It can be a struggle to find the right mix of services for applications and any technology professional with portions of their workloads in someone else’s data center is going to need a game plan. Join Microsoft advocate Phoummala Schmitt and host Alex Navarro as they discuss best practices when migrating to the cloud.
Take a look behind the scenes with these photos.
So, I’m going to put you on the spot now, if you had to break it down into five major steps to successfully migrate to the cloud, what would they be?
At SolarWinds HQ in Austin, Texas, I’m Alex Navarro with SolarWinds TechPod. This is Tech Talks. My guest today is Microsoft Advocate Phoummala Schmitt, also known as ExchangeGoddess on Twitter. We’re discussing hybrid IT and multi-cloud deployments.
Visibility is huge.
Whether by way of multiple vendors and/or multiple “as-a-service” types, multi-cloud deployments are becoming increasingly common. It can be a struggle to find the right mix of services for applications and any technology professional with portions of their workloads in someone else’s data center is going to need a game plan.
You got to sit back and go, why am I going to the cloud? Like, what’s, what’s driving me? Too often, everyone just kind of dives headfirst in the cloud because somebody read something in the magazine or a blog. Oh, the cloud, that’s so great! But it’s, it’s the organizations, I think, that tend to jump headfirst. They seem to want to come back because they haven’t developed that plan of that strategy of, you know, what am I going to take to cloud and why am I going to do this? If you develop those plans, you’re going to be a lot more successful because now the cloud has matured. You know, a few years ago, it was new and everyone was jumping into it and all those pain points that the headfirst divers, they’ve experienced, and now they’re telling everybody, hey, you know, I wish we could have done it this way. You know, we should probably should have planned this a lot better rather than try and do, migrate an application in two weeks. So, that’s not realistic either.
That’s definitely one of the benefits that people can take away from these “early adopters,” air quotes, because they can learn from those mistakes that were made and really start to use that to mold a customized strategic approach that’s best suited for themselves without having to go through those pain points.
Yeah. So you know, you’ve got to have a strategy and then understand that I’m going to say this, not everything is suited for the cloud right away. I think in the future, yes. But there are certain applications that I think still need a little bit more time. Um, you know, voice streaming, things like that I think could be tweaked a little bit more or you’re AS/400. I’m sorry, but that’s just one of those things that just won’t go away. But there are certain industries like the healthcare, most medical records are AS/400 and yes, a lot of organizations are refactoring these applications for you to be “cloud-native” is, air quotes, but it takes time to do that. So you know, developing that strategy, but also understanding that I may not be 100 percent in the cloud. We may be hybrid for years or that may be that our permanent structure is having a hybrid environment and sometimes that’s the best scenario for certain organizations.
When we’re talking about migrating to the cloud, what were some key considerations that you kept top of mind when you were evaluating workloads?
One of the things that we always looked at was obviously security, but also, you know, how highly available it’s going to be. It, it all comes down to the data. What we’re going to do with the data, you know, is it, is it going to be available for everybody? And then securing the data, backing it up. Just because you’re taking it to the cloud doesn’t mean that it’s completely hands-off, you know, Oh, the service provider would take care of us. No, it’s, that is your data. Ultimately you are responsible for your data because let’s say if there’s a breach or, I don’t know, you have data loss, and your customers aren’t going to yell at, you know, Microsoft or the other group, the other providers, they’re going to come and yell at you because you know, they are your customers. So if you lose that data, I guess there is certain scenarios where technically you could go back to the service provider, but ultimately, you’re responsible, so you got to make sure that the service contracts you have, you know, have the right SLAs and then if you need to do backups of your cloud, definitely do it. I know there’s some talk with like Office 365. Do you need to back up Office 365? Depending on your business requirements, you may need to do a point in time recovery, that may be a requirement in your industry. Then you may have to look at, you know, backing up your cloud, or if you’re, you know, you’re going to Azure and you have an application you’ve migrated, you basically just, re-hosted. You may still have to continue to do backups of that VM and not completely go, oh, well it’s in Azure and I have redundancy across, you know, different zones. Well that’s, you know, redundancy, high availability is not backup.
So you need to take stock and make sure that it applies to you.
Yes. And also, just because it’s in the cloud doesn’t mean that you change your existing processes and procedures. Yes, you, you modify them to reflect how it works in the cloud, but you still have policies. You still have governance. Like, you know, governance doesn’t go away when you go to the cloud. I think a lot of people forget that. You still have to govern that data. You still have to govern those VMs, you know, you still have to govern that process, you know, who has access to what subscriptions in Azure, who can do what functions that’s still the same like on-premises, you’ve got developers having access to certain applications in your environment. Do they have access to servers? Some don’t. That’s still the same. It’s still the same way in the cloud. I mean, regardless of where the data resides, you should still follow that same principle that you have applied from on-premises to the cloud.
I really like that you brought that up actually because I feel like it’s almost comforting to hear that, right? Because yes, maybe I’m losing some control because we’re moving or we’re migrating, we’re lifting and shifting. But going back to those processes that are tried and true, going back to the policies that we have in place for a reason, that’s something that’s familiar. That’s something that I know how to do.
Yes. So let’s say you have governance, you know, and you’ve got requirements where developers don’t push code into production. Maybe operations does that. You could still apply that same concept in the cloud using Azure. RBAC permissions, policies. You could still apply that same process in play in the cloud and you’re meeting your governance. You haven’t thrown that out the window. And you know, oftentimes we forget about that. We still have to think about governance. Permission, security, that’s all. I mean, yes, it’s in the cloud. I used to hear it all the time when I was a customer. It’s in the cloud. We don’t have to worry about it; it’s secure. Yeah, it’s secure, but did you enable two-factor? Because ultimately that’s, that’s our data. Right? And if you get breached and it turns out we didn’t have two-factor enabled, that’s our responsibility. That’s not the service provider. Regardless if it’s Azure, Google, AWS, it doesn’t matter. The tools are available. We as the customer need to turn those tools on because we are responsible for our data.
Absolutely. And so, are there any experiences from your past that you can pull from, whether there were some unexpected consequences that arose or maybe even some aspects of the migration process that surprised you in any particular instance?
Prior experiences. Usually the developers would work with the business, you know, they would like, oh, we want to create this application, we’re going to get into the cloud and one of the three clouds, the largest cloud providers. And one of the biggest mistakes is they would go and do it. They set up their test environment and, you know, they’re throwing code out there, and then halfway through, they bring infrastructure teams in and like, OK, we’re ready to go into production. How can we integrate this into our existing environment? And then infrastructure teams like myself go, well, wait a minute, how long have you been working on this? Um, we’re not sure if we can actually integrate this in and depending on what you want to do, our existing environment may need upgrades. And I swear, like three, four projects in a year. It was always about, OK. Um, why didn’t you include you know, why didn’t you include infrastructure teams in right away? Because now infrastructure teams are labeled as OK, we’re the roadblockers. We’re the ones are saying no. We’re not saying no. We were just saying, OK, let’s take a step back, you know, let’s rewind it, you know, tell us what you’re trying to do and see how we can integrate it. If we didn’t know from the beginning, it was kind of like, we were shoehorning. I say I’m an expert shoehorner because it was like, oh my gosh, we’ve got to, you know, we’ve got, we’re going live in six weeks. I’m like, well that’s, you know, I think that’s great. That’s ambitious, but we’ve got a lot of things they don’t mean. One example was, um, it was a CRM tool and they wanted to go live within two or three months. But, the requirement was a particular version of Exchange, which at the time I didn’t have, I was, yeah, we were planning to upgrade Exchange. Basically it escalated my project, my upgrade project, what would have been a year into like four months.
So, you know, we had to refactor our project and go, how do we, how do we quickly do this? So that this other project that was already being worked on for over a year by the development team and the business and they were ready to go. I mean they were like, we’re going live. And I was being the roadblocker, like, OK, you know, I need, I need to pause. So we basically shoehorned an upgrade in so that they could go live and time and time again, you know, it continued to happen. So I mean, my biggest takeaway for people is always involved infrastructure and operations before you do anything like that because you just never know. A lot of times developers don’t understand infrastructure and ops and there’s definitely little key… every organization has gremlins, you know, gremlins cause the custom tweaks that they’ve done. So a lot of times the developers don’t know that and if you involve infrastructure right away, you’re less likely to get roadblocked and delayed on your project going live, you know, start startup upfront. We’re, we’re thinking about doing this application. What do you guys think? And then, you know, then you can start developing with the infrastructure teams in play. Go, OK, here’s how we currently are. In order to meet your requirements, we need to upgrade, you know, x, y, z. While you’re, while you’re testing your code and building it, we can start implementing changes. That’s an ideal world.
From the get-go. So obviously there’s multiple layers to take into consideration here. Measure, monitor, then migrate, for example. But, let’s focus on visibility for a moment. Why, in your opinion, is having visibility so important when you’re considering or you’re migrating more over to the cloud?
You may not know what’s all in your environment. A lot of SysAdmins think they know, but oftentimes they find out things like, oh, well we didn’t realize that server was floating around. Or you find random payroll servers under these somebodies desk collecting dust, so when you bring these teams in altogether, you get all sorts of viewpoints together and then, all those hidden gremlins, and you know, the, the weird tweaks that we’ve done to make things work in our environment are all exposed and then that pretty much helps eliminate a lot of the roadblocks. And the visibility is also… it challenges us to be better IT professionals and developers too because not only getting key stakeholders but also the users. It’s about their experience as developers and the infrastructure. A lot of times we just think about ourselves, you know, and how we work. I mean, I laugh about it too, but…
We do. We sometimes forget that user experience, bring the users into it as well. The business users, what do they want their experience to be, what are their expectations? And then once we have an idea of what their expectations are, we can start developing an application or environment that will meet it to the best that we can. But if their expectations are way out of line of what we can actually provide, at the end of the day, you know, your project may not be as successful.
I like that. That’s a good point. Talking about moving a cloud-native workload on-premises or maybe even taking one from the cloud, bringing it back on-prem, is there a scenario or anything from your experience that you can pull from where this would be beneficial?
You have to ask yourself, why am I bringing it back on-prem? Is it because it costs me too much money or is it because it’s not working right? If it’s costing too much money, then you have to look at it and go, OK, did they right-size my VMs or the service? And then, you know, you have to look at the why. So we looked at the why, if it’s costing too much, but if it’s not working right, did we develop something that the user experience just isn’t quite there? Yeah, the coming back is always a, it’s always a tricky one because I think as IT professionals, we can pretty much make anything happen. We’re miracle makers. So yeah, there is a problem. Do we fix it instead of coming back? Because sometimes it’s going to be more expensive to come back.
OK. So maybe re-evaluating it and taking a step back and saying instead of just, OK, well let’s bring it back because that’s what we know. That’s what is familiar. Maybe there’s something else that we can do that we haven’t considered yet.
Yeah. So you got, you got to look at why we’re coming back and then once you figure out the why, identify. OK. If it’s a problem with the application, maybe we can fix that in because bringing, moving data costs money, whether you’re pumping it into the cloud or pulling it out because now you’ve got to consider, OK, if I’m pulling out, I’m going back to on-premises, do I have the infrastructure on-premises available to support that? You know, maybe when you migrated off, you got rid of that hardware, said you’re thinking about bringing it back. Do you still have that hardware and if you still do, is it still compatible? Maybe you need to upgrade because if you refactored to go into the cloud, now you’re bringing it back. That’s slightly different requirements. You may have to upgrade your hardware, the support, what you’re bringing back. So it’s not always going to be cheaper to come back. My advice is determine why you need to come back and try to fix it in the cloud first because sometimes it may not always work to your favor to come back, especially if you decommission the whole bunch of hardware and you’re like, I’m going to bring everything back and you go in the data center. There’s nothing there. It’s all crickets, you know, good luck.
Good luck with that. So what about in terms of deploying a private cloud versus a hybrid cloud versus the public cloud and in your opinion, when does it make sense to deploy one over the other?
Oh wow, that’s a loaded question. It really depends on, on your business requirements. I mean, you know, private cloud, I always think private cloud is just virtualization. Honestly. That’s pretty much what it is with some self-service involved in a hybrid. My opinion is everyone’s going to be hybrid at some point. It mean that’s, that’s just a fact of life that the cloud is here to stay. It’s not going anywhere. It’s not a fad. It’s one of those things. I think organizations just have to understand that you’re not there yet. You’re going to be there. So hybrid and then full cloud-native smaller organizations, awesome. Most large enterprises, if you’re established on-premises, you’re most likely going to be hybrid. That’s from a cost perspective, a learning perspective, and just compatibility with certain applications.
I feel like that is a perfect segue into just discussing performing historical baselining when we’re looking at year over year changes. Why is that so necessary?
So the cloud is real money. On-premises, I’m not saying it’s not real money, but you know, it’s capital expenditures and like with virtualization did we actually really do chargebacks? I mean the concept was great. You know, you could charge back, but most organizations didn’t really do that to the cloud. It’s real money and it can get expensive really quick, so you need, you need to monitor, you need to maintain, you need to make sure that what you’re utilizing is really what you need. And let’s say you, you’ve migrated an application and it’s just really consuming a lot of resources, but on-premises it didn’t because you have that historical data you should probably look at why is it consuming so much? Why, why? Why is this much bigger in the cloud? Maybe I’ve written it slightly different and it’s not as optimal. It’s not as efficient and that will actually help you save a little money if you, you know, if you look back and go, OK, my on-premises world was a lot more costly, resource consumption was lower. Why is it different in the cloud? Do that, monitoring, do that baseline and then do a comparison. I mean, it’s great when you have your existing environment, you can compare, you know, what it, it, it performed this way here and it’s performing this way here. What’s the difference? Because ultimately, you know, you want to be within the same range. It shouldn’t be too far off from each other and sometimes the cloud can, you know, I want to say most of the time, it can be more efficient, but if it isn’t, let’s take a look at the code.
Ask the why.
Yeah, ask ask the what, why, why is this costing me more? Am I consuming too much data? And maybe if you are, then maybe it wasn’t the right application to go, you know, headfirst in the cloud or maybe certain parts of it doesn’t need to be in the cloud because you can have an application that is partially in the cloud and partially on-premises. I mean there’s so many ways to do this. It’s, it’s, the possibilities are endless.
If you can think off the top of your head, what would be some of the most common types of applications or workloads you’re seeing move to the cloud?
Everything from retail applications, identity. CRM is huge. CRM. I think almost every CRM now is in the cloud, each, each organization has their own specific application that’s particular to their environment. Like had the healthcare industry, you’ve got insurance billing, processing. You’ll start to see a lot of that flow more into the cloud, but also a lot of providers are actually creating their own cloud services. I could probably, there’s one organization, they are an insurance provider and what they decided to do was instead of using Amazon or another cloud provider, they are becoming a services provider and providing insurance billing and processing for other insurance carriers so that it’s secure and it follows all the industry requirements for healthcare. And I’m thinking, yeah, I’m thinking, wow, this is great. And the company is called Highmark Health Services. It’s actually based in Pittsburgh, and Hershey, Pennsylvania, where I’m from. They are a Blue Cross provider, but that’s their take. So yes, the cloud is here, we’re going to join that by giving, other insurance providers the ability to offload their hardware and be in the cloud and provide those services. You know, I think it’s pretty neat that they, that they’ve taken on that venture and that seems to be really working.
When I heard about, I’m like, oh, that is really neat. So they’re going to be an Amazon, but for health insurance and do all the billing and the processing because that takes a lot of compute, and not only compute, but interactions with other systems, the government, Medicare, all of that comes through. And then there’s time requirements and then the approval process. All of that’s done through technically the cloud by this one provider that does it.
Is that something that you’re anticipating to see a lot more of?
Yes. Lessons learned. People talking about their experiences is huge. You know, telling your stories. I like the war stories. What did you experience? Because if you’re telling people what you know, what you’ve experienced, others are going to learn from it, especially the ones that haven’t dived into the cloud. There’s a lot of fear. There’s this unknown and they’re just not sure like, will this work for me? You know, it’s going to cost a lot of money. And then there’s all these questions where if people start sharing their experiences, their wins and their losses, like, OK, this really worked and this was awful. Others that are doing that journey will be able to learn and kind of help avoid, you know, it helps avoid those situations and then I think people will be a lot more successful and we’re definitely going. We’re definitely seeing that trend where people are actually talking about their experiences more now that they’ve had time to assess what, what they’ve done with their, you know, with their journey. And they’re like, oh, definitely lessons learned, you know.
That’s when the community that IT community is just so impactful because you, you can let your wall down and you can exchange those war stories like you mentioned.
Customers I think are starting to talk about it more. You know, we’re, we’re being more open about it. Previously at, you know, you’ve had to get like all sorts of waivers, legal to talk about your, you know, certain migration details or just certain, you know, customer base stories. I’m seeing a lot more customers actually going to larger conferences and telling their journey.
I feel like the more stories we hear from other people, that’s just really going to open up our eyes and our ears to new experiences and think, oh, that might actually work for us. What do you think is going to be the key factor here? Is it going to be having streamlined visibility? Is it going to be using the right combination or mixture of tools and services?
It would definitely be a combination of tools and services. There’s no one way to do something. I think we just know that naturally from any industry, if you think there’s like, you know, oh, this is the only way to do it. No, I mean it’s, I think it’s just human nature. There’s so many ways to do anything. So just with like the cloud migration, you can do multiple things. You don’t have to always refactor. You know, you can refactor part of your application and still keep part of it on-premises. Maybe your databases stay on-premises or maybe you move your databases to the cloud and then the front end, you know, is in the cloud, and maybe the middle portion is on-premises. It just really depends on your business requirements, your situation, and also money.
And also, yeah. What, what, what can you afford that? Right? Because that’s going to that ultimately that is the, at the end of the day it’s about cost. So if you, if you think about, you know, oh, I’m going to migrate and I’m going to completely refactor, I’m going to re-host, you’re going to look at it and go, can I afford this? It may not be affordable. So then you go, hmm, maybe I do a portion of, you know, of the application in the cloud and I keep the other, you know, maybe the back-end stays on-premises to keep it more affordable. And then you look at a plan where, OK, I’ve, I’ve implemented hybrid for this application. Maybe in the future, as costs go down, I could move everything or maybe you just keep it, you know, hybrid. Because you’ve got to think about the user experience. Let’s say the experience is better with some of the data on-premises rather than in the cloud because if it’s in the cloud, it may have to traverse through several hops.
With the fiscal aspect of that, I feel like it definitely will lead people to take a long, hard look at the tools that they’re currently using and really re-evaluate them and say, OK, am I getting the most out of this is or something else? Maybe because I am hybrid, maybe because I have moved some in the cloud and some is still on-prem. Maybe I need a tool that is going to give me that streamlined visibility and make my life a little bit easier so that way I can monitor and have that deep visibility right there.
Tooling is definitely important. When you’re thinking about the cloud. Just not the cloud, but your on-premises environment, you’ve got to see everything. It’s not a one way street. It’s a four lane highway. Yeah. You got to see everything because they all in the work together. Somehow, regardless, if someone says, oh, well, you know, this application is completely in the cloud, but your users are actually on-premises, you know, when, when your users go to the office that’s on-prem, that we’re not.. the office isn’t in the cloud, unless you are using VDI, VMware rising on Azure, where you actually had your desktops in the cloud, then yeah, you could say, you know, I’m completely in the cloud, but your users are still on-premises. Most likely you’re probably going to have possibly a domain controller on-premises, faster logins. I mean, it’s user experience. It’s costs, you know, but it’s, it’s also comes out that does it really work better in the cloud or on-premises? You’ve got, you’ve got to take a look at everything. Re-evaluating I think is key to growth and learning from your mistakes, you know, just sitting back and go, OK, what did we learn? It’s, it’s that retro that lessons learned. You know, same thing with your policies.
We actually had 94 percent of our respondents for our IT Trends Report indicate that cloud and/or hybrid IT are one of the top five most important technologies in the organization’s IT strategy today. Obviously we’re expecting that trend to continue. So looking at how they can potentially simplify this, what would be your advice?
To simplify anything that mentions multi-cloud and hybrid, that’s a loaded question too. Because in essence it’s, you know, it’s complicated, but one monitoring tools and tools that make the clouds work together. You want, you want to send the tools that work to, you know, I guess — agnostic.
They have the ability to work against any cloud. Like you said, multi-cloud is pretty much on everyone’s mind and I even know that not everyone’s going to use Azure. There’s going to be components of AWS and there’s going to be components of Google and IBM and you know, you name it. I mean my last environment that I came from, we had some Azure and we had some AWS.
So if you have a toolset that’s vendor agnostic, it doesn’t matter.
Yeah, and that toolset has the ability to basically plug in to all the different clouds and it gives you visibility across. Because like your report says, I mean there’s bonds, it’s on everyone’s list and I, my personal opinion, we are all multi-cloud now. Some may not think SaaS is the real cloud, but it is. I mean, think of all the SaaS offerings that people are using right now. You know, you’ve got Salesforce, that’s cloud. You have Office 365, that’s certainly cloud and you know, you’ve got Concur, you’re got all sorts of SaaS offerings and then you have your on-premises environment and then you have your custom apps that your developers are developing in, you know, one of the big cloud providers. That’s multi-cloud. And I know a lot of people say multi-cloud, they’re thinking infrastructure but like, no, it’s, multi-cloud’s everything. And we are, we’re in it. So there’s already stuff out there. So how do we, how do we gain visibility into all of that? And then managing that data through all the clouds that you’re in, all the environments that you’re in.
I’m going to put you on the spot now. If you had to break it down into five major steps to successfully migrate to the cloud, what would they be?
Okay. Five. So the first one is planning. Get an idea of why you want to do it, what you want to do. Who. So the who is going to be, you know, your stakeholders, once you get all the Whos, then, you know, it’s the Hows, and then it’s the Whens. When are you going to do it?
Nice. Thank you again for being here, Phoummala Schmitt. We really appreciate it. For some behind the scenes footage of our visit with Phoummala Schmitt, aka ExchangeGoddess, we’ve provided some useful links on our landing page. Thanks for visiting. I’m Alex Navarro, and we’ll catch you on the next episode of SolarWinds TechPod, Tech Talks.