Welcome to the Fasthosts ProActive Podcast: Spill the IT. Each episode, we'll sit down with some of the amazing ProActive team and chat through their experiences of the ups and downs of IT infrastructure management in small businesses. There's always plenty to chat about.
Shadow IT - our experts discuss the impact of network landscapes that aren’t what you thought they were on your infrastructure and your TCO. Clue: you’ll need more whiteboards!
Listen on your favourite platform!
Want to listen on your go-to platform? We're on those too...
Welcome to the Fasthosts ProActive Podcast: Spill the IT. Each episode we'll sit down with some of the amazing ProActive team and chat through their experiences of the ups and downs of IT infrastructure management in small businesses. There's always plenty to chat about.
Welcome back everyone, for part two of the Fasthosts ProActive Podcast on total cost of ownership. When we left you last, we carried on the conversation as there was just too much to talk about. So, on our recording we have Gary, who's the solutions consultant of Fasthosts ProActive, and Dan, who is the senior service owner here as well. Sit back and join us as we continue our conversation where we start to look at where people are going wrong when it comes to analyzing costs and impact to the infrastructure of their IT setup.
So, with so many applications being added to IT infrastructures and the management and the cost of these being so difficult to keep on top of, where is the industry of cloud computing going when it comes to the transparency of applications and in fact, if they are even being used? Is it all just about better reporting?
I think there's a term that we certainly use in the office, which is shadow IT, and I think there is a subset of that that is shadow apps. You might know what your infrastructure is and you might have good asset management and good control over that, but that doesn't mean that the software vendors, the developers, whoever it might be, are going in and other things are being installed, other things are being used, and it's difficult. It's a commonly used term because it happens pervasively, overall, different organizations, certainly different organizations that I've worked in through to smaller ones, up to large government ones. Somebody's just installed something because they thought it was going to be useful. How do you keep control of that?
Coming back to your point, is it purely reporting? I don't think it is purely reporting. It is some element of control and having the control all of those different layers. It's not just control of, somebody can walk into a data center, stick a bit of hardware in a rack and network it up. It's also about controls into those environments as well. One of the benefits of going with a managed service provider is somebody else is managing your software and the control of your infrastructure on your behalf. So with tighter control, you get the benefits of knowing exactly what's running on your infrastructure and you're tightening those patio doors as we were talking about, to stopping other people just coming in and just installing something because they think it's going to be useful.
I guess it's a second pair of eyes, isn't it? So I guess it's almost like, yeah, we know what we've got going on, we know what we got in our infrastructure, but it's almost like somebody coming along and saying, hang on, we've been having a look at this, and what about this, this and this? You don't seem to be getting the best out of this. Is this really needed? Is that what you're really saying? Is it where there's a deeper sort of element of analysis going on in the background that people are reassured by?
I think there's two elements to that. I think from the point of shallow IT, we often think about some of the examples that Dan's given, where a user will think, oh, this would be really handy, and they go ahead and install that and you magnify that on a mass scale and all of a sudden you'd have stuff running across your network that you weren't aware of, possibly [inaudible 00:03:39].
Big cost going out.
And yeah, it's... The number of organizations that do network scans and then suddenly realize that TikTok is their largest consumer of bandwidth on their entire network. And this happens a lot. For me, there's a more challenging aspect of that and that's the shadow IT from within infrastructure departments. Now, for the most part, without generalizing, IT professionals are pretty creative individuals, and I use that term with a grin, but I've seen examples where organizations have come to end of life, a server for example, have a good, in inverted commas, understanding of the function of that item. It goes offline and then seven other things break.
And that's because somebody somewhere went, we need a thing, they've installed it on the server, there's no record of that. And all of a sudden you find that things break. Now, where we can prevent that is that we manage the configuration for all of our infrastructure items. It's not just a case of, hop on the server, install service X and away you go, it sits away running in the background. We define all of that configuration in code so it doesn't just slip onto the server, it's auditable, it's well defined and well understood. So that if you have to take a piece of infrastructure down, you know exactly what functions that infrastructure is and [inaudible 00:05:04].
And what it's going to affect, I guess. Yeah, yeah, yeah.
... knock on effect. In the same way we were talking about observability. We go into, I suppose really as deep as the solution demands, into analyzing what's running, how is it performing? We have a fleet of metrics that we can collect, which will enable us to provide insights into all of our solutions, whether that's from a very fundamental, do you have enough disc space, level to, how is your web server performing? What queries are taking time on your database? So we can go in at quite a fine level and help analyze what's running on the solutions.
So where are people going wrong when it comes to assessing their own costs? What are you seeing? Where are people going wrong within their own businesses? Are they too reliant on their service providers?
I think for me, the top one, it's the one that always used to make me groan as an IT manager, was, number one is understanding your assets. Especially a lot of the security functions all start with understanding what you've got first. And not many IT managers are lucky enough to have, as Gary said earlier, kind of virgin playgrounds to start with, greenfield sites. You've generally inherited something. So in terms of assessing your own costs, put the time aside to sit down, understand what your assets are and speak to people. Don't just rely on your CMDB or however else you're managing your infrastructure. So that would be my big kind of where people are going wrong first.
So a big whiteboard, get it all on there.
Get everybody in the room and say, let's look at this, yeah.
Bring out your dead, yeah.
And I think there's another aspect to that, isn't there? There's the intangible costs. How much are you spending to operate this function? How do you measure the amount of effort? Most people when they are resourcing IT departments will have a certain set of functions in mind, the more common ones. So you know, doing your updates, deploying new things, et cetera. What's rarely thought about is, what happens if something goes wrong? All of a sudden these costs can spiral. You can't really plan for them particularly well.
Again, I've seen issues where internal IT department resource has been consumed due to an unexpected issue to a point where it had a wider impact on the entire roadmap for the following 12 months. The other concern around that is burn-out. When we're talking about resources, and certainly in IT, we quite often find it easy to forget there's a difference between machine resource and people resource. And whilst our highly available clusters can run 24/7, 365, at full capacity, most people can't. And what if? How do you resource for worst case scenarios without putting your team at risk?
People just aren't sleeping anymore, are they?
I think go going back a couple of years to the beginning of the pandemic, I don't know many IT people who were doing less than 14 to 16 hour days for extended periods of time, because there was unprecedented demand that people couldn't resource for. So the danger is when working at that kind of capacity is suddenly seen as both possible and repeatable.
Yeah, sure. So looking at that in a wider way, so what's driving costs up right now? So we've seen utilities obviously go through the roof, increasing salaries, because we do have skill shortages and probably that's another podcast for us, to talk about skill shortages and the issues around that. But that's massive right now. And so salaries being pushed up. We've obviously got the demand and supply of hardware, so that's obviously increased costs as well. And maybe, dare I say it, bad management of infrastructure. So you see they're all the things that intrinsically are pushing up costs within organizations.
Yeah, completely. And that happens at many layers, right?
Not only are individual businesses' costs going up, everyone's paying more for their heating, for their electricity. As you mentioned, staff costs have increased. But then so have all the vendors'. They all have those increases in costs, and guess what? That gets passed on as well. So not only are you paying more for your heating, you're paying more for your vendors' heating or cooling. So yeah, it's massively pervasive.
So I remember when I was an IT manager back in the day, and this was before the current cost of living crisis, and you try and manage and understand what your opex costs are. And I remember I had a certain network provider whose name I won't mention. For me it was keeping on track of your deals. So you thought you got a... You know, you signed up for a five year deal, they were going to provide dark fiber from A to B, and either you've inherited this deal or you've set this deal up, and five years down the line you're now where you are. And you've got to come back and look at those contracts that you've got in place, because while those deals were brilliant for the first three, five, whatever it is, years, you're coming out of that period now and the prices jump sharply.
Understand your terms and conditions, understand when it's the right time to renegotiate as well. And I remember having conversations with, again a certain provider and saying, "We're five years down the line, I want to renegotiate." And them saying, "You've lost that window of renegotiation. Your contract's rolled over, you're on a yearly deal now." And the price jump was quite significant for the organization I was in. You've got to be strong, you've got to barter with them, you've got to work closely with your account manager, your TAM, your client success manager, whoever it might be, so they understand how important that particular element is to your business and so that they can work with you to find the best ongoing solution or the best ongoing deal.
An interesting point you make, and just one additional question that comes to my mind is, so how far out should you be looking at those costs at the end of your contract? Are you looking at it six months-
Six months, yeah.
Yeah, for me it was always the six month period. I remember having a massive great big whiteboard in my office after I learned this lesson, having those significant contracts. It's your networking, your compute, your storage, mapped out and you'd have those dates on my whiteboard. So I knew six months before, this is when I need to start.
Set your calendar alarms.
Yeah, yeah, definitely.
What meeting am I supposed to be in? Oh my god, no, I've got to renegotiate.
It's sometimes the simplest of solutions, right.
So let's talk about something a little bit differently now, shall we? So there are some horror stories around the cost of outages, especially in the online retail sector, reporting millions of pounds are being lost due to outages. Surely with cloud and moving away on-prem providers, this is one less headache and something that mitigates risk and loss for the business.
I don't know if it is one less headache. It's just a headache in a different space. But what cloud does provide is some unique solutions to help mitigate and provide the paracetamol around some of those headaches essentially.
Oh, I like that. Paracetamol. Yeah. Okay.
So I mean it's things like redundancy, it's things like alerting and monitoring. It's about preventative maintenance, all of those things that you can do much more easily in the cloud that would just be a headache to do kind of on-prem. So I don't want people to think moving to the cloud's going to solve all of your solutions, but moving to the cloud does provide some benefits and some mitigations that really makes those headaches small little niggles rather than migraines, essentially, to continue that analogy.
You can make things as available as however much you're willing to spend. So if you need 24/7, 365, without fail, then there are ways to achieve that. They might not be particularly cost effective. Now where I see cloud providers, like I said, being able to assist with that is, we can manage some of those costs. So your average organization isn't going to want to pay and maintain a alternative form of power generation, for example. And electricity supplies do fail. I've been in an organization where one feed into the building failed. Whilst the people from the electric company were repairing that, the secondary feed blew, and at that point everybody went home.
Now bearing in mind that was in an organization where a 15 minute period of downtime cost several tens of thousand pounds, when you're talking about an entire afternoon because you've got no power, that gets really expensive. Similarly, with multiple network redundancy, most organizations running on-prem will have one or two connections into a building. There are still single points of failure within those type of solutions. With cloud and certainly with what we can provide, we can pay for that alternative power generation source. We can have those multiple layers of resiliency that an individual organization can't possibly afford. So you kind of get those benefits without having that massive, massive overhead of having to run that level of redundancy yourself.
And it's only going to become more intense, isn't it? As we continue and the demand to be online and to be up and always be able to be interacted with, it's only going to get more intense, isn't it? I mean, we're talking about this today. Let's think about this in another five years time, 10 years time, it's only going to become more and more important to organizations, isn't it? So I guess I wonder what that future state will look like. Do you think the future state will be all cloud? Do you think it will be a hybrid, do you think... Where do we think that future state's going to be? It just pops into my head a little bit. Where's it all going?
I always say that business IT is driven by consumer IT. And what I mean by that is, people get used to being able to perform certain tasks in their personal life and then have an expectation that things can be done in a similarly easy fashion in their working environment.
The Amazon experience almost. Yeah, yeah, yeah.
Exactly that. And I'm guilty of it myself. If I look at a store or online thing, I'm like, "I want it tomorrow." If I find vendors that don't have that kind of service, I'm not saying I wouldn't necessarily use them, part of me is like, "Oh, what, I have to wait three days for my thing?" But it's the same with IT functions. If I want my email, for example, I can use my phone, I can use a laptop, I can use a tablet, I can use somebody else's device. If I need to, if my phone explodes, I can borrow somebody else's, sign into my email account and I've got that access. Why can't I have that with my business data? And that's where there are a number of challenges around that. So I think limits on on-premise solutions, limits on availability on-premise or self-hosted solutions get in the way there, but it's massively jarring when you can do something that becomes then part of daily life that you then can't achieve in a business environment. It's very difficult to try and see, why can't I do this?
So it's more around human expectations, I guess is what you're saying?
And that demand to always be available. Always, always.
Business has changed completely. We all work different hours, we all work in different locations. What was seen as an option a little while ago now is absolutely 100% mandatory as an employer. You've got to offer people that level of flexibility. We're going to talk about skill shortages a little bit down the line, but if you are not offering people the opportunity to work in different locations at times that fits around their lives, you are going to lose staff or fail to recruit them. And so that, I think, the human element, is one of the big, big drivers. And I genuinely think that in the coming years we won't have on-premise equipment, because at the moment we don't always have on-premise people.
Oh, what a good phrase. I think that could be an actual title for another podcast.
So total cost of ownership, do you look at it more over time, over a period of time, or is it more short scale? Is it immediately immediate, you are looking at that total cost of operations and ownership in a short term, say six months? Or are you looking at that over a 12, 24 month period? And are you looking to reduce those costs and see those costs coming down perhaps?
I think there's a really easy answer to that question, which is yes.
Next question. On we go.
It depends on how much data you've got. You can use what's available to you to make informed decisions. If you don't have 24 month’s worth of data, looking longer term is impossible unless you have that crystal ball. What is important, however, is that agility. So if you don't have that long term data, or even if you do have that long term data, but you are rightly aware that things can change at the drop of a hat... For me, it's about having the agility to change the game very quickly with minimal input.
So when we're looking at inflationary pressures, so strong right now, Dan, what do you think about the current environment? So we look at the current environment of MSPs. Is it pretty stable? Are we going to see MSPs faltering? And is that a big concern to customers, do you think? Are they going to be worried about that?
As a consumer of MSP services, you've got to be concerned about how long that MSP can support you and can be... As we talk about them here, Fasthosts Proactive should be an extension to your team. So I think for some MSPs, they're going to struggle because they're not going to be able to maintain the pace and the growth for the businesses that they're supporting. Or to be fair, maybe they've been bought in for one particular project and task and naturally the two will diverge. MSPs move into other divergent markets. They might be going down all the desktop IT route. But also the businesses may kind of outgrowing them.
And it goes partly back to your kind of question that Gary answered just so brilliantly a minute ago. It depends through which lens you're looking. If you're an IT manager, IT director, and you've been tasked with setting up the IT strategy for the next five years for your organization, that's a very different proposition than, oh, I've got a contract, it's expiring, and I need to work on some specific compute storage problem that I've got here and now this week. So to echo Gary's brilliant yes answer, it depends, right. So I think there will be casualties, there will be MSPs that specialize as well and provide really great niche services. So it depends on what you need.
And I guess it's, when we talk about lens, it's a little bit like when you go to the opticians. It's lenses, isn't it, it's putting different lenses over the front of other lenses and what are you looking at that particular time, which I think is probably quite challenging. We're almost coming to the end now. So if we're talking about that and the possibility of MSPs not being so stable or being more specialist, I just want to just cover off on the cost of migration and the business disruption of that. I mean, obviously if your business is going away really nicely and then all of a sudden you've got to swap, you've got to change, you've got to go now, oh my god, we've got to put everything over here, that surely is going to have a big disruption to business. It's got a massive cost associated to it. Are people sort of worried about that?
Yeah, they are. And rightly so. We've already spoken about that desire to offer that on-demand service to clients of clients and so on. Nobody can really afford downtime. Again, depends on what the function is and what your requirements are, but it is a concern. But I think from my perspective, again, it comes down to distilling what is the function? What are we doing with the data that we're moving? When can we not do without that? Are there solutions where we can run in parallel? And there are technologies that exist like replication where it is possible to run in more than one location at the same time. And I think it really comes down to how complex and again, how much cost is going to be involved in architecting a solution that has no downtime or minimal downtime. You don't have to have disruption.
Sometimes that comes with spending a bit more money. And I think it's about understanding all of the costs. Is the cost of the downtime more than spending a little bit more on a migration solution that prevents that downtime? If the answer is yes, happy days. So it's about understanding, it's about digging beyond the IT. It's about understanding the business needs, because the IT is just a thing that helps us do other things, and what really matters is, what are the needs of the business? And if you can get a good understanding of that, then designing the IT that supports that's... I won't say easy because then I'm selling my awesome talent short, but-
No, don't ever say your job's easy, please.
It does become a lot more simple and you know that you're working from a place where you've got that understanding and education.
Yeah yeah. Dan?
Understanding was the word I'd written down on my sheet here before Gary said it. Just stolen my-
I actually saw you write that down.
Did you? Brilliant. So definitely stolen my thunder there. They don't have to be worried.
Will they be worried? Course they'll be worried. These people are the accountables, responsibles, within the business, so naturally they will be. Do they need to be? No. Gary's spot on the money there. It comes down to understanding what a migration or a change might actually mean for your business. You know your appetite for risk or for downtime, and you work that out accordingly.
Well, guys, that's been a fascinating podcast. We probably could sit here for about another half an hour, I guess, and talk through some of these subjects. So it brings to a close this month's Fasthosts ProActive podcast on total cost of ownership. Next time, if everyone's feeling great about managing their own on-prem setup or actually sold on migrating to the cloud, the next burning topic must be security, and so we're going to give that some airtime. There's a great deal to consider around about security. Dan, Gary, you've been great as usual. Thank you very much for your input today. And for everybody listening, have a great month.
Thank you for listening. We hope you enjoyed this episode. You can subscribe on Spotify or Apple Podcast, or visit proactive.fasthost.co uk for more info. See you next time.