Hyper-Convergence Gives Partners a Better Look Hyper-convergence is where we're bringing together siloed stacks of infrastructure, storage servers, sometimes networking, and even some other layers going up the stack, converging them together into one ...
maxta, matchett, hyperconverged
Hyper-convergence is where we're bringing together siloed stacks of infrastructure, storage servers, sometimes networking, and even some other layers going up the stack, converging them together into one solution to make I.T. simpler, to make infrastructure simpler, to make management simpler, and to really help folks that are having to look at the use of that infrastructure more able to focus on business value than focus on integrating parts and maintaining different parts.
Mike Matchett: Today, I've got a longtime hyper-convergence vendor with me, Maxta, and I've got Kiran Sreenivasamurthy. Welcome. You're the V.P. of Product Management. You're going to tell us all about Maxta today, right?
Kiran Sreenivasamurthy: Yep. Thanks Mike.
Mike Matchett: So where did Maxta start and come from? How did you guys get into hyper-convergence?
Kiran Sreenivasamurthy: Maxta started off as a hyper-convergence company. It was not a pivot from anything. We started off as a hyper-convergence company. And also our business model from day one was a software-based hyper-convergence model. Although we allow customers and partners to sell as an appliance, they integrate the software, and they sell as an appliance. The partner would sell it to the end customer. That's another go to market plan. But predominantly, Maxta is a software company and we always have been a software company and we enable customers and partners to deploy HCI software.
Mike Matchett: And an awful lot of your market, in your route to market, your go to market is through partners, right? And it's two partners around the world globally. So you actually have a pretty big footprint, maybe unknown here in North America.
Kiran Sreenivasamurthy: Yes, absolutely. And one of the other key benefits that Maxta also provides is the ability to white label the product and that's very helpful in other parts of the world where they can get into organizations where it would not be possible if it's not white label. So the partners sell the solutions as their own hyper-convergence solution which will help them to get into and sell into these organizations which they could never have tapped if that was not the case.
Mike Matchett: And we were talking in part of really providing great channel support and partner support is, of course, listening to the channel and figuring out what it is they're having issues with. One of the things you were telling me, the channel was reportedly having issues with was actually seeing what was going on technically in their clients sites, right?
Kiran Sreenivasamurthy: Yes.
Mike Matchett: I mean they have great touch with the client, great support with their client, and the clients actually like working with partners. But there wasn't a really good insight into what was actually happening on the infrastructure.
Kiran Sreenivasamurthy: Yes, absolutely. And we used to hear, or even now we hear that very often, especially outside the United States, they always say that hey. I know I have very good insight into the customers business processes, their models and everything else but I don't know how the cluster is behaving or how the HCI infrastructure is behaving, whether it's running low on resources, whether it's storage resources, compute resources. I have no idea or insight into sort of the tactical aspects of it.
Kiran Sreenivasamurthy: So that's where the MXIQ start off as a product to help them and also help our end customers and so on. So now with MXIQ, the partners can get full visibility into what's happening at their customer site and they also can get aggregated reports across the user base that they have sold. So they become more smart and they can add a lot of value to their end customers.
Mike Matchett: All right. So tell us a little bit about MXIQ. MX is for Maxta, MXIQ.
Kiran Sreenivasamurthy: IQ is for the intelligence.
Mike Matchett: So what is MXIQ? As a hyper-convergence service or as an offering?
Kiran Sreenivasamurthy: So MXIQ is a service offering by Maxta. And so what MXIQ provides is sort of the data analytics platform. So all the clusters that are connected or rather running Maxta software gets or send the data information back to the cloud. And Maxta does a lot of cloud based analytics on the data that we collect. And the server has multiple user privileges that gets assigned. Say, it could be an end customer privilege or it could be a partner privilege or it could be an admin privilege.
Kiran Sreenivasamurthy: So if it's the partner privilege, now partners get full access into the information metadata, especially metadata information of all the customers that they have sold to. They can't see other partners or they cannot see other customers but only the ones that they have sold to. So now they get to see they have more information about their cluster behaviors. The next time they call into their customer, they could say, "Hey Mr. Customer, we just saw that you were running low on capacity resources" or "We just saw that the performance on your cluster was a little off from normal, let's either dig deeper or would you want me to come and take a look and add a new node to your server, to your cluster and so on? So it provides them a lot of visibility into it. So that's essentially the current offering of both -- and this is available for both partners as well as for end customers so they get a lot more visibility.
Mike Matchett: And you mentioned before that you're very comfortable as a company and this happens a lot that you're white-labeled and that's part of your go to market. Is MXIQ also white-labeled for these partners to deliver?
Kiran Sreenivasamurthy: Yes. So partners can white label MXIQ and in addition to it, I mean there are certain, of course, the government regulations and so on, wherein they may want to deploy the MXIQ on their private data and MXIQ allows that too. And we can back this up in a VM form factor and it can be deployed locally in their own data center. And, of course, there are some caveats to it, wherein the data would not be accessible from outside but they get to maintain their data and we can share that importance on that VC to give them best practices on how it should be configured, or what are the best ways to go configure the cluster.
Mike Matchett: So we're talking a little bit about what you could do when you have that kind of data stream coming back. Obviously, you could do call home support but now you can do more sophisticated data analytics. And in particular, in an HCI environment, you could do some better job than just maybe if you were only looking at a storage solution now. Tell us a little bit about what you can do if you have data coming from the HCI level.
Kiran Sreenivasamurthy: Oh, that's an awesome question. So as we know, traditionally, a lot of storage vendors have always been collecting this information and providing some intelligent analysis on the storage part of the data that they collect. So with the HCI model, we get to collect more than just storage. We get to see the hyper visor information. We get to see the server information. So now we could do a lot of data correlation between the entire stack, whether it's the compute stack, the storage stack.
Kiran Sreenivasamurthy: For example, let's say a customer is running low on storage resources. Let's say they are hitting close to 80 percent capacity utilization. And in the meantime, they're also seeing some performance degradation. Maybe from, let's say 3 to 5 millisecond latency to say 10 to 15 millisecond latencies. Now, we could correlate the information and say, oh, this capacity -- the higher utilization of -- or low -- minimum I should say, less utilization of capacity, is it related to the compute also? And say, hey, look at the compute and say oh the compute utilization is close to 200 percent on a 24-hour basis. So maybe instead of just adding some additional drives to the cluster and expanding just the storage part of it, it's much better to add an entire node into the cluster. So now they could scale out both their compute resources as well as their storage resources.
Mike Matchett: And we talk about scaling performance, a little bit orthogonal to capacity sometimes and people criticize HCI as being monolithic blocks. But this really allows you to make some decisions to scale capacity versus performance by putting in a disk drive, putting a faster disk drive, a bigger disk drive, a compute, a core, faster core, a whole other node, so you've got to still have more knobs to switch, right?
Kiran Sreenivasamurthy: Exactly. And we could correlate and give them the right information and not have to say, hey, let's just go increase storage and that didn't work, let's go add another node or so we could give them a lot clearer and precise information because we have the matrix, correlation matrix so that we could say, look by adding another node, because we see this correlation, you will in effect increase your compute resources as well as your storage resources.
Mike Matchett: Awesome. Alright. I think we only have time for just a little bit more insight here. Where someone interested in Maxta hyper-convergence then go look for information? I assume your website has a lot of information but if you were to tell someone who's interested, to go look for something, where should they start? What would you recommend they look for?
Kiran Sreenivasamurthy: Yeah, I mean definitely, as you pointed out, Maxta.com is a good option. And in addition to it, Maxta offers a freemium model wherein the exact same software that they can request to download and they can run it in production too. I mean, of course, we limit the cluster size and the capacity that they can deploy be limited to 3 nodes and 24 terabytes in capacity, but they can go right out even on their production data. And when they are ready, they don't have to -- see, traditionally, what we have seen as they do a POC, they are very happy, and then they need to transition over, they have to do the complete data migration and so on. Here, it is just a flip of the license key and they could continue to keep what they are tested on. So it becomes a lot easier to even try it out and say after three months, they decide to say I need to buy full support for this and so on, they can flip the key and they get to use it in production with full support and so on. So the only limitation with this as support is limited, if you buy a freemium license and deploy it in production, it's just limited. But if you go with the full support, you'll get 24x7 support.
Mike Matchett: That's awesome because I'm thinking there's a lot of people out there where three nodes and 24 terabytes would take them pretty far in kicking the tires. And then when they want to get support and go to real production, it's just a phone call.
Kiran Sreenivasamurthy: Yes, absolutely.
Mike Matchett: All right. Well, thank you very much, Kiran. Thank you guys for being here from Maxta.
Kiran Sreenivasamurthy: Thank you very much, Mike. It's a pleasure.
Mike Matchett: Yeah. And thank you for watching another episode on hyper-convergence topics. I'm sure we're going to come back with some more. Take care, guys.