Mastering Cloud Migration with Google Cloud: Best Practices and Insights with Stephen Orban
Transcript
(dramatic music) >> Hey, guys, how's it going? It's Keith Townsend, principal of The CTO Advisor. This is a special video sponsored by Intel. We get to talk about hardware, software, and cloud providers coming through. If you haven't watched the video that we did with Intel, talking about how they're doing the low-level work to make Istio run faster on Sapphire Rapids and accelerators, check that out because this is the follow on to that conversation. Just by adopting Istio on Intel processors, your abstracted infrastructure runs faster.
So if you're adopting cloud-native technologies, there's really no optimization you need to do. Now, you couple that with a cloud provider, such as Google Cloud, who has adopted and exposed the accelerators to Sapphire Rapids, you get magic. You have your cloud-native applications using the service mesh, such as Istio, and it just runs optimized. But how do we get there is the challenge. And I have the perfect guest with us, Stephen Orban, VP of- >> Migrations. >> Migrations- >> And, yes.
>> at Google Cloud. And you're going to talk to me about RaMP, which is a process for getting our traditional applications out of the data center into Google Cloud, running in a optimized native cloud fashion. >> That's right, thanks for having me, Keith. >> Thanks for hosting us here at the beautiful Google campus in New York City. It's pretty cool here. >> You bet, it is a cool campus- >> Yeah. >> like many of our campuses. It's awesome to be here, and thank you also, Intel, for your tremendous partnership.
Yeah, so I joined Google eight months ago to help us really build a world-class discipline around helping customers do large-scale migrations to the cloud. Prior to this, I spent eight years at AWS in a variety of roles, also helping customers migrate. Before that, I was the CIO at Dow Jones and led a big migration. And prior to that, I spent a little more than a decade at Bloomberg doing a bunch of things in and around the cloud. And I've had the opportunity, fortunate opportunity, to talk to more than 1,000 customers over the course of the past decade, all of whom are thinking about moving some, or even, in some cases, all of their IT estate to the cloud just so that they can take advantage of the agility and cost benefits that they get.
And most of the conversations I have with customers go something like, "Stephen, I'd like to be 75 to 100% to the cloud in the next three to five years," and I've been having that conversation with the same sets of customers for more than three to five years. And so, what is making it take longer than some would like? And if you kind of think about it, a large-scale cloud migration in a big enterprise, in particular where they have people, process, and technology that all work in unison with one another that have been operating a particular way for 10, 20, 30, 40, in some cases, years, there's a lot of change management that's required not just in the technology.
So the way we're thinking about this is how can we bring the best practices that we know help a customer not just understand what's in their IT estate and move it responsibly and quickly to the cloud, but what are all the other things that we need to bring to bear, both from Google Cloud and from our partners to help a customer through that change management? So there's kind of five best practices I really think about along these lines. The first is making sure we have a great business case and executive alignment on that business case, so why are you doing what you're doing?
What is the cost savings you're looking to get? How do you quantify the sort of speed savings? One customer we worked with has 2,000 developers, and their business case hinged on the fact that they believe by moving to the cloud, their developers would be 50% more productive. That's 1,000 extra developer years per year that they'll free up as an effort of moving. So everything around their move, that North Star that they set, hinged on that business case. The second is having a training plan, training, and I'll call it enablement plan, enablement plan.
One of the reasons I became an engineer is (Keith laughing) 'cause you couldn't read my writing. >> Your handwriting's slightly better than mine. (Stephen laughing) >> So, you know, look, people who have been in, IT folks who have been in, let's say, a sys admin or a network engineer or a security operations role for 10 or 15 years pre-cloud need to, you know, upskill and uplevel, okay, learn new things in order to operate in the cloud really well. We haven't changed physics or computer science.
Like, it's still the same core principles. However, we've made it easier to build distributed systems at scale in a cloud operating model. So having those folks have a clear path for how they can get trained and understand what the cloud's going to mean for them. One customer I worked with had a Cloud Warriors program, where they said that there's a scientific belief that if 10% of a population holds a unshakeable belief in something, it'll create a tipping point for the rest to follow.
And so, they took a goal of having 10% of their IT associates trained and certified on our cloud technologies in a three-month period and since have migrated multiple data centers to the cloud. So that's the second best practice. The third is having what I'll call a CCOE or a cloud center of excellence. A diversely functions team at the center from developer, operations, financial, FinOps, as we often call it, financial operations, HR, to help with the training plan to really think about harvesting the best practices in the learnings as a migration unfolds and take shape.
In a big enterprise context, you might have 1,000 applications in your portfolio, let's just say. But of those thousands, there's usually a couple of dozen patterns within those thousands. You might have a lot that are like a CMS system, a lot that are an N-tier application, so the CCOE team would really be responsible for setting the blueprints and the best practices for how to move those to accelerate the sort of migration and set principles for the migration on an ongoing basis. The fourth is setting a great OKR.
So once you have your business case and your executives are aligned on your North Star, make my developers 50% more productive, let's say, what is the actual measurable results we're going to take on a monthly, quarterly, annual basis? Move the first 100 apps by X date. Train my 10% of my team by Y date. Make sure that the landing zone where all the applications are going to get moved is set up by Z date in accordance with these regulatory practices if you're in a regulatory environment.
Having a very clear line in the sand is the fourth best practice. And then fifth is great governance. So how do we meet on a regular basis? Weekly, at least. In some cases, some of the larger migrations we're doing, I was just meeting with a customer earlier this morning, big media company moving all of their mail and analytics services to us. We meet on a daily basis, looking at the OKR and what's our progress against it? What are the risks, issues, and challenges?
Stuff is going to come up, so the point is, how do we deal with it very quickly and make sure we're escalating really quickly to unblock issues? So, as you can see, it's not really just technical. Yes, we also need to understand what's in the IT environment, have a good plan for what's going to move in what order, how do we segment all the different applications into different migration strategies. But I feel like most technologists can get the technology right. It's when we miss these things that have to do with the change management to really get a program and discipline in place to move it, to have a migration operate at world-class is really where I see a lot of these things stall.
>> So, Stephen, I really appreciate you taking out the time. We've covered this topic a lot here on The CTO Advisor, talking about cloud migrations. Optimizing your existing workloads for the public cloud is much more than a technical problem. This isn't moving a VM on premises to the public cloud. We've leaked that problem. It is about moving people, process, and technology so that you're getting the most out of your public cloud investment. We appreciate Google Cloud sponsoring past content talking about this.
We'll link to that past content and research, but specifically, continuing the conversation, going through and seeing how technology itself is not enough, and we need to have people process around that. And, Stephen, we appreciate the work you and the team have done and educating our audience. You want to find out more about The CTO Advisor and this project that we did with Google Cloud? com/google-cloud, for our research specific around Google migration tools of that time. I'll talk to you online @ctoadvisor on most social media platforms that you want to engage.
DMs are open. Stephen, thanks a lot. >> Thanks for having me, Keith, appreciate it.
