ALISON BROOKER (AB): According to Flexera's 2021 State of the Cloud report, 92% of businesses have adopted multi-cloud, but not all of them are getting the most out of this increasingly popular model for cloud consumption.
I'm your guest host, Alison Brooker, and this is IT Availability Now, the show that tells stories of business resilience from the people who keep the digital world available.
On today's show, Leon Godwin, Principal Cloud Evangelist at Sungard AS, does a deep dive into everything multi-cloud, including the benefits, key motivators driving mass adoption and tips to getting the most out of this cloud strategy.
Leon, welcome back to the show.
LEON GODWIN (LG): Hi, Alison, it's good to be back.
(AB): So, I want to get into why more organisations today are adopting multi-cloud. But first, I think we should level set a bit on what multi-cloud means to us just because different people use the term in different ways. So can you kind of give us sort of an overview there?
(LG): Absolutely. And you're right. Multi-cloud means different things to different people. If we think about cloud and its cause for consumption models, we have infrastructure as a service (Iaas), PaaS and SaaS. So multi-cloud can mean different things. I personally wouldn't describe consuming YouTube and Netflix, which are SaaS services, as being multi-cloud. For me, it's when organisations consume multiple cloud services from two or more providers of computing and storage services from a single network architecture.
(AB): So what do you think makes multi-cloud a good strategy?
(LG): For starters, different cloud environments have different features, advantages and benefits. Multi-cloud isn’t a one size fits all solution. By having choices, you can make better decisions for your business. When you only have one choice because you're dedicated to one cloud, then you're limited to just the capabilities there. So choices give you options, options give you better opportunities to make the right decisions for your business.
(AB): That makes sense. Can you give us an example of what this might look like?
(LG): An organisation might prioritise cost predictability over other considerations. However, the applications might also have dynamic needs and they need elasticity where they can scale up and down at different times of the day or based on different workloads. So depending on those needs of those workloads or their business’ priorities around cost predictability, and a whole bunch of other factors, they may choose for private cloud for maybe their more traditional workloads or maybe their crown jewel workloads, but may opt for dynamic workflows to go into the public cloud or software as a service or platform as a service.
For example, lots of organisations will consume development environments that they can power down because their developers are only working nine to five, and they can power it down at those other times to save those cost efficiencies. And they can scale up more rapidly within those public cloud environments.
(AB): So speaking of cost, how does multi-cloud use really affect budget planning and what should organisations be considering around that?
(LG): Different cloud platforms have different consumption models. Private cloud, for example, is typically billed for a committed amount of consumption per month, regardless of the usage, much like your broadband connection. You pay a fixed fee and whether you use it or you don't use it, you're going to pay a fixed amount.
Whereas the public cloud tends to be billed per second or hour of usage. And that means if you've got workloads that potentially don't need to be as big or even available outside of core business hours, public cloud can be very cost-efficient because you can power those systems down or shrink them down in size out of those core business hours, thereby optimising the cost because effectively, your consumption outside of core hours may be significantly less than in your core hours, and that may deliver better advantages that can ultimately deliver a faster return on investment.
(AB): Can you explain that a bit more? You know, how does choosing different cloud platforms really allow for businesses to realise that faster return on investment?
(LG): By having different cloud platforms that have different consumption models, you can align those consumption models more closely to your budgets and your needs. So for example, having certain workloads that are available. Let's say it's a virtual desktop that's being delivered out of the public cloud, and then that can be shrunk down massively. And maybe provisioning those just to enable a third-party workforce or supplier to come in and do some discreet work for you. You're not having to go through the cost of potentially buying laptops and what have you where you might be left with legacy debt.
Equally, when you go and you purchase hardware equipment within data centres, where you have quite peaky requirements, you're buying capacity based on these peaks. And when outside of these peaks you are effectively left with quite a lot of legacy debt.
So by purchasing capacity when you need it, whether it be in the public cloud or the private cloud, bursting up on demand to meet those capacity requirements, it means your total spend is reduced. Which means if you're developing an application, you can develop it faster with a lower cost base thereby receiving a faster return on investment as opposed to going out and buying a lot of CapEx based hardware, where you have a requirement and you're left with a legacy debt situation. And because your cost base would be higher, your ROI would be longer.
(AB): Right, that definitely makes sense. So we have costs, we have faster return on investment. What else should organisations consider?
(LG): Security and geography are very important. If we look at other areas such as storage needs and availability, those are huge considerations within the multi-cloud universe.
(AB): Let's look at each of those, kind of break them down a bit. And let's start with security. Can we dig in there a bit?
(LG): So, security. When you look at your data, not every system necessarily requires the same security requirements, especially when it comes to the compliance element. So for example, you may have personal identifiable information or medical data, those crown jewels that you want to keep particularly safe. Now you may choose to deploy those into an environment that has a higher security posture. That security posture costs money, so you only want to put inside those areas, those datasets that are relevant. Other systems, maybe you want to share some of the information with other business to business connectors with other organisations or external users that may have a lower security posture.
So identifying your security posture and identifying the cloud platform that is best suited to deliver that security posture is critical. With private cloud you get a good level of security, but the speed of innovation in private cloud may not be quite as fast as it is in public cloud where they have marketplaces where you can procure more services. And because of that speed of innovation, when new and emerging threats happen, actually reacting to these can often be faster in the public cloud than within the private cloud.
(AB): Got it. That makes sense. How might geography influence a multi-cloud strategy?
(LG): So, interestingly enough, I was on a virtual boardroom where we were discussing this around how organisations that operate globally have to comply with different rules and identifying your data and saying this data has this classification, whereas this data set has a different classification. Being able to architect your applications in such a way that your data is able to be stored in a particular geography and be compliant to the prevailing rules within that geography is critical for achieving compliance within that region.
But it also needs to be done in such a way that you don’t have lots and lots of instances of your application, which becomes very expensive. So you end up with your data being spread potentially all over the globe, which plays into the storage element, but equally, you've got to have that data protected, so that it is classified and available regardless of the prevailing compliance elements.
Also, furthermore, certain data sets you want to move closer to the user. So, geography is also critical if you're delivering web services. I mentioned Netflix at the top of the call. If I am consuming something here in the UK, it's much better if the data that I'm consuming is also in the UK because I will get a better experience than if that data is coming across the Atlantic over to here. So having distributed data sets that are closer to the end-user can provide better experiences.
(AB): That's great and you did mention storage a bit. So what about storage needs as well as availability? How do these factor into the decision making process?
(LG): Storage and availability are very closely aligned. When you've got a multi-cloud scenario, private clouds tend to be more geographically constrained, whereas public clouds, there tends to be many more data centres available to go and store that data in, giving you more options to move that data around. But equally, you want that data to be not just available by having high levels of availability of the actual storage, but if it goes wrong to have that recoverable, which means having other copies of that data, potentially immutable copies of that data stored, in other locations. So should one location become compromised, the data and the business systems can be recovered.
The advantage within the multi-cloud scenario is you can deliver, say your primary environment on the private cloud but have the data replicated and stored within the public cloud. So should the private cloud become available, it can be recovered to the public cloud and then once it's in the public cloud, it can be recovered to any other geography around the globe within the public cloud.
(AB): That's great. So one area we haven't talked about, Leon, is network connectivity. Does the network become even more important when you're using multiple clouds?
(LG): So this was one of the biggest barriers that has existed in the sort of evolution of cloud. Networks used to be very expensive, and the amount of bandwidth and latency that was available and the locations in which they were available were expensive to obtain and would take a long time to receive and they weren’t particularly agile. Things have moved on massively now and we have software-defined wide area networks that are available in practically every hyper scalar in the majority of data centres around the world, which means you can provision capacity incredibly cheaply. And because it's software-defined, we can do it incredibly quickly.
Many organisations will offer an SLA to implement a changing capacity in under one minute. And that means because of the ubiquitous nature of networks, it's become an enabler for multi-cloud because actually, if the network is incredibly cost-effective, low latency and available everywhere and resilient, all of a sudden it becomes the enabler, the backbone and the highways that connect with different cities, except in this metaphor, the cities are the clouds.
(AB): Yeah, that's great Leon. I have one last question for you. When choosing a multi-cloud strategy, how can organisations determine what clouds are right for them?
(LG): That is possibly the most important question. You need to go and understand your applications, your servers, your infrastructure, your estate. Start off with an assessment. What are the needs of an application and try to be application-centric within your view. You can look at these things from an infrastructure perspective: How much CPU and RAM am I using? How much CPU and RAM do I need? Procuring what you need is much more important than repeating what you already have. And then once you've assessed both your infrastructure and your application needs, the dependencies between those applications - the website might be dependant on an application server and a database server, but all of those may be dependant on some kind of authentication system, so collectively, they become a discrete system. And then once you identify these groupings of systems, these dependencies, you can start working out okay, well, which ones of these would be best served in different clouds? What are the applications, what is the security posture and the data classification of those systems? And therefore, should they be better served in one cloud or another? Do they have dynamic scalability requirements? As I just mentioned when we talked about web servers, maybe you're an organisation that runs campaigns or you are subject to seasonal loads that generate a lot more traffic and therefore the website needs to scale at a particular time. That may align it more toward a public cloud.
So assessing your needs is really, really important. But also, don't just assess your needs in terms of where you are today. Any cloud journey should think about the digital transformation that might be taking place in the future. Look at the long term business strategy. Are you just looking to do exactly the same thing you're doing today just cheaper, faster or more available? Or are you actually looking to develop a competitive advantage and do things better and smarter and deliver more value and new value to customers? When you start understanding the business context and the roadmap of the business and you understand where you are, you can start creating a gap analysis and how those different cloud platforms can enable that change.
(AB): This is fantastic Leon and just really incredibly helpful to our listeners.
92% of organisations use multi-cloud and for good reason. Each cloud environment offers different features, advantages and benefits. Adopting a multi-cloud strategy eliminates the one size fits all concept and provides organisations with choice making it easier to select the clouds that achieve their business goals and application requirements. So whether you care most about avoiding vendor lock-in, ensuring cost predictability or securing your crown jewels, multi-cloud gives you the flexibility to meet your needs and adapt as those needs change.
Leon, thanks so much for joining us today. We really appreciate it.
(LG): Oh, you're welcome anytime. I really enjoy these sessions.
(AB): Great. Leon Godwin is Principal Cloud Evangelist at Sungard AS.
You can find the show notes for this episode at SungardAS.com/ITAvailabilityNow.
Please subscribe to the show on your podcast platform of choice to get new episodes as soon as they’re available.
IT Availability Now is a production of Sungard Availability Services.
I’m your guest host, Alison Brooker, and until next time, stay available.