How to Implement Enterprise DevOps in a Huge Organization (US 2021)

“What’s it like to move from waterfall and legacy technologies in a monolithic architecture to a highly responsive DevOps-based delivery model in a large organization? Join this session to learn about the experience at Sasol Limited, an integrated energy and chemical company based in South Africa. Bramley Maetsa, Head of Delivery Enablement at Sasol, will share the Sasol journey covering these key topics: - The strategy – focused on velocity and delivering business value - Organizing teams to be DevOps-ready - Establishing the right toolchain and DevOps framework - Shifting left and adding DevSecOps to build high quality products Bramley will share an example of how they identified and eliminated bottlenecks by completely automating change approvals, leveraging their Azure DevOps toolchains tied to ServiceNow." Guus Hutschemaekers will be on Slack to answer questions during the session. This session is presented by ServiceNow.

breakoutuslas vegasvegas2021

Bramley Maetsa

Head of Delivery Enablement, Sasol Limited



Good morning. Uh, good afternoon, uh, colleagues, uh, wherever you are in the world. Uh, my name is Bradley I'm from Sasol. Um, very excited to be here today. Uh, it is, uh, the first time, uh, that I'm presenting at, um, DevOps enterprise summit. And I'll be talking to you on a subject that I'm very interested in, which is implementing enterprise DevOps in a larger organization. So this is a case study, uh, for my company, uh, Sasso. And I'll take you through the steps in terms of what I do accessible. Just a little bit of more details. I'm looking off to dev sec ops and the chapters, the chapters are the, um, center of excellence of responsible for standards, uh, and setting up a policies. Um, my background is an electrical engineering software development, um, presales, and, um, I have experience in it consulting. I have a passion about education, uh, mentorship and anything disruptive. And, uh, when I do have free time, I play golf.


Alright, so what's some of the agenda, uh, latest in German. Just to go back a bit, I want to walk you through a full items, uh, that I have planned for you. Um, oh, we'll share. Uh, the business case that we had is also I'll Sydney, uh, walk you through, uh, division, uh, from it, uh, is a function and we'll touch base on how we organize ourselves for, for success and look at, uh, some of the processes that we've put in place. And finally, the technologies that we have, uh, implemented, uh, uh, implemented, oh, what is, uh, well, what's going on within a big organization? Why did we have to embark on the journey for digital transformation? So, um, Sassel like many organization, uh, that were born, uh, and, uh, in the fifties or even much earlier, uh, those, uh, that his study, um, um, technology to pollution, uh, you will realize that, uh, oil, um, you know, came just on the fish second, uh, revolution, uh, Sydney, uh, you know, in the 1950s when we're not far off.


So you will find that we have legacy systems. We invested a lot and in SAP, which to a large part was, uh, when we bucked in the Jenny, uh, was on prem, I wasn't working was what's a full base. And when it comes to decision making was difficult. And of course, uh, finding, uh, data, uh, was a huge consent, uh, for, for business in the end, uh, business was, you know, was out the little boats additives. So is it, that was challenge. We had to, uh, find a way, uh, for us to be relevant the way for us to, uh, to stay really current. So what did we do? We put together, um, the division that is in line, uh, with, with business or vision was, uh, the one that is aimed at delivering great value, uh, for, for Sasol and, uh, by doing three things, uh, deliver great value, uh, by focusing on the customer, um, investing in more than technology and embracing you was working with.


So this was very key, uh, to a large part or, um, uh, for Sassal we, of course at the time sessile was, uh, embarked on, uh, setting up, setting up, uh, you know, digital office. So we had to make sure that it is geared of, or, um, uh, the change and we put fee enablers in place. Uh, we invested in modern, our ERP. We, uh, invested on, uh, cloud Jenny, uh, primarily here. Uh, we focused on moving, uh, to, to Azure, but a strategy, uh, for cloud really is, um, is a hybrid. And, uh, we are, uh, embracing, uh, both, um, uh, Azure and an AWS is just, we have more maturity, uh, in, uh, in Asia. We also recognize the organizational, um, uh, culture shift that we had to, to look at, and certainly our people, uh, their coal of, of the journey, uh, that we'll look at from the people lens perspective.


We're so, had to organize ourselves, um, organizing ourselves in that the org structure had to follow the strategy, um, is Sussel today is organized in two portfolios, uh, is organizing in energy and, and the chemicals. And this is where you'll find a lot of execution and delivery of, uh, technology taking place. And certainly if you look, uh, specifically in energy, you're are going to find, uh, two VPs that are sitting in the space responsible for, uh, platforms and others responsible for, but Florida's in the idea is that the two of them work and, and, um, uh, devil's way of working. And of course, the corporate center, we had a CIO, uh, and this is where my team, uh, lives, um, and all responsibility, uh, is to set up a strategy, a roadmap, standards, policies, uh, templates, pinch framework, uh, quick system is to ensure that we create that digital backbone for business to be able to distalize.


Now, if you look at the second part of how we organize ourselves, this was really focusing internally into how the teams will go into the idea of, uh, collaboration and, and also the idea of continuous learning. And here we set up eight chapters. Um, chapters are a way, uh, to promote, um, experience here, best practice collaboration, cross pollination of ideas, uh, and innovation across agile teams. We're also coupled with, um, chapters, we had committed to the practices. Um, these are the communities that get together, um, on monthly basis to share some ideas and practices on what's going on. Um, uh, spice for leadership, uh, is concerned.


So what is dev ops anyway? So, and I feel strongly that, uh, although, you know, um, dev ops movement has been in existence for, for quite a while. Um, but each organization has to define what it means to them and to us certainly, uh, we do recognize that if we all are to more faster at customer speed, uh, with quality, uh, and stability in production, we have to define DevOps as a UNM of our people, processes and products to enable continuous delivery of value to, uh, add users. We were primarily targeting the old was working, which at the main we had, uh, uh, silos between, uh, or amongst, um, cast Evans, uh, developing teams and, uh, and operation, and our, um, view here was to make sure that those silos are taken out. Um, and we believe that by doing so we'll, um, achieve really dev ops as a way, or a mode of delivery by maximizing that, uh, flow once massive flow, uh, amplifying feedback, uh, and lost the, have the idea of continuous continuous learning.


We also looked from technology lands, uh, folks, uh, from service now will be familiar with this slide. Um, uh, we adopted it, um, uh, and make maybe it or Sassel, uh, specific, uh, but really, um, the desperation is from a service now. Um, if I can just spend a little bit of time here, colleagues, just to show you, um, what, what I actually have on the slide for you, uh, is that accessible? Uh, the ambition is to have, uh, two things, um, uh, doing two things, instead of, uh, not a lot of things that we used to do in the past one was that the demands, which are, um, project specific, the go through idea puzzle, uh, prioritize by portfolio managers, uh, and given over to, um, folks that have gained, uh, at the unified backlog and prioritize with the delivery teams that will actually use tool chains to, uh, deliver, uh, the product in an automated way.


Um, uh, specifically here targeting, um, you know, uh, key concepts like automated, uh, changes, which I'll get into details on later on, and also making sure that our products are compliance by design, and we are able to see what, what makes our, and of course, uh, if, if things are transactions, transactions are, uh, cataloged items, uh, incidents, uh, those will go through your traditional ITA, Sam service portal. What also want to make sure that all of these pieces are connected. Uh, uh, we're looking at automating a GRC, uh, looking at the concept of SecOps baked that in, within dev ops to really have pure desk of, and also look at the, um, the feature, uh, world of I ops, uh, we already implemented observability platform here. Um, but at the core of it really, uh, if there's one thing that I want you to take away from this slide, uh, it is that, um, um, everything starts with data and answer with data at the core of what we do is a CMDB.


We, uh, looked at CMDB, we looked at Aligna, I see, uh, uh, a CMDB with the CS, uh, DM model, uh, which is the recommendation we've got from my colleagues, uh, in six. Now we also worked with a partner, um, a platform mission, uh, out of, uh, Europe, uh, that, um, lid helped us to make sense of setting up the base, uh, for us seemed to be, uh, we did that analysis that looked at the data, uh, that looked at how we organize ourselves, uh, in terms of people. Um, and we looked at procesees, um, that we have, uh, as far as , um, and we also looked at the technical debts. Um, so, uh, it TSM, uh, incest, or has gone through a number of iteration. Um, and it was heavy, customized as only recently when we managed to, um, you know, uh, bring in more and more out of the box functionality, but there's a lot of technical debt, uh, that I think it will take us quite a bit, how to what I just quickly on technology.


Um, we also realized that integration is key. Every piece, uh, that I have talked about, um, is that it has to be connected, um, uh, when you that, if we want to have it. So it's a visit by the tape. Uh, we had to make sure that, uh, the items that are being, uh, uh, put into a catalog, uh, need to be connected to a pipeline, the pipeline, the CIC D needs to be connected to service now with the automated changes, uh, takes place a lot of lenders here, um, in terms of, uh, really how the flow, you know, uh, what's going to go, um, how we map, uh, things like epics, uh, the user stories and tasks on both sides of the world to make sure that there is that continuous flow. Uh, in fact, the picture that he looking at, uh, from, um, vegetative ops was also connected to citizen now, ISO model.


So one is to make sure that that visibility, uh, happens, um, uh, in order for us to achieve automated changes, which are, I'll give you some details right now. Uh, the DMR, uh, automated changes, quite simple. So we'll look to add, um, with a bunch of next saw, uh, and most of the books next really sits, uh, and, um, uh, in cap, uh, the idea, uh, that, uh, you wouldn't have to unload gain, uh, uh, coal, uh, that about, uh, five people need to get involved depending on what CPI are you working on may to get involved, to, uh, to approve that. Uh, and also you have to sit in a of making, uh, to present, uh, we find that a lot of this can be automated, and that was really the key vision, uh, around automated changes.


The automating that, uh, for us was practical. It was about moving, uh, from weeks and days to seconds and minutes. What you're looking at is one of the product, uh, quite quick, product that with bold, uh, super, uh, chat bot that handles, um, uh, thousands of possessions, uh, for our customers today, uh, from service desk to, you know, having your passwords reset, uh, we're using, uh, the, the chat bot, um, that is based on the net framework. And, and, and what do you see in front of you, but is that from packaging, uh, the solution, uh, from one environment on the, looking at the quality gates, where there is to security, uh, and to, uh, making sure that we still get the approval from, um, uh, from the, um, service owner was key, but he can see that everything is actually, uh, um, we went from, uh, seconds, 19 seconds to just, uh, waiting for, uh, about 40 minutes for an approval when that is done.


We took about 10 minutes, uh, to put it in production. So ladies and gentlemen, this is what I call moving at a customs speed. And in what's possible, I must mention because we have, um, since now, uh, devil's module, uh, integrated, uh, with Azure, a DevOps, of course, uh, if you, aren't going to implement technology, uh, you also have to measure, you know, when FedEx deliver, uh, although measurements, we want to move more, uh, you know, from proxy metrics that look at concessions, uh, to, to value, uh, to even, uh, these are the initial metrics that we thought would give us a better view in terms of people, uh, the culture here, people are all adopting the technologies that we have put in place, and I will recommend you to the same, uh, more insights in terms of what you get out of, uh, uh, serious now, um, um, uh, DevOps module that the small coming, uh, it would w what the new version of wrong coming in, um, in terms of ticket, uh, take away points, um, recommendation.


Uh, I would like to leave you with, uh, three things. Um, one is that, uh, you really have to work on the vision, uh, make sure that this is linked, uh, to the business strategy. It's, it's very, it's very key. Uh, people would, uh, encourage, uh, those folks that are different masters, uh, to really innovate and make sure that when they do fail, uh, folks do Fayetteville. What also, if you're going to have a strategy as, as it needs to be powered, uh, by, uh, the ox structure, have the right people around you and see, uh, do, uh, set aside by the team best, uh, in modern technology. My last name, uh, if you want to know more about saves now, um, Dave offs, uh, go to Sage, boss dash, um, DevOps. Thank you. And appreciate that the time goodbye.