Expand – Fordway – Your Choices as we see them
Hi, what we’re gonna do is just take a quick run through the current state of IT infrastructure as we see it. When you think about infrastructure at the moment, this includes everything from in-house hosted and obviously cloud.
So what are we trying to achieve? Well I think nobody would deny that at this moment in time it is a lot more likely that public cloud use is gonna increase.
Now I don’t think anybody could possibly contradict that statement. There’s various learned analysts who said things. But effectively we’re looking at pretty much a 20 to 30 percent compound annual growth rate of the public cloud and I think alongside that we’d also all agree that the internal IT spend is going to reduce. As more and more services are migrated to the public cloud. But we must remember that the public cloud is actually a range of different types of services and capabilities and there are three core elements of the public cloud the first and most obvious one is software as a service SaaS. And really that’s where we see most of the growth is and also in our view where most of our customers are actually heading towards is to get better capabilities by renting the actual applications and services directly from the users – from the from providers. And this is typically your office 365’s, your G suite, your Salesforce; where there are ready-made solutions which you can customize and tune where necessary to your own requirements but give you that capability on tap and on demand. But the public cloud also includes platform as a service which is very very useful for developing your own applications and obviously infrastructure as a service where effectively you rent the capability to host your own product services and the existing applications on.
These definitions are quite important because realistically where you’re going to is where your help.
But let’s have a look first of all how ready you are to move to cloud. So we’ve run a slightly tongue-in-cheek but very quick test that gives you the opportunity to say is cloud gonna be suitable with my organization. So, yes please.
Well first of all do you need to do something with your current IT provision. Have you got to the point where you think we must do something. So we use a team in our organization Fordway called the compelling event. So this is something that said actually need to do something at about the way we provide our current IT service, and if so what do we need to do.
So it’s that rethink and for server infrastructure for applications for core services then typically we need to be looking at; well is cloud a suitable way and in our view it is an alternative way of delivering. It is not the only way but it’s a lot of cases, it’s a very very good way. But typically what you’re looking with cloud is to say; is there a way that we can use the capability of the cloud to provide a better IT service than we can in the way we do it at the moment?
There are organizations conversely that typically cloud may not be a very good answer at this moment in time and every decision has its appropriate time to be made. if you’ve just invested heavily or you have superb capability or if your IT services delivering everything that your organization needs for the right cost for the right capability, for the right performance – If your team is absolutely settled and capable of delivering anything your organization needs to deliver then what’s the point of moving the cloud because you’ve obviously running a superb IT service at the moment and long may it continue.
And you’ve probably cracked a lot of the answers the cloud that can help you help you address. So, in a generalist in a very very general view what we would normally say is that if there is a suitable SaaS service available, that is the right price, it has a right capability and is fit for purpose then there’s no real reason not to do it. What you’re giving yourself there is the what cloud is really supposed to deliver it’s just that instant capability that gives you that opportunity not to have to worry about any of the services – about making that application work.
Everything is done through the provider. Obviously, you still need connectivity and security into those services but once those fairly basic elements are cracked, you’ve got what you need. PaaS, Platform as a service, is also extremely useful if you’re trying to develop digital assets or create applications or Digital Services that allow you to create value for your organization. So typically, Wonderful if you want to develop new web presences if you want to develop mobile applications, if you want to scale up a worldwide service, so if you’re delivering content. But the one that we’re saying actually are you absolutely sure you want to go to? Is infrastructure-as-a-service because what an infrastructure service from a cloud provider gives you is that platform which basically is the hardware and a hyper biter. Which perfectly honestly isn’t actually that difficult to provide running in your own Data Center and keep running. What do you want to do? How do you want to do it? Where are you planning to go now? And why how are you planning to get there? and this is really where organizations like Fordway you can definitely help because we can come and say well actually what is the most appropriate way for you to deliver the service.
So, in the mix, there’s a few new technologies coming along. So, the first technology that is really changing in our world is networking. Networking typically boxes linked together and each of those boxes had its own capability. You had to work out how to connect them, how to integrate them and then if connected correctly they routed traffic to the correct place, but each device had its own intelligence and the intelligence was typically embedded in the device, in the control plane in that device. One of the key things that virtualization has brought, both initially for servers and more recently for desktops and now heading towards networking, is that ability to take that capability away from each physical device and actually run it has a policy for all the network traffic running between those devices. So, if you think most of us will use an SSL VPN to connect into secure services, in fact, most of the major web providers now mandate that you use SSL to connect. Now that creates a secure encrypted tunnel between your endpoints and the service or website that you’re accessing or consuming. SDN, effectively does that but allows you to create policy-based tunnels between any endpoint and any service that you can manage and control centrally which is in principle a really really good idea.
In practice it is pretty complicated and obviously the key thing is defining how various systems and endpoints talk to each other, how they’re allowed to communicate, and you need to define this and also once you come out of your control software and divine networking environment how that then interconnects with somebody else’s environment, so there’s got to be interoperability. However, it brings major benefits because once you’ve taken that capability of each individual device, you can plan an architect Networks. From a core infrastructure and the boxes as with server virtualization just become the hosting platform, the carrier for the service but the service can be defined in the service can be changed dynamically. You can also add new capabilities into the service is very easy because you don’t have to upgrade 50 boxes or 50 Routers or 55 Firewalls to allow that service to happen. You simply upgrade the software capability and create the policy between all the end points and allow it to be passed through all the intermediate points in the network. It also allows you to provide much better visibility because you can then review the traffic from a single central collector Point without having to forward traffic from lots of different devices. So potentially a really good thing. In practice extremely extremely difficult to implement now but becoming easier. It’s coming in with hypervisors, it’s coming in with the providers. So, these days were starting to see more of a hybrid where the capability to run the software is on top of the existing Network infrastructure. But ultimately, we should be able to buy a white box switch in the same way as you buy a white box server today and run our choice of software across it which will deal with all the networking connectivity.
So, networking. It’s changing. It’s always complicated. Also, inside our data centers we have the cloud. It’s available. It’s there. It’s ready to its fit for purpose. But again, a lot of people need or want to continue to run Services in-house and what we can do is build our own clouds inside our own data centers without having to rent the capacity on demand from other providers and if you’re doing this at some sort of scale or if you’ve got security requirements then. It’s still a perfectly valid option. What we’ve done is we’ve converged between infrastructure as it existed several years ago with a typical SAN and server architecture. That’s now become converge into converge units which can run both the storage and the hypervisors to provide the servers you need. One thing we would say is you don’t necessarily have to have a very expensive hyperconvergence product provider, providing that service. Simplistically most of the core operating systems you can buy, be that VMware be that windows or be that Linux actually have that capability. So, if you take your industry standard 2 or 4 socket server with a decent amount of storage and if you put on a copy of Microsoft Windows you can then add system center virtual machine manager as your pack and you can effectively build your own private Cloud inside your own data center likewise with VMware. It has all the components and capabilities. You need to build that hyper-converged unit without paying for a Nutanix or a Simplivity extra cost to run that if you’ve chosen purely to run Windows or you chose to run VMware and likewise with red hat or any of the major Linux distributions. Inside your distribution of red hat SUSE, Ubuntu or whatever there is that capability that allows you to create that hyper-converged unit which will then allow you to run your choice of a Windows or Linux server in that environment?
The key thing with this is if you’re building effectively your own cloud inside your data center because the way the vendors are going once you’ve got your Cloud you can then start integrating it with a public cloud and this is how it fits into the picture whether the infrastructure layer is just the basic facilitation layer you now have choice and through that choice. You can then choose to potentially to move workloads. Where the most appropriate. So, your development environment may well be in the cloud because the cloud is very well suited to an infrastructure service or platform-as-a-service to running inside development environment. But once you know what you need to run and how you need to run it. You may find that if you have good existing capabilities and facilities it’s cheaper to run it on your own infrastructure and your production systems conversely the cloud can offer you the ability, for example, of temporary environments such as a disaster recovery environment to provide a disaster recovery service or potential long-term archiving service much more cheaply. So, what that all gives you is this option to say, you don’t have to pay someone else to host it you could will be doing it very effectively yourself. But be warned whichever hypervisor you choose is going to impact your choice of cloud. You do need to watch this. Okay, if you run Microsoft Hyper-V and you choose to put as your pack or even start moving towards as your stack then realistically your best and maybe most favorable and easiest integration options is going to be into a Azure and Microsoft with licensing policies and the Technologies are trying to make that happen, and is almost certainly in your interest and realistically there isn’t a huge amount of difference in pricing between the major Cloud providers because what you choose on one, they’re going to provide that capability in a similar fashion and comparably a similar price. So, Microsoft, if you’re with Microsoft chances Azure looks like a pretty good option for you. VMware is starting to look a little bit like an orphan like they have this tactical alliance with Amazon web services, it’s certainly capable but, actually what it is at heart is, is effectively you renting some AWS Hardware inside a Data Center and as a minimum configuration. So, unless you’re looking at typically 40 or 50 virtual servers or greater, it doesn’t make any sense of the moment because you have to rent physical Hardware nodes to run the VMO underneath the Amazon Cloud environment. It’s not that cheap it certainly only really worth looking at if you have scale if you have a large investment in VMware, and if you’re prepared to look at making a long-term commitment one thing there is in the VMware world is there are several of the smaller, sort of less well-known Cloud providers who actually run VMware under the V-Cloud air program.
So, maybe looking outside the big three as we call them the Amazons the Googles and the Microsofts and saying do you want to go and look at a smaller more local provider? And if you’re an organization without sort of global reach and scale it may be that one of them smaller local providers actually provide a better option and we’ve been playing around with Linux for some time now. Realistically, it can deliver the service, but you’ve got to do most of the work yourself. With both VMware and Microsoft, you get a ready-made operating system that works in a certain way with all the support you need if you can’t make it work properly yourselves, and if something goes wrong. With a Linux distribution, you do get a level of that but primarily the tools are much less sophisticated and much less well-developed than if you were working with one of the major providers, but it costs an awful lot less. So, if you know cost is the overriding concern and time is not one of the biggest issues on the iron triangle between cost quality and speed then Linux might well be a viable option and certainly, that’s what most of the service providers have chosen to do. With OpenStack there is now the capability with OpenStack to integrate an in-house OpenStack environment very easily with public Cloud environments. So, both AWS and Google support and OpenStack integration and even Microsoft are now starting to become capable of integrating within the open site. One thing about the Linux environment I would say is it is at the moment probably more geared towards containers because all the development work around Docker and Kubernetes primarily is starting on the Linux platform and is now starting to become available both in Windows servers being there for a couple of years now and also with VMWare.
So, once you’ve chosen your platform public Cloud, convergence, infrastructure, running whatever hypervisor of your choice and it really doesn’t matter all are good options. It’s just helping you work out which the best one is you’ve got to look at if you have gone to Cloud what does that actually mean? Well, a lot of our customers are seeing already this concert we call Cloud shot, which a lot of people have talked about. I’m Cloud shot basically is we’ve moved it over to the public cloud and public cloud is cheap and then you get the bill and you go ” oh, hang on a second. That’s not quite what I was expecting” so public Cloud infrastructure as a service has a number of characteristics that you need to be aware of if you’re choosing to move towards it. First of all is that everything is metered and it’s very finely granular so that you might buy a virtual server but to make it work in the AWS world you actually need five other services storage, IP addresses, potentially domain name Services, management monitoring Etc. to actually make that into a workable server. Also, in the infrastructure-as-a-service world, you are always moving into AWS’ way of doing it or Microsoft way of doing it because they have designed these Services specifically to work in a certain way. If you need a particular type of storage, you have got to hope or check that that provider offers that type of storage. Now let’s be perfectly honest they are working very hard and they are very capable to make sure that most requirements are going to be catered for but you have to design your infrastructure in the way that the cloud provider States you need to design that infrastructure and do it in a certain way. So, you’re limiting your choice is potentially and also as I say you have got this issue where you are running potentially many many different metered services and some of these Services might be just a few pennies per instance or per month. There might be a large amount of traffic and one of the key things we need to look at within the public Cloud world is data transfers particularly out of the cloud provider.
So most of the cloud providers will happily let you put data into their Cloud for free, that’s what they’re there for. They would like to get you on their platform and I would like to retain you. When you’re taking data out or moving data between Cloud providers or even between availability to Mains within Cloud providers that is again metered. So you need to understand how your application behaves whether there is keep alive or significant data transfers between servers, between domains and one thing, is it’s not actually that cheap. So, this is admittedly a fairly large instant size. But you know, this is typically something that might be a SQL Server. So, in AWS you have a decent-sized database server, you know, number of cores, decent log of ram and a bit of storage. You know, that’s a few hundred pounds a month just for a standalone server without backup and before you start putting resilience in and one thing I would say is if it’s infrastructure service on the public Cloud. All they do is provide the hypervisor. Everything above the operating system there is your responsibility. You don’t get patching as part of the service. So typically, if you want multi-domain resilience between two availability zones inside the Amazon world that single server that starts off at 5-600 pounds a month could potentially be ending up costing you 16-1700 pounds a month. In Azure, because it’s architected differently, the costs are comparable but done in a slightly different way because they actually give you a slightly higher level of Core resilience as part of their basic service in which case you do pay that little bit more. But as soon as you start running High availability or multiple instances clustered between availability zones again, the same issue arises.
So, clouds cheap is it? Well, Fordway runs a private cloud service that includes patching. So, for just slightly over two-thirds of the cost of running on AWS per month and these services are unmetered. This is a fixed cost per month irrespective of how much you use it now granted if you were able to bring those servers up and down on demand and only use them for say 50% of the month the cost would decrease but who brings their core production database servers up and down on regular basis? I mean we work with hundreds if not thousands of organizations and I know very few of them who would even contemplate bringing their core Services down over the weekend if they didn’t absolutely have to to make some major infrastructure change. So, do you really want that metered service for your core it? It’s great for temporary. It’s great when you can put automation it and that you can bring Services up and down. But if that service absolutely has to be there all the time every day potentially is going to look a little bit more expensive than you might have thought and other options exist. We would suggest looking at private Cloud providers and Fordway categorically one of them might be a better option because we also provide a lot of added value services in with the basic instance cost.
So, how do we make sure we do it well, Fordway work with our customers in a particular fashion. We need to understand what you’re trying to achieve. So, with a public cloud provider, they give you a menu you pick off the menu what they want. That works if you know exactly what you want and how it’s going to work. But let’s say that this is the first time you’re going into the cloud world. Do you really understand what you’re going into and why you’re going into it and what is going to look like? Well, we do because not only do we run our own cloud services. We also work very closely with the major Cloud providers and we understand. So, the first part of our process is to engage with you to help say can we actually help and be useful to help you get a better result faster. If we can if we believe were aligned and we can help we’ll then work with you to work out if you are trying to make that move into another environment that you’re going into for the first time or maybe even the second or third time that can we help you understand exactly how that environment needs to work and what it might cost so we can then analyze and say well here’s your options. By doing those options we can help to refine down to let’s make this the best way of doing it that we possibly can. Once you’ve done that analysis phase, the key thing is from the options analyses maybe we say well this is what we think we should do let us go because typically you’re going to have presented a business case to make some sort of change that is going to have cost or risk or impact on your organization to allow you to do that. So, should you do that? You say yes, once it’s approved Fordway unlike most consultants, our work is not done at that point. That’s only the start of the journey we can help you with because you want to make that change, well typically most organizations need to make these changes fairly infrequently, but we’re an organization that was set up to deliver and love doing change.
We have teams of people specifically to help organizations Implement change successfully. So, the first part onto case improved is how do we plan to ensure that this change goes in successfully using all our expertise and skills we’ve gained over 27 years of existence to actually help you put that change in place? Once we’ve done the planning then it’s bringing the team of experts and Specialists to sit alongside your teams to actually make that change happen and that provides you again with this thing. This mantra we keep saying ‘the better result, faster’. The key thing is help, you get to where you want to be faster than you could on your own and the cost of using our services typically will be considerably less than the extended time it will take for you to do this on your own. And lastly in the most important part is we really want to help you realize the benefits of making that change because the reason you’re doing that change is to get some business benefit from doing so maybe reduce cost may be more capability whatever but until it’s actually in and working and actually delivering the benefits for you, it hasn’t realized any benefit, it hasn’t proved it’s worth. What we can also do is say well actually there may be elements of your IT service that you choose not to run in house or you don’t feel they have the capability to deliver successfully. So, therefore, we can help you by taking responsibility for delivering those elements to ensure that it continues to work. So, it’s a whole life cycle both of change and delivery that gives you that confidence and peace of mind to say. Well, we’re here to help you achieve your goals. And as we say help you achieve that better results, faster.
So, thank you very much. Hopefully, that’s been a quick overview.