Cloud Crunch
Cloud Crunch

Episode 5 · 2 years ago

S1E05: Hybrid cloud computing

ABOUT THIS EPISODE

Today we welcome our first guest, Dusty Simoni, Sr Product Manager at 2nd Watch, to discuss hybrid cloud computing. We dive into what hybrid cloud is is, examples of hybrid, benefits, complexities, and how to get started. For this conversation, we look at hybrid cloud as on-premises infrastructure and public cloud - specifically around AWS, Azure and VMware - and exclude private cloud services.

Involve, solve evolved. Welcome to cloud crunch, the podcast for any large enterprise planning on moving to or is in the midst of moving to, the cloud, hosted by the cloud computing experts from Second Watch. Jeff Aiden, cofounder and Executive Vice President of marketing and Business Development, Ian will be chief architect cloud solutions, and Skip Berry, executive director of cloud enablement and now here, are your hosts of cloud crunch pay good afternoon everybody. Jeff Aiden, cofounder sech watched here and I'm joined today by Ian Willoughby, as skip Berry is out with a client. On today's episode we also have a special guest, dusty Sumoni, senior product manager of highbred cloud computing here at Seco Watch. We're glad to have his experience and expertise as he shares today as we discuss highbred computing and what it is, examples of how to get started and some of the viewpoints from cloud born and cloud native providers like ourselves. So welcome to the show, dusty. Thanks for having me. Jeff, yeah, it's great, too, great to have you here. Well, let's start off, you know, with kind of defining highbred computing from you and INN's perspective, because it is a vague term. It can cover quite a big ground out there. So let's just start off with what hybrid computing is and where we're seeing some of the use cases today. Sure, I think you hit it right on that. There's a lot of confusion out there. Everybody has their own definition. You'll see some people think hybercloud is using just two different cloud providers, of Azure and it to us, while others tend to think of it as more of using a private cloud and the public clouds, and that's typically how I like to think of it, is using the private cloud being mostly vmwhere, and that in the public cloud being your main providers of aws, Azure and GCP, and working that way. Use cases, you know, span beyond multiple use cases and you tend to agree with that definition of hypercloud. I do absolutely. Yeah, it's a combination of one or many cloud providers with it could be a managed data center, it could be their own data center, whatever combination it is, or just some reference of on on premise equipment that just can't seem to move to the cloud. Yeah, I think that's a large driver of it. Where people tend to have those sometimes lacy applications is just can't move. But they are evolving and becoming more Dev centric and more cloud applications being put in place. Digital transformations are happening, but they're still those legacy apps that can't even be lifted and shifted into the cloud. They have to remain in the data center. Yeah, those are those are great definitions and you know, we've seen it kind of morph and change, right, I think you know over the ten years, early on highbread and private cloud were used to combat some of the growth that Amazon was see it. But now we've even seen Amazon embrace it without post and new products like VMC vmwhere's cloud on aws? Right, and so I think it's continuing to Morphrom commans different things to different people. Right, absolutely, and about it. I would say several months ago I was at a security forum within it was all CIOS and cteos of Publica trading companies and large enterprises, mainly in the southeast, and they all were saying hybrids not going away, it's here to stay for the foreseeable future. Yeah, we really saw that shift the last two years, I would say two or three years right, where it's definitely not going away and and really even out in the field. There's lots of use cases like batch computing. We're using the cloud to burst and expand your capabilities or were compute power, as well as high performance computing. A lot of those are high bred applications in...

...many cases. Right. Absolutely, Yes, in a lot of cases, because data is never going to move entirely to the cloud, because that's the secret sauce and there may be too much of it. It. Take an insurance company that's been in business for over a hundred years. They're not motivated necessarily to move all that data to the cloud. It is exactly what makes them worth while and they want to keep it close to it. Maybe mainframes or other it's listing things that just do not lend themselves necessarily great to the cloud at this point. Well, and you said secret sauce that they may not move to the cloud. It's not that they believe that the cloud is insecure. It is it's proprietary information. It might be formulas, it might be a way of doing business or data that they have that they choose do not want anybody, even providers, to have access to. One or two people in the organization may have access to some of that proprietary data is what we've seen in the past. Absolutely, and yeah, again, it's not security, because most people do understand that the cloud is in most cases far more secure than you can do it on Prem you've got people working at twenty four hours a day that you don't have to employ on your own side. So that's fantastic. But yes, it's in some cases, if you look at just the raw cost of storage and the type of storage that they need in the formats that they need, there may not always be at r Wi specifically for the cloud. Now I've run into that a lot as well. We're what we've seen is that, you know, as you get your data center equipment and you're running near eighty percent, are above utilization. At that point there the cost savings and your own data center might make sense to run that workload there. But for that first capacity as well, as you're non consistent workloads that you can turn off at night or or don't need full seven, those in the cloud definitely you can experienced the cost savings there. Yeah, that's an interesting point that you both bring up, that you know, why don't companies just move to the least expensive right so in many cases cost is not the only driver. RELEASED DOWN, not to an application level that enterprises are looking for. Is that we're seeing dusty. Yeah. So, especially not at the application level. They get the ability to scale up with the cloud and scale down and and that versatility and flexibility give them the cost savings. But for those always on applications that are pretty stagnant, if they're running a large environment like that, have applications like that, keeping it in their data center can be cost effective for them that way, especially if it's also not being egress traffic going out into the if it's all internal it t APPS. From that Standplat, then a nonincurrent bandwidth charges that might be happening if you go to the public clouds. So there's definitely a use case of their strongest case for a mixture of applications being placed outside into the public cloud and then maybe internal I tapps remaining in the data center. Yeah, and one major driver we've seen over the last couple of years is data and the amount of data is increased right is companies are not able to build, or nor want to build their data centers as fast as the business needs. Right. So we had very large financial services company that that used it to run money Carlos simulations on their actuaries to understand risk exposed. Right. There's tons of use cases where they're either in the queue or they can't quite get the it resources where hybread application or service may make sense. Absolutely. The other thing that you know your data center equipment, if you've made the capital investment to purchase it, it's a sun cost already. So why you still have your power and cooling that you have to maintain the data center. But if it's already they're getting the longevity of the life out of your existing premise equipment. Can extend the life, especially in some of the use case I'm sure will probably talk about later. Yeah, that's that's a great point. So...

...what are some of the security concerns that does your in? You see that companies are concerned about with hybrid computing, but I think definitely and in the past, and it's getting less and less, which is which is good, a lot of the companies were very concerned about who would be able to see their data, would it be able to leak? How do they know that it's locked down? As the tools have gotten better, there's better visibility in it but if you just take a look at all the compliance that cloud providers have, they already have PCI, hip, ISO, nist, fed, ramp, you name it. It goes on and on and on. So there are constantly looking at this and bringing in third party auditors to verify that their practices and their architectures are up to the compliance that people need. We also know that the cloud providers can work with secret government agencies and if you can work at that level it's going to be far better than what you could probably do on Prom having all these eyes on you all the time. And you know, if Amazon or asure Google makes a mistake, it's going to be on the Wall Street Journal. And that alone is a lot of pressure that a lot of people would just would never experience unless they kind of outsource that shared security model to a club provider. So it's more perceived security concern than a reality. It is, you know, and they do decommission resources correctly or reprovision them. They're making sure that data is white and again, but it doesn't get people off the hook from doing the best practices, that is, making sure that you have security at rest to security and transit and if you fall the practices in the principles that you're supposed to be doing. anyways, you're going to be far more secure in the club. Yeah, he does. You think Amazon's be a release of outpost is to address some of the security concerns or to bring the high red capabilities closer to the customer? I think it hits on both of those. Actually, they're bringing outpost to the data center itself for those clients that are just uncomfortable with having their data a lot of the times or their secret source. We talked about, you know, O side of their control. You know a lot of the chief security officers. You know they're the ones that are jobs are on the line if there's a breach. So they want to have their fingers and control and they know if it's in their data center they feel mill secured. That just sense of security. But then also those applications at have latency, so Amazon bringing those into the data center of the customers so that they can then get rid of reduce the latency that might be there for some of those legacy apps that just can't make the migration into the cloud. But I want to go back and E and hit well one point that I thought was really good earlier about that. Sometimes the public cloud is more secure than your own data center. I think in the regards security concerns, customers that are not adopting a hybrid of strategy run the risk of shadow I t going on. If they keep them try and isolate them into their own data center on their own infrastructure, you're going to see the rogue operator doing shadow I t where they don't have the visibility. But by embracing hybrid cloud strategy you can see then take advantage of the tools such as control tower sentinel, using blueprints and landing zones and making sure that all of that is under the purview of the Sisso so that there isn't that concern of less security in the public cloud. It's actually more these tools that azure and aws have built have now made it where your public cloud probably is a lot more secure, other than your sister going in and just locking everything down and not allowing any egress traffic. Yeah, and we talked about this on the previous episode, one of the benefits of outpost rite is Amazon ships that Rax it and stacks it all for the customer and then the partner and or employees can access that services which is pretty handy. They switching gears real quick duste. As you look at public cloud and...

...you have on premise resources. Let's talk about management, both from, you know, the associate side or partner side, as well as to like orchestrate management of these workloads and different environments. What are you seeing some of the tools and how our companies being successful in doing some of that? So the successful adoption of a hybrid strategy and also one of the complexities that goes along with it is exactly this, the the management orchestration. How do you get your hands around that? And the Marcus really started to see an explosion of clow management providers coming out developing tools. Companies such as well, vmware has their cloud health tooling and their V realized sweet but then you know, right scale has been around for a while. They were bought by Flex Sarah, but they have tooling that allows for your code to be deployed to whichever cloud that you want. Does the orchestration, but looking at your automation and government's life cycle management, usability, the access, all of that can be done through a lot of these cloud management tools that are on the market now and they just continue to advance in their abilities one partner scaler. They're heavily connected the terraform and folkusing on you like you can your terraform a scaler to deploy the multiple cloud yeah, we love terraform. Anything to add to that one now, I think absolutely right. You know, everybody's always has this holy grail of one single painted glass. The toolings definitely got better. We tend to recommend not going that direction but using there are great tools like cloudhealth that could you visibility across the board. But as far as management, a lot of times you want to use something like terraform. If for infrastructure is code, definitely you need infrastructures code. Let's get that clear. But as far as using the unique tooling for each cloud provider, I think that's important because if you try to do a single pin of glass for, let's say, service catalogs or launching off new resources, some cases you can kind of simplify it down too much and not be able to take advantage of the best aspects of each cloud. So you just have to be careful with that as well. Yeah, and you know where we've seen companies successful right is where a lot of times they leverage partners for some of the public cloud and will manage their internal resources where their experts are. By all means you don't have to use a partner. We've just seen customers become really successful when they do to ensure that they don't have to train everybody initially up front on public cloud services. That comes over time as they start to work with it and get educated with it. Obviously. Hey. So let's talk about some of the complexities. The management piece obviously deals with some of that, but where we seeing customers get stuck or the complexities as they start to implement either multicloud or hybrid cloud computing? I know we touched upon it in previous episode, but let's you know, dusty, get your perspective and then he and of course short. So I think one of the things that I've seen clients get stuck with is not doing a full assessment of their applications and their dependencies, and so they move their application to the public cloud but they keep their database back into their private cloud er on premise data center and then that latency drag, you know, and we had one client that had sap that they kept in their data center and they moved everything else out into the public cloud and the latency and the hit. You know, SAP's a hog as it is on resources and trying to then come back the performance hit they took. It wasn't a good experience for them because they didn't map out all the dependencies. So looking at your dependency mapping, seeing what's related to make sure that when you're making those moves and doing that planning of what you can move to the public cloud...

...and what you're leaving back in your data center, you know how is it going to be impacted when you make that migration? Hey, does stay on that example. Was a case that the client just didn't understand how to leverage the pub the cloud or how to create that seamless hybrid experience? Or was it you know that they they had other issues. They had other issues. They had worked with a consultant prior to bring Second Watch in and and that particular instance there wasn't dependency, proper dependency mapping done and they were quite upset with the implementation was done. They came to us and then asked us to fix a form and that's what we found out that it wasn't the solution. They were actually doing a VMC on aws solution and it wasn't the fact that V m where what was poor performing, but the way they had implemented it. So once we made some changes and implemented hyberd connect the performance increase some. You still have that latency, you know, but we're taking steps to improve that. And so as vm where right. So you have some complexities with the technology, but you also have complexity with some maybe traditional partners or partners companies of leverage in the past that just don't have that depth of knowledge. Absolutely, and you know, a tuting our own horn here for a second. You know, we are one of only five master service competencies on vmcn aws in the United States. So, knowing having done numerous implementations that way, we understand the complexities that go along with running a hybrid vmc implementation. In anything to add on complexities? Yeah, absolutely. I think sometimes it gets a challenge to figure out, let's say, rationalization of where to put things as well, so in a hybrid world that you keep that workload on prem and that goes back to some of the analysis that Dosty was talking about as well. But you know, we're exactly should this reside and it gets even more complex when you are dealing with the multiclod environment. You know there's cost implications, there's in just different apples and oranges in some cases, but also just the networking components. It's getting a lot better, but you know, you still those skills that you still had on prem related to when connectivity. That still exists today and it really have to think about how to monitor that traffic in Gress, egress East West and those types of things as well. Great, let's talk about hybread architecture for a moment. That often involves silos, right. So you got the infrastructure, that orchestration, the application, the data, that item management, and there can be issues when companies are trying to architect with these silos. Let's talk about some of the challenges and how companies over should look at overcoming those to be successful. Sure, I think one of those go back to the previous point I was talking where we, you know, having a dependency map, understanding two ens point your traffic flows, what the networking information is, but also your data. Data is one of the biggest obstacles to clients moving to the cloud just because of the the cost of moving it in and moving it out and then accessing it. So they abundancy of it impacts that. So getting a clear understanding of these silence and the data is often siloed off to the side and every department has our own data. You know, some of that can be resolved by it data legs and data platforms, and that's a great topic for a future podcast. So and not my area of expertise, but definitely something that we're seeing people have challenges with understanding of how do we get access to data were we might have had a big EMC or our net APP that everybody had their own carved off portion. But how is it going to be impacted as we go into the cloud and sometimes Ip Ministry is or even realizing all the different people accessing those big sand array storages. I also think that a lot of the silos are introduced or they're created by or existing because of the human...

...factor, and a lot of it is, you know, we have various groups. They've done their own thing. Networking did networking, storage did storage, sysadmin's did sysadmin work. As this you can go to the cloud. This is an opportunity to do a cultural shift inside your organization as well, and that usually involves making sure that you have the buy in from the top level down of an organization. Most developers they know about devops, they want to do those types of things, but it needs to work from the top to the bottom and really trying to break those silos down. And there's a lot of patterns out there that we didn't create but they've organically created in the industry. It's the formation of Cloud Center of excellence or Cloud enablement engines, depending on what you want to call it, and it's important to have those together during this transformation time period to bring these groups together and the stakeholders to really understand each other in the challenges and then develop a strategy around the hybrid deployments of how you're going to address those things culturally and within the groups and break down the silos. Great. Well, let's maybe also address what kind of training and obviously it's going to make a difference if we're talking about you know aw as or you know vm where or asure or Google, right, but let's talk about addressing some of the basic training of staff or how they work with partners like us to help get them up to speed as they're starting to think and build around hybrid computing so dusty. You could even go back to your example of that assessment and listening the dependencies. You know, that's not new to it. But how come that stuff gets missed as they start to look towards public cloud and and by all means and Chime in there as well? I think some of that has to do with you know where people and the development especially that's happened over you know. I think he's been around for forty years, but clouds only been around for ten and you have a mixture of employees who are who have done things you know all in the data cent are not adopting the cloud and don't necessarily understand the DEV offs type approach to deploying while then you have your newer, you know, employees that are application builders that have fully adopted cloud, and trying to marry some of those definitely becomes I think you see your people who are fresh, newer developers. They won't even think about doing a stuff in the data center unless their force. They feel it's our Arcic, while the legacy staff that is a history of running, you know, Siss Admin and knowing what's done there, and so as getting those two work together and we typically do that by starting off with, whenever we have an engagement, getting all the stakeholders in the room understanding what their goals are and how do we achieve those goals and bring everybody up to speed. But I have a colleague here, Stefano, who'll tell you that. You know, devops is not just going in and saying, Oh, we're going to do the devops. It's coming in and getting the buy and throughout the whole ortization it starts at the top down getting that buy and but also being implemented, and then it is as a new skill set. And whenever we complete an engagement with there are several artifacts that are left behind as well, and obviously we could talk more about what goes on in between, but getting everybody up speed, training, understanding how the tooling works. tooling something part of it, though, because it is a concept and mindset adoption. Yeah, know, and you touched upon a grade point that I can say with confidence. You know, for over ten years of doing this, I have yet to see somebody from what we would call old it embrace the cloud and lose their job. They've actually been promoted many times within their company and have gotten hired away to lead cloud development cloud transformation. I've never seen anybody get fired. However, I have seen people to get moved out that did not embrace the cloud. And so for those of you that may becoming a little late to the cloud party or adoption, rest as sure...

...if you embrace it and learned it, you're actually going to propel your career and have a chance to make a lot more income and change your lifestyle because it's in high demand. So rest assured if you embrace some of the new you will experience benefits from that. And anything else to add that dusty might have glossed over or something else you want to add on that piece. Yeah, you know, I think it's you've made a very valid point, Jeff, and the companies that we have dealt with, you're absolutely right. The people that do embrace the cloud, they win, they do well, they go on, they get promoted. We kind of keep an official running total of the number of people get promoted that we work with and it's very vast and growing every day. But I've also never met anybody yet that decided that they wanted to learn the cloud and they're an it that hasn't been able to adopt. It's really just a choice at that point. So get those training, get the certification as a as an individual. You put that on Linkedin, you're going to be amazed at what happens to the number of calls you get by recruiters. So keep that in mind as well. All right, Hey, we've talked quite a bit about vmwhere, and you know vmwhere on a tows aws and outpost. Let's talk about add your stack, which could have similarities to outpost as far as you know physical gear. It's a little bit more custom but he and why don't you talk about some of the use cases or you know similarities between some of those? Yeah, absolutely, I think as your stack's been around for a couple of years at least and it's it's a great way to get started when you're trying to test cloud deployments and it rates very well into the development system, as Azure does itself, into the different Microsoft tools, and you can start small and you can really try out what you're trying to do in the cloud and as great for APP DEV testing. What you get there you could push into azure. It looks just like the stame environment and you're good to go. It's also same use cases as well. If you've got latency issues but you want one platform and you're got to stay close to the data or you got to stay close to manufacturing equipment, it's a great thing to have right there on site. Hey, guys, before we wrap up the shows, there anything you want to add around vmc on aws, outpost, Azure Stack, high bred cloud computing? That that we want to get in in the last couple minutes of this show. I just think it's really exciting at this particular junction of the industry that there's a lot of options now for a hybrid environment. If you want to extend your vm where platform into the cloud, fantastic. You've got that option now. You've got the people in the capabilities already in place. But if you're already on aw last or you want to move towards that way an outpost solution, we'll look just like aws. Same thing with azure stacks. So again, just fantastic opportunities to really really nail down a hybrid strategy. Yeah, I think it's a really exciting time as well because you're seeing, you know, your private clouds moving into the public cloud and your public clouds moving into the data center that clients own on their own, so that mixing and Meshing of the world's there's going to be a lot more coming over the next two three years. Hey Dusty, thanks for joining us today. It's always great to hear from really knowledgeable experts within Second Watch and appreciate what you're doing to move hybrid cloud computing forward with our company. Always a pleasurey in. We'll talk with you next week, obviously as we kick off the next episode, and if you have any questions, comments and feedback listeners, please send it to cloud crunch at second watchcom and we'll see you next week. You've been listening to cloud crunch with Jeff aiden, Ian Willoughby and skip Berry. For more information, check out the block second watchcom company block or reach out to Second Watch on twitter.

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (33)