VxRail Proof of Concept - SaaS Controller Surprise
Transcript
(air whooshing) >> Hi, I'm Alastair Cooke. I'm an analyst here with the CTO Advisor group. In this video, we're looking at some VxRail and some multi-cluster VxRail. Hands on the tools will be Mr. Tom Green as usual, and also joining us is Mr. Curtis Edwards from Dell themselves to give us a bit more of a view beyond just what we do. But to start off with, I'd like to ask Tom, just to frame up what the problem is we're trying to solve here because the CTO advisor hybrid infrastructure has been a little overloaded.
>> Hi, yeah, sure, Alastair. We are in a bit of a dilemma in the hybrid infrastructure. 7 and we're running Visa, and Visa has been completely solid, a great platform for us for storage, but now we're looking to upgrade. And there's a lot of caveats and pitfalls with the actual hardware itself that we have to consider. So we decided, instead of trying to do it ourselves and doing all that research, we would look at a hyper-converged solution like VxRail to help us solve the problems that we're having with the upgrade process.
>> And then we have this perspective that New Zealand seems to be a really nice place to be at the moment, and so we might open a data center here with me in New Zealand. And we need to think about a more global management of our multiple data centers as well. You spent some time hands-on, what's going to be really different about doing your job, if you're using this Dell hyper-converged solution with VxRail? >> The job will significantly change as I'm working through the day-to-day of, you know, first off we can just talk about lifecycle and an upgrade, doing the planning and the research on our existing hardware, which is just a Dell R730xd.
We bought them, they weren't in spec. We had to do some upgrades. I know Keith has described this a little earlier. So we had some stuff from the GetGo. Now we're running, but we don't want to have to go back and do that research and the planning for how to get to a complete validated design, everything that's in the hardware compatibility list. So we're looking at a solution like the VxRail, just first to get us over that hurdle and that's kind of table stakes with the solution as I've had hands-on with it.
I know that they can do lifecycle day in and day out. The real benefit that we see and that we're getting from that is the potential for the global expansion of being able to manage everything from there. So there's a whole lot more to the story than this lifecycle management from my perspective, and in particular we're looking at the SaaS multi-cluster management tool as a way to manage an entire infrastructure. As, like you said, we're looking to expand into New Zealand and how are we going to manage that day to day?
And how does the VxRail platform help us there? >> So this is about getting beyond a single hyper-converged cluster in a single location, and trying to unify a view across multiple clusters in multiple physical locations. >> Yeah, exactly. So it's really a hard thing to do whenever you're managing your own infrastructure and you're speccing out your own hardware, especially whenever you're doing it into another country. So what we're looking at this platform holistically and we're saying, okay, the the day-to-day lifecycle management, yeah, that's taken care of.
So how do we do health management? How do we make sure everything is running well? And how do we do that from Chicago if we're trying to manage infrastructure in New Zealand? So whenever we're looking at the SaaS multi-cluster management, that really appealed to us as a way to make our lives easier. It's got a global health map, you can see exactly on a map where your hardware is, and you can drill down into that all the way from that to an individual node to see what's going on.
What's the problem. >> So does this translate to you spending more less time getting from a problem to enough information to be able to resolve it? >> Oh, definitely. It goes into, it uses like ecosystem connectors to reach into the actual hardware itself and the ESXi layer and the vSphere layer. So it can pull errors and alerts and events off of the virtual machines and off of the host. So whenever there's a problem going on, you can check and look at this health score inside of the management software and say, okay, well, you know, this is the problem.
This is how we can resolve it. And then you can get on the case to actually fix the problem quicker because you see it all in one location and it's holistic, it can tell you the status of an individual hardware component all the way up to the vCenter level. >> So, yeah, just joining us. >> Hi. >> Curtis Edwards from Dell. >> Thanks. >> Quick introduction for yourself, please. >> Yes, my name is Curtis Edwards. I am a engineer technologist for Dell and I am part of the VxRail team focusing on the success of VxRail.
>> Seems like you've been doing a good job because the product seems has been going absolutely massively out there for customers. >> It has been growing by leaps and bounds year over year. No complaints, just constant growth, and we are constantly innovating and moving the product forward. >> So what else has been really transformational for customers about using the Multicloud Manager beyond just having that really rock solid VxRail platform underneath? >> So that's a very good question. And as Tom was mentioning, he has the ability to see on a map where all the clusters may be located around the world, whether that be wherever he is, whether that be in New Zealand, elsewhere in China, et cetera.
One of the aspects that many of our customers take advantage of is not only just being able to see where those clusters are, but to also understand what version of code each one of those clusters may be running because not every location is kept up to date at the same speed with each other. So being able to notice, oh, you know what, I have three or four different locations that are not running the most recent version of code or I have three or four different locations that's running a version of code that Dell or VMware has said that they may have a hotfix that needs to be applied for.
So being able to quickly determine where are my clusters are, what version they're running and what I need to do on those clusters to get them to the latest version is one aspect. Another hidden gem that's there is, you know, sometimes there's bandwidth constraints between the core location and your remote location. So maybe in New Zealand, you guys may have a much much lower bandwidth for being able to communicate to the internet or being able to communicate back to the core data center.
And with the multi-cluster management, you have the ability to create custom bundles or custom to the version, custom to the hardware that's located at that location, whether it be in New Zealand or in China, et cetera, so that it can take advantage of very low latency, very low bandwidth to download that bundle. >> And that custom bundle is going to be really useful for anywhere there's bandwidth constraint. because I can imagine that-- >> Correct. >> There's places like, I mean, the kind of examples we talk about is ships at sea and exploration or mining operations that are really remote.
And so that reduction in bandwidth is going to make a big difference to your ability to actually deploy changes out and not back down. >> Yeah, we have multiple customers that have thousands of locations, hundreds of locations, and many of those locations are still on basically dial up, right, are very, very slow connections, but they still have a need to have a hyper-converged solution at the edge. So being able to give them the solution and at the same time, give them the ability to create those custom bundles for their download is key to the success for them.
>> One of the other things you mentioned was that there's an ability around identifying that a specific version of software you've got has a hotfix release for it. Can you go into that a little bit because I really liked this proactive notify me before there's a problem rather than an ambulance at the bottom of the cliff. >> So slight correction. So the ability, so when you now get an alert, whether it be an alert from Dell or from VMware that there is a potential problem or VSA or whatever the case may be in regards to a certain version, then you can quickly glance into your multi-cluster management interface, your My VxRail to see, is my version affected by that or see how many locations I have that are currently running that version.
>> And Curtis, whenever I was doing my testing and looking at this, we were doing a lot of the lifecycle management with the VxRail Manager. I didn't see those options in VxRail Manager to make a custom bundle. Is that exclusive to SaaS multi-cluster management? >> That is exclusive to SaaS multi-cluster management. That is one of the perks of utilizing the SaaS multi-cluster management. When you utilize the lifecycle management within VxRail, it is simply downloading a holistic composite bundle. And that holistic composite bundle will include all the different revisions, all the different firmwares all the different drivers, et cetera, for all of the VxRail platforms, all in one holistic bundle.
And then when it goes through, its life cycle it skips what it needs to skip accordingly and applies just the components that it needs to apply. Now, when you take that same mindset and you now go to the multi-cluster management interface, it says wait a minute, why should I download all these other components that don't necessarily pertain to me, right? >> So that multi-cluster management is definitely going to give me that benefit around consistency and getting versions consistent and reducing my support cost.
Tom, you spent some time also doing some performance troubleshooting in a sort of broader scale than a single cluster. How did that play out on the multi-cluster manager? >> Actually, went very well. So a really nice feature that I liked quite a bit was the ability for SaaS multi-customer management to baseline the actual virtual machines that were running on side the VxRail nodes. So it was actually aware of all the virtual machines running there and it kept baselines based on how they've been performing.
So it worked as a very effective monitoring tool beyond just doing any sort of updates or health status. So, you know, I'm a monitoring engineer, I really like the metrics and the numbers. And I love seeing the graphs running well and I love seeing utilization that's uniform. But something that this actually did was would tell me whenever something was abnormal. So it uses some intelligence to say this vitual machine trends at 70% utilization. All of a sudden it pegged up at 90 and it's been there for a day.
So it would actually call out that anomaly for something to be detected quickly. So it's not just looking for spikes, it's looking for when things are different. >> So again, trying to get ahead and resolve an issue before there's actually help desk tickets coming in to say that the service has been degraded >> Right, yeah. So there could be a runaway service that's not going to create an issue today, or a memory leak in certain applications. So you actually see the memory growing and holding steady and maybe you can get into restart service or boot the machine before it becomes a major outage.
>> And then there was also long-term trend capacity monitoring in there as well, wasn't there? >> Yeah, so whenever you're looking at that, you're seeing this performance, you're seeing this health but also it manages the growth over time. And then it can create an intelligent estimate of saying, if you keep growing at the rate that you have for the past, let's say seven days or whatever, you're going to run out of hard drive capacity in a month. And so, if that growth rate is what you're expecting, then it'll give you enough lead time to say, all right, well, we know we're doing a migration into this, we need to just start the procurement process to add extra nodes because it's telling us, hey, you're going to run out of capacity.
And since that's a part of the SaaS multi-cluster management if you're shipping environment to the New Zealand branch that we're thinking about expanding into, if we can use that as our primary monitoring tool while we're getting it set up, so we can actually see the performance gains there as a part of this environment that we're going to be running and using already. >> Excellent. Well, thank you both Curtis and Tom for joining me in this conversation. I think it's been really interesting to see that beyond just the basics of managing a hyper-converged infrastructure with VxRail, that there are tools for getting together multiple environments.
And particularly, I like this getting ahead of problems, being able to resolve them before they turn into user visible issues both from the release of patches but also from the performance management and health management. I think it's a really interesting story. And I think people should be looking very closely at these capabilities and whether they're the right solution for managing that scale, particularly across a lots of sites. Stay tuned for more of the content around the VxRail platform and it's large scale enterprise usefulness with the CTO Advisor team.
And thank you again both for joining me. Stay tuned for more of Tom. Tom and Keith are going to have a chat in the next video about why you might want VMware Cloud Foundation. And it's more than just Kubernetes.
