now because they've invested in Office 365 going to Azure because it’s natural state and there's a lot more organisations looking at Saas-based offerings because they’re available. And I think that those things are happening. Then there's the operational element where you're looking at technologies like backup and you’re looking at technologies like monitoring and then there's the next stage which is the critical apps and this is where we seeing some stalling in some of our traditional customers and some of the public cloud providers realise that and I think what you've seen over the last year or so a lot of those public cloud providers are now partnering up with those traditional on-premise vendors, and obviously you've got a lot more reach back in from the public cloud into the datacentre so I think the hybrid or multi convergence with vendors is happening and I think that's essentially to accelerate potentially some of the challenges there are around those applications that maybe can't go to cloud native or be refactored.
Craig Lodzinski: And I think particularly in the research sector in areas like AI and machine learning, there’s a lot of secret sauce hidden in the different clouds, so we’re seeing organisations might want to use Google Cloud Platform to get access to their TensorFlow processing units. And that's a unique hardware proposition that nobody else has. In other organisations you’re going to see them using Oracle Cloud for their Oracle workloads, not only for performance but also for cost benefits. You may see it as well, perhaps more so in countries like Germany where you have a much greater awareness of things like data privacy, using something like an IBM cloud platform for the data storage and breaking out into something that's a little bit more flexible around the devops side of things or potentially storing that data on premises and brokering it out that Google Kubernetes service.
Jaspreet Singh: I think from my point of view the cloud providers are trying to specialise now. The code of public cloud is commoditising from a computed storage point of view, so how do you specialise? My money is if Microsoft and Oracle try to specialise as an application provider, Saas providers, which they call DNAs with the partners called DNAs, and Google and Amazon AWS will try to specialise as developer clouds, as they build more and more devops tools for machine learning or building applications, build on AWS and then they will lift and shift clouds which will be into the stack providers which will lift and shift applications from datacentres to public cloud like VMC and like IBM. And then a few localised players in Germany or Russia or Japan which will be more localised to local privacy laws. The way of specialising for cloud providers helps the customer to make their choices easier and build this interoperability model much more seamless and better.
Michael Bird: So for an organisation why would they consider multi-cloud? Is it something that they just can't avoid?
Dean Gardner: A lot of the customers we talk to, they just don't want to be running data centres. It's not the first choice and as more applications move towards Saas, we’re seeing that's the first choice - cloud-first we hear a lot with a lot of organisations – we feel it's more cloud-appropriate because it is an application centric decision and at that point you can start making decisions on where best from them to run. So it's the fact that people just have more flexibility and that for me is the driver for a lot of organisations.
Craig Lodzinski: And I think certainly architecting for multi-cloud is absolutely essential because if you tie your ship to purely native services on a singular cloud provider, you’re just buying the same stack of infrastructure in a different way. Whereas if you architect for multi cloud, architect for cloud native. Decouple your data and your applications from each other rather than running monoliths. That gives you the flexibility, it gives you the opportunity to use what was really promised in the cloud and that flexibility, that ability to react, to be agile, to be the types of organisations that are succeeding in today's economy.
Jaspreet Singh: From my point of view I’ll make two points here. One, is cloud being a complete force of nature right now, and number two being multi cloud or cloud in general is actually not just a technology trend it's more a business imperative. Let me walk you through both of them. In 1903 when JP Morgan went to Thomas Edison and they said, “let's invent electricity, let's show the world what an AC generator can do,” they invited a bunch of people to show them how they can light a bulb. The next day there are multiple people showed up to place orders to buy AC genset, and a few weeks later when Nikola Tesla showcased how AC could kill a full-blown elephant, JP Morgan came back and said, “you know what, the idea of pulling a genset makes so much sense but the wife is kind of scared. I don't want to blow up my backyard by trying to light a bulb, why don't I put the genset in your back yard and I can lease it from you?” I think that's a genesis of cloud. IT is too much tired of blowing up the backyard by trying to chase innovation but it's not code of their business, so cloud as a 30 year trend, it’s not a small, it’s a 30 year trend, and we’re only in the genesis of it, just like software was, just like internet was it's a megatrend and it will power the next trend called AI. That’s my first point. Second point is if you are the nascent stages of cloud, given the fact that datacentre were so fragmented, the cloud naturally will be very much a heterogeneous environment. And if your go back to my first point the whole promise of cloud is not so much software defined datacentre promise of technology change it's more of a promise saying at a fixed price point anywhere in the globe I can buy a class of service at a predictable SLA. With that in mind it's only natural the whole cloud will be a multi-cloud world and every business has to adapt to interoperability model that they can use multiple class of services, multiple class of SLAs across the world. That's the true promise of cloud.
Craig Lodzinski: Without wishing to coin a phrase and steal from Jaspreet’s example, it's the lightbulb moment. Cloud is going to be that utility, it's going to be a utility model that you bring in and you consume but all these services, the data that's flowing through everything, the AI services that now seem very fancy but in the future as this market matures are going to be just fundamental underpinnings of applications that we don't really see, in the same way that you don't really think about how electricity works in your house, just, it is a thing, it's kind of ethereal. I think cloud services are going to become that because - Mark Andreessen’s always said software is eating the world, as software and technology become so deeply imbued within everything, the cloud is a hub and it's about what you do with it, about what sits further up the stack. Now with public cloud we’ve abstracted away from the hardware layer so you don't deal with the servers anymore, somebody else manages a warehouse scale computer. You never really need to think of how that design works and that’s similar to you don't have to go down the store and buy a load of magnets to build a generator, the electricity comes in on a pipe. What we have at the moment though is that first layer up. So if you go back to the early 1900s in the UK you can see various different types of plugs for example and we have that at the moment, and we have a lack of interoperability but there will be this convergence, we will agree on a common standard for the basic fundamentals and then it's about what we place on top to use it, what devices are consuming this utility, what services we provide off the back of that, that's where it starts to become really interesting.
Michael Bird: So then looking at multi-cloud in a little more detail why would an organisation avoid it, what are the disadvantages?
Jaspreet Singh: From an application point of view there's absolutely no avoidance of it, they have to absorb multiple applications. From infrastructure point of view, there is a tax to understand multiple stacks, and how do you build on top of them. There was point in time you had x86, you had Intel compute on your Macs, and you had the ARM processor on your mobile, the trade off had to be made for a particular experience, so some things are not avoidable but there is a tax to run parallel platforms, to understand interconnects, to understand standards, to understand security, to understand billing but it will not be avoidable in most circumstances.
Dean Gardner: I think we’re seeing that actually, where if you are building an application then that application, if it's developed in a cloud platform it's one thing, it's got multiple things in that app but if you go to an enterprise organisation or in public sector in certainly in the UK, you could be looking at 100 plus apps and that's a lot of applications that run in a traditional sense in a datacentre. If you wanted to then move all those to cloud that’s a lot of complexity that you're having to move, or build to replace. And with that complexity comes cost. Not only the cost to transform those applications but actually to run them in those platforms. You have to look at it and go, “is it actually worth it? It may be better but is it worth me doing that and transforming it or is it just worth me having a hybrid or multi-cloud with the datacentre still part of that conversation?”.
Michael Bird: So how would an organisation manage all of this?
Jaspreet Singh: There's a real skill based shortage in the world, which definitely the enterprise have to carefully assess and understand. Working with the right cloud providers, working with the right partners to really assess the transition datacentre there are to lack of datacentre anymore and how the cloud powers the enterprise has to be assessed how the core business is still selling what they do, primarily, and IT is just a way to get there. So assessments of transition is of paramount importance.
Dean Gardner: And we're seeing a fair few of the companies we talk to, they're doing two things, they’re partnering, in some cases, but the likes of Microsoft, Amazon, Google, they've got some incredible training material that they've opened – they’re just saying, “look this is how you do stuff,” and they're building frameworks that you can adopt and put them into your own cloud centre of excellence, if you will. There is a lot of help now and you can up skill your existing staff because they don't just exist - people don't just go, “I've started IT, I know how all these platforms work,” it doesn't work like that so you need to plan in how you're going to change and build training and plans to be able to get those people up to speed.
Craig Lodzinski: I think there's two elements to it. Firstly I think there's the idea, and you see it in sports coaching and stuff, whereby you spend 90% of time in practice on what you’re great at, and 10% on what you're bad at. So first of all, focus on what your business does brilliantly, what your IT teams do brilliantly, build that out, if you've got rockstar developers already, use them, let them lead the charge and get you to the cloud and augment their skills, don't try and take somebody who's fantastic in one area and convert them to something else because you really have a passion project over here. Really focus on your business skills and building what you're great at. I think also, as Dean and Jaspreet have put, it’s a community play. The world and technology as a whole is very complex, even simple items today require a cast of tens of thousands of people around the world to build and the same is true of organisations. The systems that we have require a lot of different people, a lot of different skills, a lot of different resources, so absolutely, partner, bring different people in, build out this super generalist model whereby you have people who are ‘Jacks of all trades’, and look to focus on a much wider perspective rather than just trying to do everything yourself because it's just impossible.
Jaspreet Singh: With any change or disruption there are two elements, one is a reskill. As Dean mentioned, you have to focus on reskilling your organisation and your core part of team by adding more skills, moving people around to relevant areas and second thing is change management. As Dean mentioned the whole framework for cloud adoption as multiple cloud providers have put forward, there’s the whole notion of testing, foundation, migration reinvention, how do you assess which workloads make the most sense? How do you put forward a framework to test the thesis, prove it out and then build a business case to follow the model across your entire organisation?
Dean Gardner: And doing all of that while still running a business. Because you've still got to facilitate what's happening within that business and there’s such an appetite to change, there’s such an appetite to go and move to these cloud first models and just meeting that demand is a challenge for us as a business, but it’s also just a challenge for our customers.
Michael Bird: So we’ve talked a bit about the advantages and disadvantages of multi-cloud. What are some of the myths or the common misconceptions that people have?
Craig Lodzinski: The myths of multi-cloud are the same myths that we had in original cloud and hybrid cloud and cloud light and diet cloud, and diet diet cloud with vanilla… it's the same as we had but magnified, because multi-cloud is on a larger scale. And some of the myths are self-perpetuating, we see it from certain leadership and people who want to be seen to be doing the cool thing but not having a path on how to do it. So the idea of multi-cloud is that if do it correctly you'll have greater agility, you'll have a better cost base, you’ll be able to be a more innovative, more agile organisation that can really take advantage of all these possibilities. The big myth is that you can just put stuff there and it happens. It's kind of like buying any item and it’s suddenly better. Going out to buy $1,000 worth of Japanese steel knives and £500 worth of pans suddenly turns you into a great chef. It doesn’t, having the best tooling is great, but it's about what you do with it and a lot of the myths around it are about how moving from, “oh we were on premises and now we are in a multi-cloud environment, we’re fantastic”. It’s cloud washing and it doesn't really stand up.
Dean Gardner: I think a few years ago it was that security thing, it's not as secure, and I think now there’s things like shared responsibility model that you have to adhere to, your application ultimately is your responsibility as a customer, the platform itself is probably the most secure platform and I’d say all the public cloud providers they've invested more than any organisation could on security. Ultimately you're still responsible for the application so I think the security myth of a few years ago was, “it's not secure,” it is, from an infrastructure perspective. I think the other one is cost, it's cheaper, we’ve seen that a lot as well and ultimately it can be, it cannot be, it depends on you running it and it goes back to what Craig was just saying, you have to still run the platform or the application on the platform which you're playing someone for, so you need to secure it, you need to make sure it's cost-effective, so the operational models just need to evolve to accommodate that. So all of the comments and the myths essentially, of those things - cheaper and it's less secure - I don't think they're true, but there are still things you need to do make that happen.
Craig Lodzinski: Perhaps the biggest myth around multi-cloud is that it's a choice. It's not, it will happen whether you like it or not and whether your IT department in your organisation embraces it or it happens as part of shadow IT and as part of just the way the market goes, multi-cloud is going to happen to every organisation that consumes technology whether on a small or a large scale, and the myth that this is something that you can choose one day that suddenly you’re multi-cloud, it's just not reality, this is coming and it's about preparing for it.
Jaspreet Singh: From my perspective there's two core myths, one is a myth of inter portability, people think that they won't have the option to import their workload between multiple clouds and it's a must have before you choose cloud vendor. I think the choice of interoperability or best of breed is a good one, where trying to make applications port between multiple clouds it's too expensive and probably never happened in datacentre itself. People did not port between Oracle or SQL, VMWare or Hyper-V, they migrated, but they never made it portable, so trying to keep applications portable between multiple cloud is the wrong thought process. Second is the myth that I can do what I'm doing, buy the next piece of hardware and think that the hardware will support simply out of box support multiple cloud. This is a change for enterprises. The change has to happen by very likely not taking your legacy to the cloud area, but trying to re-think what skills and what software and what applications are needed in the new world. So going to your legacy to support two new flavours of cloud isn't going to work, you have to think about how the world looks different now and what tools make sense in the new world.
Michael Bird: So let’s look to the future. More multi-cloud?
Craig Lodzinski: I think it's going to be more of everything - more Saas, more agile, more cloud or cloud style technologies, there’s obviously the risk of big exogenous shocks whereby a country just clamps down and says ‘No you can't have data anywhere,’ or vice versa a wide-reaching act similar to the cloud act in the US but much broader powers actually forces organisations to entrench, but unless those unexpected shocks happen it's going to be more of the same. What is new will become old, AI will no longer be the latest greatest cool scary thing, it will start to drip down into being this utility model, the things that are emerging tech now will become established tech and new and elaborate things will emerge to take their place as the new kids on the block. Most organisations are still only at the tip of this and not really taking the best advantage of it and we're seeing different markets really ripe for disruption I think if you look at HR software, a tiny fraction of organisations still run HR software on premises, there are some niche providers out there, but the elephants in the room, guys like Workday who are out there and have built such a great platform that everyone's jumped on to that kind of thing. We’ve seen the Salesforce affect in CRM – with Salesforce, Dynamics, with the Oracle proposition as well, everything’s moved into that space. There absolutely will be small use cases for niche providers, there will still be people who go out there and do things on their own, who build themselves but in the grand scheme of things on a very large scale we’re only going to see a very small set of niche providers serving very niche markets. In a similar way to the car industry as we’re seeing consolidation, you have your Toyotas, Fords General Motors, huge global organisations and there's always a place for Ferrari and Bentley and these very specific manufacturers, but they will not be the normal, they will not be what everyone is using the weight of the market will generally continue in the direction that we’ve already established.
Jaspreet Singh: I think Saas will diversify quite a lot, I agree with Craig on that the Saas market, application market will diversify broadly. I think interest in the market on the contrary, I believe will shrink to a few players, I think the providers will be more like semiconductors, the biggest players will gain more momentum and there will be standardisation there. My money is that one of the big players we know today will fold in a year or two. They won’t be able to compete with the general momentum in the broad space but people will also try and pivot. I do see there's a space for someone who is AI specific cloud or a very specific niche expertise, someone with a general-purpose infrastructure devops and building application. It will be a market of, just like semiconductor, market of one or maybe two.
Dean Gardner: I think we've seen certainly the public cloud arena has been, I wouldn't say it’s one, certainly, but it's market share growing within the major players and even we look at someone like Alibaba a few years ago, they didn't exist in a public cloud, well now they’re on the top five list because they are specific to a region and now they're actually going out from that region into the global world, but I don’t think that’s going to happen a lot, that's going to be fairly unique but I agree with Jaspreet, I think there are going to be specific clouds that do specific technologies that we may not even know exist yet. So I think that’ll evolve, but I also agree that the infrastructure market… people are still buying infrastructure, but is that out of a need to do so? It's probably because there’s stuff that’s needed to facilitate today but in five years time when it becomes more common to do what is happening within going towards public cloud or using multi-cloud then suddenly that infrastructure market really isn't as necessary as it has been, so I agree with that as well, that will shrink and you'll be down to maybe one or two vendors that specialise in that area that integrate into those different cloud platforms. I think we’re already starting to see that, actually emerge.
Michael Bird: So to summarise?
Dean Gardner: So organisations are going multi-cloud, if not most organisations are already in multi-cloud and what we’re seeing is that shift to understand more about application workloads and data to determine where they should run, some more intelligence around that and then obviously operational controls, governance evolving of the traditional infrastructure teams to become operations teams I think that's definitely happening, and to be able to essentially take some ownership on how those elements are secure, how you can get visibility of those multi-cloud environments and essentially just run them as an IT function and facilitate the fact that it is happening.
Craig Lodzinski: And I think every multi-cloud deployment will be different and what multi-cloud means to your organisation depends on your organisation, what it looks like, what your business does, what your IT estate looks like, for sure, but in the same way that every business is now an IT business to a greater or lesser extent, every business is or will be a multi-cloud business to a greater or lesser extent, and how you work with that, how you make that a reality, how you embrace that and transform that into aligning with your business objectives is an absolute imperative in the modern economy.
Jaspreet Singh: I think as Dean said, multi-cloud is here and only in the right direction. The choice is now how to understand security, operations, cost, billing, data architecture, how do we put that together to make sure the transition is more seamless and the cloud is more connected?
Michael Bird: Dean, Craig and Jaspreet, thank you so much for your time it's been fascinating speaking to you all. If there's anything in the show, listeners, that has piqued your interest, or if you'd like to speak to someone at Softcat about anything that we’ve talked about, do make sure you check out the show notes, we're going to put some of the stuff that we've talked about today as well as some contact details on. Also make sure you click subscribe wherever you get your podcast and we’ll deliver the next episode to your device as soon as it lands. So thank you for listening to Explain IT from Softcat and goodbye.
Episode 1: 2019 Tech Predictions
Episode 2: AI and Machine Learning
Episode 3: The Future of IT in Healthcare
Episode 4: Security Trends
Episode 5: 5G
Episode 6: Supply Chain Attacks
Episode 7: Rise of the machines
Episode 8: Unstructured Data
Episode 9: Quantum Computing
Episode 10: Multi-cloud
Episode 11: The Future of Work and Workplace