DevOps and the Economic Downturn – David Williams, Quali

This is texturing TV. Hey guys. Thanks for the throw.

We're here with David Williams a senior vice president from Market strategy for quality. And we're going to be talking about the impact the economic downturn might be having on devops teams and one of the implications there of David. Welcome the show.

Thank you Mike. It's good to be here. We've seen a lot of focus, of course on developer productivity late, but the downturn kind of focuses the mind a little bit so I don't want to go and say it's a full blown recession yet, but it would seem that as we go along here.

There's gonna be a lot more interesting automating processes because we do need developers to be more productive. So what does that look like? Well, you see when when devops came around the the world was a different place.

So as you explain to, you know, we're looking up two things that are happening one is obviously a tightening of belts in regards to Investments and how Devil's practices are being implemented but that's being also driven by the fact that obviously the economies having a bit of a traveling time depending on what day of the week you wake up on but also the maturity of devops has grown significantly over the last five six years, I'd say that it's become mainstream. It's very unusual to do soft. Development without a devops type approach going on and ironically when devops was an egg and it was still being developed.

It really didn't have things like cost management as being a an inhibitor to execution. So as I say it is growing up and devops generally has become more accountable to the to outcomes and the bottom lines. That means that unfortunately things have got to change slightly in respect to the governance.

It's got to be implemented to make sure that devops is, you know, delivering the value that it needs to and is accountable. So to say it in a crisis things come to a point and as investment starts to be tightened up and people are looking at where to invest first second and third they need to understand what's going on. How do you go about doing that?

Because frankly when times were good developers were consuming Cloud resources, especially something that kind of looked and felt maybe like drunken Sailors on a Saturday night. But if that's the approach that had been taken how do I kind of instill a little bit of discipline without completely freaking everybody out? Well, you see this is a it's a it's a mixture of things and when you talk to the analyst I'll talk it's a mixture of processes and organizational design and as well as the technology that's applied to it.

But the end of the day when it comes to accountability to the bottom line budgets and cost management is that top down type of practice devops and the implementation around agile practices have been a Bottoms Up type of practice. So soon or later those things were going to have a collision and that's what's happening now. So what's what's really being approached by a lot of companies is that special when they're adopting infrastructure as code practices which which enable you to you know, A extremely good ability to provision Cloud infrastructure as part of the code development.

It enables lots of I'd say more sophisticated activities to be done by people who have software development as their primary objective. So infrastructure has become more as fate accountable to the business. How much Cloud am I using why when and how and when you're actually got your developers who can provision and spin up resources very very quickly using a piece of code then obviously that piece of code need now needs to be a little bit more accountable to you know, who owns it what it's doing and how how it can get prioritized against all the other things that are provisioning Cloud sources so you can budget accordingly, but that's you know that with that in in mind, you know, the the accountability to IAC or infrastructure as code or even you doing your configuration technologies that go out there and provision infrastructure.

There needs to be an ability to be accountable to what you're using and also be guided in how it's being used. So that means that you know, you can't measure what you can't see and now the business is saying we need to see what's being done. They're not going to inhibit development that's gonna cripple Innovation.

And that would be credit. You know, that would be terrible. So what they're really doing is saying we just need a weight with which to allow our development organization to innovate but without but with accountability to the bottom line and the business budgets and that is where the Crux is happening at this very moment in time.

That's where we are today. When I talk to finance people, it's not always so much about they want to reduce costs. That's a nice to have it but really it's about a predictability of costs.

And yeah, the thing that they get freaked out about is when there's a monthly bill that pops up they weren't expecting. Um, so is that part of the discipline is just making people understand that, you know, we need a regular Cadence and cost not so much that we want to drive 50% out of the budget per say. Yeah, you just want to understand what's being spent when and where yeah, I mean, it's understanding.

I mean, I don't think anybody would be saying, oh we're gonna reduce our Cloud spend by significant amount. That's not gonna be happening. In fact, if anything I would imagine clouds spend is going to increase as the need to reduce costs and they the scalability that that provides, you know lends itself, but you know, we're moving in a different world now, so we come from a world of tangible from a financially become aware from a tangible.

So tangible to things you can touch assets and things that have a depreciation life cycle, so you Exactly what you're buying and it's got an Roi and you know, you look at it. Well, unfortunately that's not how things are provisioned today. Everything's intangible.

You don't own what you use you lease it from one to another term and you use it for the time you need it and it goes away. So not only is it not yours, but it's temporary. So even if you're using in production, the cloud providers will move things around in their own, you know to optimize their own resources.

So optimization requires you to have a plan you need to understand what your spend is going to be and I think what we've done for the last say the last half a decade at least is we've got a good grip on what we've spent. So everybody does a sort of like a cat her again exercise where they're trying to gather all the other costs and trying to work out what the value was and now it's got to a point where the business is saying before you provision things. And before you do things we need to understand one who's using it why they're using it and when is it going to be used and how much resources are you going to be using?

So it means Technology that has to be associated with provisioning has to take much more of an accountability role in regards to the ownership and the allocation of resources in a smarter type of way. Can I capture that data in real time? Because part of the issue is by the time I get a bill it's too late to do anything about it.

So how can get in front of these things? Well, you know, they there's loads of way that they're organizations are looking at it. First of all them there is less entrepreneurial thing as you mentioned earlier on, you know, when it was the little more the wild west so a whole bunch of different teams or individuals within teams could provision what they liked in support of an outcome that they themselves needed that understood the value of now what's happening is governance is coming in much more into place.

So configurations and models are coming in. So instead of actually creating something from a fresh every single time. It's leveraging existing known States.

So it's having much more accountability to historical record of what's being used before and then extrapolate and that need going forward but also being much more reactive and being more accountable to all the different and nuances that that will be using the cloud. So it's different teams use the cloud that's got to be captured but also extrapolated so they understand what it would be going forward as well as what it's going back which is the difference I think today because up until recently Very much. It was what is cloud costing you and I need to account that to a cost of some description based upon a value and then I will work out whether that was prioritized accordingly, but now you've got to work out.

Okay. What are you using? How are you using it?

And you've got more planning going into it. So IAC and different, you know, the sort of things that you are used to provision the cloud. I've got to be much more accountable to an outcome.

So you've got to look at the value that you're achieving and then tie that back to what it takes to achieve it. So it's a little bit of a more God rails more governance. But again being very careful not to inhibit Innovation.

It's a very delicate Balancing Act to do to say use what you need to get what we need done but use it within reason and optimize things much more effectively. So everybody's not Reinventing the wheel create models, make sure we're leveraging known State type of infrastructure. And so it's just getting up.

It's getting smarter as they're Going out as they're going forward. So technology itself has got to be more accountable to what it's provisioning and why going forward as well as looking back and that means that the organizations have got to have much more governance around how their provisioning infrastructure. Or on their algorithms that will save us from ourselves.

Can we throw AI at this stuff and then we don't have to worry about it. I mean how smart can the whole thing get? Yeah.

Absolutely. I think that much more there's a lot more I'd say I wouldn't say I artificial intelligence is a bit of a stretch but I would definitely say the ability to be more smart and smarter around how you provision and what's being used and why I think accountability has never been really trapped. So if you and I were developing our own code and we were provisioning the infrastructure support that release of what we were doing we would do it as an instance.

So for example looking at how costs are done from a pipeline perspective really helps. So as code is continuously released into into production the ability to understand what it took to Bear to do the development the testing the releasing and into production and tying those infrastructure costs together to see the end of a pipeline release. That sort of that sort of intelligence is required.

So you need to be able to tie things together that have a common objective and not treat everything as an individual Nuance within a within a pipeline and also the ability to be able to you know, consolidate up and optimize your resources by having a greater understanding about cross-team collaboration. So it's no good every single team, you know duplicate in the wheel. So, I think the technology that's coming out is much more around life cycles understanding what's being used if it's not being used then deep provision it so no more things sitting out there waiting for someone to notice that it's costing money before they do something about it, but associating our use with an actual value and a cost and making sure that it's a managed that way subliminally in the background whilst enable the developers do their work, but they say that's the balancing act.

So it's much more around ownership purpose and also making sure that you do that with a reduced risk of of things being Know open up for attack. So you've got to be much smarter around how things are provisioned today in the cloud. It's no longer in the hands of just writing something provisioning and letting it go and it seems like more organizations are using multiple clouds.

So as it getting harder to do all that because we are trying to figure out which workload goes where based on cost. Yeah. It's it's it's tough.

I mean data sits in different places. When you look at age Computing, you know, you're looking at infrastructure being at the edge within a local environment. So retail stores, it will be sitting out there and that's more intelligence and more more information gets to the edge and uses the cloud as a centralized resource.

Obviously, you've got to have a hierarchical way of looking at how infrastructure provision from end to end. So it has got more complex, but I think that the technology today is associating infrastructure with you know, the cost of doing and objective and outcome. So it's getting smarter in respect to having different teams working or at least being seen as being a collaborative Source in support of an outcome.

And also it's looking at the end and view. So technology has got to look above and beyond One Cloud. That's why it's going to be difficult for companies that are looking purely at One Cloud infrastructure or one method of provisioning as being the Utopia because that's not reflective of our applications of develop today.

They are pretty high rockical. They have data at different levels. They share data across the board.

They have different hierarchies. So you've got to look at it really from the the application perspective and how and understand how the application is using the infrastructure and not look at it from different parts of infrastructure and try to you know, go her the whole thing together and make sure that it makes sense. So it's really much more application to Centric thinking as opposed to infrastructure and working out what it's doing thinking.

How hard is it to set all this up? Because a lot of the devops environments have been around for a while. They're a little more rigid than maybe originally envisioned.

So, how do I insert something like this into an existing process? It's really weird. The way that things are provisioned is really around ownership and governance.

So the I think once the outline has been given in regards to what the objective is. So for example, if you are going to be managing multiple modules within an application and that are they are all being, you know, delivered ongoing and 24/7. Then the ability to look at it from an application center view.

I think that what's happened is in the past. It's look from it very much from an object view or a team view or a very specific app. So now you look at it as a set of services and the services that got to be associated with a level of cost management.

So you've got to look at how much is that cost gonna be impact in my business and what that means is is that when development today is done some of it's done for the delivery of you know, active services and active applications that up front and line to the business others are done much more of the background process. So I think the way that companies are doing this is basically Prioritizing what their spending? So once you understand what your most important applications are and don't treat them as a generic then that's it.

So actually understanding to your answer your question prioritizing the workload in respect to what the outcome is is going to enable people to actually put some governance in place that enables them to provision accordingly what they need from the cloud and if they do need to optimize Cloud usage that's based upon what the outcome and then you can actually do that in a much more effective way. So it's it's gonna be a little bit of a path for people that have been entrepreneurial for the last decade in doing this but I think that a lot of companies are moving this way anyway, because the way that applications of come online recently with the modern development practices, most of them have governance now as part of their implementation and what needs to be done is more of a an organizational design practice where it's just not the technology, but the the people that are using it need to understand what their priority is and how they're going to be allocated resour. Is in respect to other parts of the business, so it really is much more of a cost management there.

I say or a governance perspective. It's no longer around a freedom of just build it and they will come it's got to be done in a prioritized way. So it's going to take a while for companies to do it.

But if they have technology that looks at the infrastructure as a provisioned application based outcome as opposed to a activity based outcome, then that really helps so that helps the end-to-end view and it also helps prioritize what you do against what work I do that enables people to actually do the cost management more effectively. Isn't surprise you that it's taken us this long to get to this point because you know what, you're really describing is organizations behaving by grown-ups right now is this is being you know, the industry is a long tail. So I think eventually when something devops is always been an entrepreneurial practice is enabled agility things 24/7, you know increased in value that's always been let loose and you have to let it Loose to a degree to make sure that the value that you are associate with it is actually being realized in some form or another and I think even though devops has changed it flavor in many ways where you're connecting the development practice with the operational values that come out of it.

I think that most of the applications are being developed now have that in mind. So even though I think that the early stuff was probably more fragmented than it is today. There are much more team like capabilities that go across the process and I think that organizational Function pretty much feeds tool dysfunction.

So as you start applying technology to individual silos, you're going to create lots of silos that don't collaborate and I think as companies start looking at the application View and as the more mature applications, not just the front end applications, but now the back end application start looking at the value. They're providing at that end in view. It's going to make the ability for the to plan all the way through to develop tests release operations that sort of life cycle that's looked at as a holistic.

You know, it's a manufacturing line. It's no longer a set of silos. So once the silos are broken down then this will happen but devops traditionally has come from more of a devopment side the operation side now in the applications have got to a point where they start to really come to fruition where developers have much more of a shift, you know, they can shift right if you like into the operations environment that's enabled these type of practices to become a little bit easy to manage and therefore the infrastructure.

Is now accountable all parts of a software development lifecycle. All right, folks. You heard it here money still talks.

Hey David. Thanks for being on the show. That's great.

It's a pleasure to be here. Thank you. All right back to you guys in the studio.