Grey Beards on Systems - 87: Matt & Ray show at VMworld 2019

Episode Date: August 28, 2019

Matt and Ray were both at VMworld 2019 in San Francisco this past week, and we did an impromptu podcast on recent news at the show. VMware announced a number of new projects and just prior to the show... they announced the intent to acquire Pivotal and Carbon Black. Pat’s keynote the first day was … Continue reading "87: Matt & Ray show at VMworld 2019"

Transcript
Discussion (0)
Starting point is 00:00:00 This is Ray Lucchese, Greybeard on Storage with my co-host Matt. Yeah Matt. And we're here at VMworld 2019 in San Francisco and we're here to talk about what's new at VMworld. So what's new Matt did you see here at VMworld? Well I think the the most interesting announcement during the keynote and again compliments to Pat who is far more comfortable in front of an audience than I used to see him but this project Tenzu which is a sort of a Kubernetes multi-cloud play to me I think is probably the most compelling piece of information I've received. Yeah, so I mean, VMware has been playing with Kubernetes off and on for quite a while, maybe more on than off, actually, with PKS and VIC and all this other stuff. Project Photon.
Starting point is 00:00:55 Project Photon, that was a good one, Matt. And so it's interesting to see that they're taking this to the next level. And it seems like they're almost embedding it in the system, don't you think? Actually, take it a step further I think they're embedding a panel within vCenter so that you can do your your composing and your your orchestration level through the vCenter your cluster level stuff through vCenter itself so I was in a session the other day where they talked about almost having a parallel level of functionality for containers within vSphere. So you
Starting point is 00:01:30 actually have all you have so this thing called Project Pacific under Tanzu that that's providing a certain level of capabilities and stuff like that that's that's all all container based. Yeah I find it a remarkable step in the right direction, recognizing how significant containers are going to be, particularly in new homegrown applications and that architecture. But I think also the idea that with this orchestration layer, with this management layer, you can incorporate AWS, Azure, and GCP into the same equation. And VMware. Well, certainly, well, yeah.
Starting point is 00:02:17 It's not necessarily going to do for your VMs what it does for your containers. I don't think you can directly migrate a VM to GCP. So what they started talking about is being able to actually take a VM and make it into a container. And then once it's a container, it's running under the Kubernetes cluster and stuff like that. Now you can start thinking about doing stuff like migration,
Starting point is 00:02:48 but how you get the data over there, data has got gravity and all that stuff. That's always been the bugaboo with containers, right? People claim that to containerize an application is a no brainer, but where does that data reside and what does it rely on in order for it to flow back and forth and and that's always been the issue and certainly none of us have had an
Starting point is 00:03:09 opportunity to play with Pacific or Tanzu yet but there will come a time when when these questions become clear to us and answerable right so they've talked about this project Pacific as being sort of an outgrowth of PKS so if you took pivotal kubernetes system I guess is the So if you took Pivotal Kubernetes System, I guess, is the term. And if you're running that, then you're running under Pacific would be very similar to that. Although they've got this whole management overlay, this Tanzu mission control thing, which looks pretty bizarre. But you said it's multi-cloud across the world, you know? I think it's very compelling. I think it shows that, you know,
Starting point is 00:03:47 what has historically been a virtualization platform for, if you can actually use the language, traditional server-based virtualization. It's growing. It's becoming something very much more focused for the future again, as the company was back when we were first hearing about vSphere. ESX2.1 was my first version. No, it's bizarre. They have actually, I guess it's all within Pat's framework.
Starting point is 00:04:21 I don't know how many years he mentioned, seven years or something like that he was been here as a CEO yeah and before that he actually worked for Intel for years right right he was a he was pretty high up there in a technical stream of that sort of stuff so yeah since then he's been they've been doing this stuff with Kubernetes. Like I say, VIC came out, Project Photon, PKS was running on vSphere. This is an evolution down this whole path of more and more containers. So what else is there? They talked about Carbon Black coming on board and Pivotal and all that stuff. What do you think of all that stuff?
Starting point is 00:05:01 I think it shows vision. I think it shows that for a company that, let's face it, for a couple of years VMware felt to me like all they were doing was selling licenses. And even when I worked there that was an issue that they were trying to resolve. It seems again that they're focusing on the technology at a far more rapid, far more developmental pace than I've seen them do in a long time. It's very exciting. It's a good time for this company. That's a great time.
Starting point is 00:05:33 And all this stuff with the VMware Cloud and VMC running on AWS, running on Azure, running on Google Cloud, running on IBM, it's running everywhere. I mean, you can effectively run this stuff anywhere you want. Yeah. I mean, I think even a pure data center play like a rack space, theoretically, even though it's OpenShift, could probably support this without any issues. I mean, they've got hundreds of VMware cloud providers and stuff like that. There are these organizations that are service organizations that can run this stuff anytime you want. I actually heard somebody mention about OpenShift running under VMware.
Starting point is 00:06:13 I'm not sure where that plays in a Pacific game or what. You're going to have to ask Cody about that, I think. But, no, they talked about Pulse IT, and they talked about IoT and OpenShift being more active in telcos, and they've been starting to focus on the telco arena and stuff like that. So I can't see it very far down the path. No, it wouldn't surprise me. And, you know, it really has been a boundary play, the telcos and the service providers, has been an area where VMware hasn't traditionally owned that space.
Starting point is 00:06:47 A KVM or an OpenShift or something like that has been far more a virtualization player within those areas. And I'd like to see VMware do, you know, far more significant work in those fields. They seem to be broadening their exposure. I mean, the Kubernetes thing is obviously, you know, a broader play into the more cloud-native space, and they've been playing in that space kind of on the outskirts of that. But nowadays, they're starting to look at bringing it all inside vSphere. They're looking at telcos. They're looking at IoT and edge stuff.
Starting point is 00:07:19 They're doing an awful lot more stuff than they used to do before. Yeah, and I've always found the IoT space to be intriguing, but up until this point, it's really been about what hardware do you throw at a project, right? What storage is the back end, and how does that data communicate from whatever its source is to the target? HPE made a couple of very interesting edge plays for IoT over the past recent years. But again, they were hardware plays. How does that software fit in? And I think that it's really intriguing to see VMware stepping into that space
Starting point is 00:07:55 from a software management layer perspective as well. It seems like they're playing more active in the IoT space. They're playing more active in the IoT space. They're playing more active in Edge. They talk a lot more about this whole ARM, vSphere running on ARM, or ESXi running on ARM and stuff like that as an Edge play, that sort of stuff. It's pretty bizarre. I love ARM, right? I think that anything that can be a purpose-built ESX server designed for low, maybe lots of data transactions, but low CPU utilization really
Starting point is 00:08:29 is a perfect play. It wasn't, again, I'm going back to HPE, sorry, but weren't Moonshot devices based on ARM processors? They do have devices. They have servers on ARM processors as well as DELTAs as well. But I mean, it's the edge stuff. I don't know what the edge components are. I was sitting on the floor out there.
Starting point is 00:08:49 There's a company called HiveCell that has, you know, these stackable ARM servers, which are about the size of a fairly sizable book that, you know, on one. It's like pagers you see in restaurants and stuff like that. They've got little indentations in each one that powers, they've got a power bus on one side and an internet bus on the other. You just plug in the bottom one into a 120 AC and it's your edge server. It's got everything. And they're moving it to x86 as well. It's pretty bizarre. I love it. I do. I have to say, I don't think it's so bizarre. I think it's almost an obvious play, maybe not obvious, but certainly predictable.
Starting point is 00:09:30 I think there's a lot of companies out there that don't want to put a big, fatty SX, you know, 256 gigs of RAM, and what, the new Cascade Lake processors are at 58 cores per? What does that look like? And it's certainly not a tiny little edge box. It's a serious box. I wrote an article, a blog post a couple of weeks back, on where data should be processed in an IoT edge kind of configuration. It's a lot of different play, lots of different parameters in that sort of discussion,
Starting point is 00:10:03 and where it should be processed is not actually obvious you know no it isn't and and you know i i would take that a step further and try to figure out what data actually does get utilized and what gets uh you know wheat from the chaff sort of separated uh i think that you could really consume a lot of storage if you took in everything and never got rid of anything. Being a storage player, I can understand that. I can almost agree with that sort of thing. But there's a side of me that says, you know, there's got to be some happy compromise here on what you actually process and what you need and stuff like that.
Starting point is 00:10:37 You know, I still believe, and forgive me if, again, I'm showing my bias here, but I think Splunk is the best software package to sort of gain data analytics against whatever that raw data is and give you, as an administrator, a functional dashboard. It's no longer certainly a SIM. It's way more than a SIM. And I love the product. I use it and recommend it all the time. I guess I'm not too familiar with Splunk. I've seen some advertisements for it and stuff like that. Great t-shirts.
Starting point is 00:11:11 Well, what more do you want from a company like that? So what else is new in VMworld today or VMware? I mean, there's a lot of stuff on, you know, obviously interacting, integrating Pivotal, what their expectations for that are, this whole Carbon Black thing. Are you familiar with Carbon Black at all? No. So my understanding of Carbon Black is endpoint security, so endpoint device security, endpoint workload security kinds of things. It provides malware detection.
Starting point is 00:11:42 It provides ransomware. What do you define as an endpoint though a pc or a thin client what about what about an ipad or a phone or because those are endpoints aren't they i think they're mobile i think they're mobile uh activated i think um i'm not sure about thin clients but you know pcs macs that sort of thing desktops are certainly in that space but uh yeah we'll have to see what they're doing. But, you know, Pat was presenting it's like carbon black is going to be integral in their whole security environment. Workspace one, yeah, particularly.
Starting point is 00:12:16 Which is where the endpoints would be, right? Theoretically. It's certainly an interesting acquisition. And, you know, Pat was saying something to the effect that security is ripe for disruption. I don't understand what he's saying there. Well, so, you know, if we're going to start talking about other vendors, I had a really interesting conversation with Apparetto yesterday. Have you heard of them? No, I haven't. So you were mentioning security security and I kind of love
Starting point is 00:12:46 these guys and their approach. I'm not a shill. I'm not getting paid for that comment. But essentially what they're doing is they're claiming that the application will point towards a malicious software behavior far sooner or far more accurately than would a firewall or a gateway protect against that information coming in through email or what have you. So from like a profile perspective, profiling the work that they're doing? They call it a thumbprint. I like blueprint better, but they call it a thumbprint or a fingerprint maybe. And the idea is you install an app
Starting point is 00:13:25 and you fingerprint that app so it sees all the standard. It's workload, IO, the memory and all that stuff. It's more about interactions between the app and other servers as it traverses the network and the services as well involved in that and authentication. Once the app sees something anomalous against the fingerprint,
Starting point is 00:13:48 it isolates that traffic and shuts it down. This is very smart, but it's also future-proofed to play in the space that containers play. And most traditional server firewall, router switch types of environments, they aren't that sort of dynamic. They see something maybe that comes in, or they see anomalous behavior, denial of service attack. That's not the same thing. Right. This is about the performance, not from a speed perspective, but the actual how the app is performing against its functions elsewhere on the network. So the whole profiling perspective is an interesting vector to try to defend against things.
Starting point is 00:14:40 Because, I mean, how can you fake a profile? You're running this application. It's doing certain things. It's doing it in a certain way. And, you know, if it starts changing that methodology or changing that functionality, then certainly something to be aware of. I'm not sure you can shut it down immediately at that point because, you know, it could be some specific function that's just happening. But, you know, that sort of thing over time, but you should be able to see that sort of workload differentiation over time and be able to detect something outside that framework, I think.
Starting point is 00:15:12 Well that's the goal here. And again, as I say, I like that they're thinking in a way that people really haven't thought before about some of the classic problems that our database administrators, network administrators, VMware, vSphere administrators have been seeing but haven't really had the tool sets to address. Obviously, if you're attacked by malware, you can see the after effects, but can you see it as it's happening? So maybe that's where Carbon Black fits into this framework.
Starting point is 00:15:45 Let's take that sort of endpoint profiling and move it into the workload, move it into the servers, move it into more intrinsic into, you know, vSphere functionality. And now you've got some more security types and networking and stuff like that. So maybe that's where it plays out. Jesus, is there anything else that you can think of that happened here? Well, dinner last night was a whole lot of fun. And all the parties are happening tonight, so that's the other side of this. Well, this has been great, Matt. I appreciate you being on our show today.
Starting point is 00:16:16 I always love it. Thanks. And we'll try to do this at other venues as well, and stay tuned to the next Greybeards on Storage podcast. Thanks a lot guys. Bye-bye.

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.