Grey Beards on Systems - Greybeards talk cloud storage with Marc Farley, product marketing for Microsoft StorSimple

Episode Date: April 11, 2014

Welcome to our seventh episode. This time we talk about cloud storage and how data centers can have both cloud and on premises storage with the proper system. We discuss all this with Marc Farley, Product Marketing Manager for Microsoft StorSimple Cloud-integrated Storage solution. Marc’s an old friend so besides talking about StorSimple’s technology we … Continue reading "Greybeards talk cloud storage with Marc Farley, product marketing for Microsoft StorSimple"

Hey everybody, Ray Lucchese here and Howard Marks here. Welcome to the next episode of Greybeards on Storage, a monthly podcast, a show where we get Greybeards storage and system bloggers to talk with storage and system vendors to discuss upcoming products, technologies, and trends affecting the data center today. Welcome to the seventh episode of Greybeards on Storage, which was recorded on March 18, 2014. We have with us here today Mark Farley, Senior Product Marketing Manager of StoreSimple, a Microsoft product. Why don't you tell us a little bit about yourself and your product, Mark? Oh, gosh. You know, I'm a gray beard.

You are a gray beard. We're actually quite happy. This is the first time that we've had somebody over 40 as a guest. Oh, boy. And as you know, my beard is gray. And I'm glad we're not using the video on this because it's real gray, and I've probably got dandruff on the front of my shirt. So that's how gray my beard is. At least you still have hair at the top of your head. That won't go there.

There is that, yeah. So, yeah, anyway, I started in the network storage industry back in 91. I was hired by Palindrome. Palindrome was a Chicago-based provider of network backup systems. And I've just ridden the storage wave ever since then. Went through, you know, the SAN and NAS development. Was at Equalogic to see the beginning of iSCSI. That was purchased by Dell. I left there and went to work for 3PAR. And of course, 3PAR, you know, was the thin provisioning and a very efficient use of large, if you will, service provider storage. And almost purchased by Dell.

Almost purchased by Dell. Yeah, there was that war there. And it was really, really fun. I had a blog there called Storage Wrap that was a wild amount of fun. My boss, Craig Nunes, was really delightful to work for. And he gave me free reign to do whatever I wanted to, which was fun and simple at that time because we were going up against EMC. So it was, you know, every day I woke up and I knew what my job was to do, which was to find something about EMC that I could knock. Of course you don't do that now.

No, no, absolutely not. I don't do that now. And, you know, 3PAR, of course, was purchased by HP after a bidding war. And then I went to work for StoreSimple. And a little more than a year ago, StoreSimple was acquired by Microsoft. And so now I'm trying to figure out how to fit in and work within Microsoft. It's a much different type of organization, of course, a large software company. I'm a storage guy. And yet I have a storage product to be the champion for. And it's been a really fun and interesting storage product.

It is a hybrid cloud storage, and its purpose in life is managing data growth for customers. It takes data that's on-site. Inactive data goes to the cloud, whether that's data that's not being used by applications, backup data, archive data, that sort of thing. And it keeps the working set of data on-site in a system that has SSDs and hard disks. So it's been fun the whole way. I've been able to work with forward-thinking technology, stuff that I think moves the industry forward, and it's been a great ride.

It really has for the last 22 years. There's certainly a lot of questions on your ability to choose the right startups, and we'll get to that later. But in the meantime... Well, I did miss a few.. We'll get to that later. But in the meantime – Well, I did miss a few. I didn't go to data domains. Yeah, you didn't miss one or two. But for the most part, you've been fairly successful in your startup life, Mark.

No actual blowups. You could call them all successful exits. Clearly three-par, more successful because of the bidding war. Yeah, that was crazy. That was just crazy. All right, let's talk about StoreSimple here. So you mentioned hybrid, cloud data storage, data storage growth, fixer, interactive data.

So what is StoreSimple actually? Talk about, let's talk hardware first and we'll talk software later. What does the StoreSimple box look like? Is there a box? There is, yeah. Yeah, It's a 2U box. It's got disk drives, a pair of controllers, power supplies, you know, a pair of network ports. So it's got fully redundant, all of that stuff. And unlike other storage systems that are scale up, you know, where you add components to it or scale out where you cluster. This is a scale across, and it grows by capacity into the cloud. So when you need more capacity, it takes the old stuff and moves it to the cloud.

Actually, I'll describe how that works later. So that's it. It's a box. It's an appliance. It really is a storage appliance. Wait a minute. You can't create a new scale mechanism. You're scale across? I just minute. You can't create a new scale mechanism. You're scale across?

I just did. You can't do that, Mark. There's too much at stake here. I know. I have a problem with this internally, too. I remember I bring this up, and it's like, is that an industry term or a mark term? I'd have to say, well, that's

a markism all the way. What is scale across? That's just because you haven't worked hard enough to make it an industry term. Yeah, apparently. Yeah, thank you, Howard. That's exactly right. Yeah, and if I could get some of these guys I work with to jump on the bandwagon, I might have some fun with that. Yeah, yeah.

But the thing is, right, it doesn't scale up, it doesn't scale out, but it does scale to the cloud. So I don't know what you call it. I call it scale across. Well, I mean, you could call it cloud scale and really confuse people. Yes, yes. And, well, you know, we did a little cloud washing here for a while, and we call it, you know, cloud is a tier or cloud storage is a tier, which kind of works.

Right. But it's not – that doesn't really describe scalability. You know, if you're going to do scale, you've got to do scale with one other word. Okay. But scale, in my mind, there's capacity, there's performance, there's connectivity, there's redundancy. All those sorts of things play into scale out versus scale up. Now, scale across, because you're using cloud as an infinite storage environment, doesn't necessarily mean that it's not scale up. It's not scale out.

It's scale really up. Yeah. Up to the cloud. So here's the thing, right? The system doesn't scale performance-wise, right? Really? Yeah, that dictates workloads. You can't add controllers.

You can't add network ports. It's going to do the amount of work that it does, but otherwise it scales for capacity, which means that it ought to be used for applications and workloads that have capacity scaling issues much more than performance ones. So it works great for things like document management, for archive, for a lot of file serving, for all those zombie VMs. It works great for that stuff. But it's not the kind of place where you put Exchange next to SQL Server, next to something else, next to media streaming and that kind of thing. It's not built for that. Although with the archiving built into Exchange now, would that be that bad an idea? It wouldn't be that bad an idea if we could only get the Microsoft towers to work with each other.

Let's not go there. We probably shouldn't go there. We don't want to get Mark in too much trouble. We don't want to get him in too much trouble. I'll just say this. There's a lot of things to do inside Microsoft, and integrating with the Exchange Group isn't one that's happened yet. Yeah, yeah, yeah. So on the back end of the Store Simple box is the Microsoft Azure. Is that the cloud storage environment, or how does that play out?

Yeah, so actually there's two things. Within the Store Simple system itself on-premises, we've got an object-based back-end. And so everything that comes in gets deduped, and it's block dedup, and it is content-addressable storage. So we create a hash out of the dedup footprint, and that becomes the name of the object. And so we've got object addressable, even though it's block object. And at that point, we manage blocks as objects inside the system. And it's the same name, it's the same object that we store in Azure. And we keep all this metadata for it, including pointers to where it's located. So an object that's located internally is not much different than an object that's located in Windows Azure or, for that matter, S3 or other cloud storage. It's just a pointer to it, whether it's internal or external.

Does that make sense to you guys? Yeah, yeah, yeah. But you mentioned block there. So each 4K block, for instance, is an object in your environment? Yeah. Oy. Well, that's an architecture we're seeing object in your environment? Yeah. Oy, oy. Well, that's an architecture we're seeing a lot of lately.

Yeah, I suppose. Exa-blocks and SolidFire basically use that same take the data in, break it into blocks, and then do content addressable storage on the blocks. Uh-huh, uh-huh. They haven't taken it so far as to go, well, now that they're objects, let's go to the cloud. Right. Which I have to admit is clever. I understand your reaction, Ray, right? So we've got NVRAM and SSDs, right? And so a lot of the metadata is put in NVRAM and SSDs. So the lookup, right, you've got two lookups that you have to do. One is the virtual to physical storage, and the other one is virtual storage to object storage. So there's two levels of translation that go on.

But we're doing it in storage and memories that are pretty fast so that lookup doesn't take a long time. It's not like we've got object all on disk where you've got all those mechanical latencies to do this multiple redirection. We do it on high-speed stuff. Yeah, yeah, yeah. You know, this is another thing, though, right? This is not a storage system that is made for, you know, Tier 1 workloads. It's just not. It's made for Tier 2 workloads.

And to some extent, that means more streaming than random IOPS? Yeah, I think it probably means more streaming. But, you know, I think we're all familiar with the bathtub curves that show it's really a lot of office automation data right you know and and you know that stuff's created and after a week it's not touched that much and after a month not very much and and after three months not there at all it's actually hard to find these graphs and it was uh oh i'm trying to remember your neighbor, Ray, Horizons Technology. Yeah, yeah, yeah. Fred Moore?

Yeah, Fred Moore. Thanks. Yeah, yeah. So Fred did the work on this, you know, years ago. And there's still a lot of data that's like this, right? And so when you talk about data growth, a lot of the problems of data growth have to do with the old stuff that just accumulates and piles up that never gets used. And that's really what's true, right?

Oh, so true. I used to have a client was a tobacco company. As you could imagine, eventually the lawyers took over. And at one point, an edict went out from the lawyers who were essentially running the company that said, we're not going to ever delete anything. All of our sins have been exposed. And when the next lawsuit comes around, we never want to again say we didn't keep that because we didn't think it was important. We just want to, because as soon as we say that, people will think we did new evil. And we're not

doing any new evil. We just have old evil. So we're going to expose it. It's all been exposed. So these guys had multi-mate documents. Multi-mate? exposed and it's all been exposed so these guys had multimate documents multi-made i have not heard that word that was a multi-made like a wang processor something like that multimate was a pc port of the wang word processor that went out of business in the mid 80s oh god but they still have files in that format that nobody can read but they can't delete i'm impressed i've heard i've heard some companies where they delete everything you know you get like two weeks of storage and your emails and it's gone so that's a that's a different game but i have a good story for that too but right the problem with delete everything is is the missing usb thing that

put in their pockets and then it shows up right yeah and uh actually this one was the email from a customer from a client going we decided not to renew your contract but you can work for us another 30 days for a million dollars oh really and and that email got deleted and then the customer refused to pay and nobody could find the email. Oh, God. Because it was past the two-week retention period. Oh, boy. Interesting. I made $28,000 searching through their exchange backups for that email.

Howard, are you an expert witness from time to time? I have done that occasionally, yes. Aren't we all? Apparently, I'm overpriced as an expert witness but underpriced as a consultant that's a different discussion anyways yeah so so that's another show gray beards on expert witness testimony no no no let's not go there let's not go there all right um so the store symbol box so if if I have one of these guys, it's redundant, it does its thing. I read and write to this with file protocol, is that, or block storage, or? All iSCSI.

Oh, it's all iSCSI. So it's iSCSI blocks. I'm reading and writing iSCSI blocks to this, and it's managing some portion of it. It considers active on site, and the rest, it's moving out to the cloud storage environment. If I have needs for more performance, I bring another one of these boxes in, and it becomes a different iSCSI volume. There's no connectivity between the two. There's no clustering, right?

There's no clustering. There's no clustering. Yeah, if you run out of processor, it's a new system. It's a new system. And you would know that you ran out of processor through the monitoring tools and stuff like that? Yeah, we do have monitoring tools that show that. And what we really try to do is make sure we sell them into the capacity-hungry environment so it doesn't become an issue. But obviously, it does from time to time.

And yeah, there's always there's always somebody who goes, I need space for this. There's space over there. I know it wasn't intended for this, but I'll use it anyway. Exactly. I mean, and I mean, there are aspects of this of this system that really require adjustment. There's different types of best practices. So, you know, Howard, actually, Ray, you had asked or mentioned, you know, we've got some algorithms that figure out what stuff gets old. It's really an access. It's really, you know, recently accessed. So there are some processes you don't want to run on the system, right?

You don't want to do a defrag. Oh, God. Oh, yeah. But, you know, a lot of, you know, modern storage, you don't need to do defrag anyway. You know, if you've got a lot of SSD, why defrag? Does all the storage get to the cloud? Here's the way it works. Everything that gets backed up gets to the cloud.

And so pretty much everything gets to the cloud. And there are policies. We call that a cloud snapshot. They're always incremental, so you don't take a periodic full, and then we create virtual fulls from that when you need to do a DR process or something. So running backup exec against it would be a bad idea too. Well, yeah, duh. And that's another thing.

We're not backup for other storage. We have this integrated data protection with the system itself, but we're not a backup target. Okay, so the data is deduped internally and then sent to the cloud as objects. So how do you get that deduplication map, let's say, out to the cloud? Are you dumping your metadata to the cloud as well? Yeah. With every cloud snapshot operation, we do an update to the metadata in the cloud. So it would know what the dedupe map would be, for instance, for all the data it holds. Yeah. So, you know, say your building burns down and you have to restore somewhere else. The first thing you do is you recover that map, and the map points to all the objects in the cloud that were, you know, that were there for that last

snapshot, and then, you know, you download those. So is there a virtual version that runs in Azure so that if my building burns down, I can recover there? There is not a virtual version that runs in Azure yet. Okay, okay, that's good. And we've agreed not to drill you too much on futures, so we'll leave it at that. Yeah. There isn't one yet, no. Would that be great if there was, though? Yeah, I would think.

I spent way too long doing consulting for midsize companies on DR. Yeah, thinking that there's a need here, right? Oh, it just makes so much sense to say, you know, I'm not going to go to some colo and maintain a whole other set of hardware just in case. But I'll spin it up in the cloud when the disaster strikes. You know, Howard, I remember talking to you about this very thing about two years ago when I joined Store Simple. We're having the same discussion again. Well, you know, it's still a good idea. All right.

I was going someplace. You know, since the Nervonix debacle, a lot of people have been talking about having multiple cloud storage purveyors. Now, not like Microsoft Azure is going to go under anytime soon, but you might want to have multiple. Do you have some sort of backup? Backup. Backend support for multiple clouds? Yes and no.

The practical side is that making a copy from one cloud provider to another tends to be pretty costly. And it's also time consuming. So practically, not really. No. Yeah. The only way I've seen that done reasonably is by putting a virtual appliance in a data center and doing the transfer without sucking it back through the customer premise. Yeah, right. Sucking it back through the customer premise is just not – Yeah, it is. Bandwidth limitations will eat you alive.

Yeah. Speaking of bandwidth limitations – What we have had, though, I mean, what actually was interesting that happened when Nutanix went under is we had a customer that had some of their stuff there and some of their stuff on Azure. Right. And then they had to move it off of Nutanix and put it back. Okay, just a quick correction. Nervonix. Nutanix is still in business. Yeah, yeah. Nutanix is a very popular going venture, and I apologize to everybody at Nutanix for saying that.

Derive was our first guest. It was our first guest. We had a good session with that. What do we categorize that under? Is that Freudian or Greybeardian or what? We'll put it down to Freudian. There's no such thing as Greybeardian. Maybe. It's our own form of senior moment. Oh, ouch. That hurt. That hurt a lot. So you've got SSD in this thing as well, and you're using this as sort of, you do, you know, let's call it automated storage tiering with cloud as one of the tiers here? Is that how it plays out? That's right, yeah. So when everything comes in, it's first written in raw form in a segment of SSD, and then it goes

through the dedu process, and it's written again to SSD. So there's two SSD tiers, basically. There's the raw, which is, you can think of it like a cache, but it's actually if no more data comes into the system, it doesn't spill over. It just stays there. And then it's deduped, and it goes to SSD again. And then the next tier internally is a hard drive tier, and it's also compressed if it can be compressed when it goes to the hard drive. So we've got these three tiers internally, two on SSD, one on HDD. And then as the HDD tier fills up, it spills over to the cloud. But we don't have to move any data at that point, though, because it's already been backed up there. And we know that it's been backed

up there. We have references for it. All we have to do is change the pointer. There's no real-time tiering that goes on. It's just pointer swap right it's it's you're just ejecting it from the spinning disk tier right we return that to the free space yeah more more like a cache and I noticed that you like most of your erstwhile competitors astute easily avoid the gateway term I like the gateway term what's wrong with the Cloud Gateway stuff? Gateway is so many things, right? And it can be anything.

It can be a piece of software. It can be hardware. And what I think we've got here really is this object storage that extends over these two spheres. And, yeah, you can say it's – but I have a hard time. To me, gateway is something that you export and import. And in this case, we're swapping objects that are the same objects. We're not exporting anything. We encrypt them and send them to the cloud, but it's, you know, there's no other there other than it is in the cloud. Right. Yeah, so the old SNA gateway where it's doing conversion. Exactly. It's just stuck.

Yeah. Exactly. There's no conversion, and that's why I don't like gateway. Yeah, it's just funny how everybody in your business avoids that term astutiously. I heard the word encryption here. So the data is encrypted on-site, on-prem, before it's sent to the cloud? Yes. We don't encrypt internally, but we encrypt it when it's coming out of the box. Customer keeps control of the keys.

We don't have them. Windows Azure doesn't have them. They're responsible for you know key escrow key management all that stuff changing keys periodically and uh but they do that and i think i think realistically that's where all of this cloud stuff is going to go is customers own up to key management and do it themselves i mean who do you who do you want to manage the keys for you anybody else yeah no yeah is that the is that a new business for Graybeards? Key management? Yeah. Actually, there's a whole discussion there. We'll go there some other time.

So are you integrated with some sort of external key management protocol, or are you just providing somebody just types it into a monitor screen and everything goes out with that key at that point. I mean. Customer enters them. Yeah. We don't have that level of key management integration. The customer enters them. At the scale of the customers you're selling this to, I find it hard to believe they'd have a Kemp server sitting around someplace to do key management. Yeah.

Yeah. I don't know enough about key management to even know what you just said, Howard. Key mip all that stuff. Yeah, yeah, yeah. Okay, all right, all right, all right. All right, so... So if I've got appliances in multiple locations... Yes.

...and I use the same key, do I get global dedupe? No. The dedupe domain... This is an interesting question that I got from Ray. What is a dedupe domain? Okay. It is a volume. It's not a system.

Actually, what it is, Ray, it's the cloud bucket where the data is going. So if the data is in, you can choose to use multiple cloud buckets. And if you do, the domain stops there. So if you've got different accounts going to different buckets, we don't dedupe across those buckets. Okay. But I can have multiple volumes going to the same bucket being the same dedupe realm, right's right okay I mean and that makes sense that's you know when you think about the cloud snapshot the cloud snapshot is the is the thing that goes up if you will that sinks all the data in that bucket right and so we don't have a way to sink across buckets okay something about yeah

so how is the so cloud storage costs you know a dollar per gigabyte per month I'm not sure what the number is how does all that build out from your perspective? So, I mean, are you selling Store Simple as a service or is it sold as a product? How is the cloud storage component of that integrated or not with your selling process or licensing process or billing process? It doesn't make sense. No, it does. It's a good question. And I think we're sold, for the most part,

as an on-premises enabler to the cloud storage service. I think, especially here at Microsoft, it's the cloud storage service is really the most important thing, and the way that customers use that cloud storage service is with StoreSimple. So there is a charge for the system, but then there's the cost of the service. And there are different ways, you know, there are different deals that go on in that. I don't want to talk about the deals necessarily. And Azure storage is, you know, from $0.04 to $0.11 per gigabyte per month. Yeah, I'm not even sure what it is right now.

I just looked it up because Ray said $1, and that was a little high. A little high. It was high enough that, yeah. That was the Greybeard's cloud storage service, you know service with highly reliable replicated objects. Now, we charged $10 for that. Yeah, I think so. That was like me saying Nutanix. Okay, all right.

Hit me over the head some more. Go ahead. So it's iSCSI only, block storage, deduped, compressed, encrypted. God, you guys do everything. Yeah, there's a lot going on. It's really a pretty sophisticated little system. What it doesn't do is high-speed transactions. You know, it doesn't do tier one apps.

But there's a lot of business out there for especially unstructured data management. So that's our slice. Right. management that uh you know so that's that's our slice right and you know and i'm sure if we were talking about a remote office and a 10 user exchange server you could handle that yeah i'm not too concerned about that yeah yeah most of that would move to the cloud as a service anyway so one should only hope you would think you would think you would think yeah so i think you know the thing is though at microsoft we talked earlier about Exchange archiving, that sort of thing. There really is this notion here that, well, why wouldn't somebody just use Office 365?

Right. Really? I have run into that. For Exchange archiving service? No, just for Exchange. Just for Exchange. Why run your own Exchange service? Right, just for exchange. Just for exchange. Why run your own exchange service? Right, right, right. And I've certainly started recommending that to people. If you've got less than several hundred

exchange users, running your own exchange server nowadays just makes no sense at all. The skill set you need to do that well is rare at that size. Okay, so I want to move on to another topic, Mark. How can you choose the proper startup to invest your time and effort and money in that's going to make a gazillion dollars at the end of the five or six years it takes to go to sale and stuff like that? In other words, what are the secrets to my success? What are the secrets to your success?

Well, let's see. The first thing is that you need to do something. You need to get on the right side of the negotiating, which I'm not sure I've always done because I'm still working. Yes, yes, yes. So there's that. You clearly did not have enough three-par stock options. I did not.

No. Nor did I have enough Equalogic. Yeah, yeah. You know, nor Store Simple. I'm still doing this. So there's that. But for what I do as a marketing guy coming in, you know, what I do is I interpret technology and I try to help other people understand it. Companies usually don't need me until, you know, until they're running ventures and have, you know, a lot of people in it. Companies usually don't need me until, you know, until they're running

ventures and have, you know, a lot of people in them. So I don't get that kind of leverage. But otherwise, your question about how do you pick a company, right? Well, with Equalogic, it was kind of easy. I was a big believer in storage over Ethernet anyway. I mean, back in the late 90s, you know, there was all this crap about, well, fiber channel's a channel and Ethernet's not a channel. It's all just rubbish. Yeah. It's just rubbish. No, they're both networks. Right. And I knew that. So, you know, and I remember, you know, there was, what was it, SOIP, there was all these different things. And I just knew that at some point, storage over Ethernet was going to work and then Equalogic came out. And my friend Don Buelens

was somebody that I knew outside of the storage industry. And the VCs in the Boston area were trying to plug him into various companies. And he would send me an email and say, what about this? And what about that? I want to say, no, no, no. And he said, he asked about Equalogic. And you know, the fact is I didn't really have anything but this antenna that went wild when he mentioned Equalogic because I hadn't talked to anybody in the company. And I just sort of knew what they were up with. And so I sent him an email back, even without any knowledge. I just said, find out who your competitors are and kill them. That's the job you want. And it was all based on five, right? Yeah. And, and, and so Don, and, and I helped coach Don through that because he wasn't a storage guy and the board was concerned that, you know, somebody that didn't know storage, you know,

could run a storage company. But my point was, if you actually know business, if you know the channel, you know, if you, if you know how to get products to market, and if you, you can be an adult managing a company, you can work. And it turned out that, you know, Paul Along on the engineering side there was terrific. And the whole engineering team at Equalogic was just outstanding. Right. And so, you know, Don was able to come into that company. He wanted to hire me and it took me about a year to get smart enough to actually follow my advice to him. You know, I bed one morning and said, what the hell am I doing? I need to go to work for Don. Once you told him to take the job, I know.

Isn't that dumb? But you guys know I was independent at the time. I love that independence. So I had to go back working for the man. And in that case, it was John Joseph. And I love working for John. and uh in that case it was john joseph and i love working for john it was it was a lot of fun and um and so then you know we got acquired by dell and there was just something about going to work for dell every day that wasn't the same yeah i remember that it can

be those big company things you were not a happy camper i i wasn't right but i had known no one known david scott and craig nunes at three par and uh and uh it was it was one of those things i was watching what was going on and and one of the things that i did at equalogic was i was working in in social media and so i was watching the space and i was seeing what emc was doing with social media and they were really nailing it. You know, they were terrific. And NetApp was doing a good job. And I was, you know, sort of looking around and saying, well, is there a place I can leverage myself here somewhere? And really, the obvious one was 3PAR. They were just getting bowled over in social. They didn't seem to have a clue how to do it. So I called Craig one morning. I just said, actually, it was after leaving a conference call at Dell. I was just pulling my hair out. The next thing I did was I dialed Craig's number. I said, you know,

is there any way, you know, I'll just do this. I'll just do this thing for you. You don't have to worry. I'll cover social for you. Don't make me go to meetings. Don't make me report to PR, you know. And Craig accepted the whole thing. And yeah, it was dynamite. And so I was just able to go in. Carte Blanche, and David was with it too, David Scott. And those guys trusted me to do battle in the social medias with EMC guys. And it was a ton of fun,

a great ride. And so that was three-part. You look for a void, right? In Equalogic's case, it was what's new and coming. In three-part's case, there was a void. It was new and coming in Store Simple, too. You know, cloud storage, you know, it's like, you know. So there were a couple of things.

But Paula Long was actually an advisor to the founders at Store Simple. So Paula had introduced me to those guys. And, you know, they thought I was the golden leprechaun. They hired me on that. Yeah, yeah, yeah. Which, whether or not it's true or not, is another story. By that time, you had reached the stage where having you around got a certain amount of visibility for a company. Just that alone.

There were enough of us that were following you. Yeah, and I would say this. I had nothing to do with this company's success. It took me a year to go to Store Simple also. When I first met Urshi, he talked to me about joining the company. I said, no, I'm not interested. A year later, I was. They had changed their business model.

They had broadened it from being a SharePoint storage solution to being, you know, general purpose iSCSI storage, which is much more interesting to my liking. But the reason this company succeeded was really, they had a very good technology concept. You know, you look at all the pieces of this, you know, the dedupe, the compression, the the cloud tiering all this stuff is really well well put together a good architect oh yeah no and and the and the whole dedupe and cloud storage you're paying by the gigabyte per month just makes so much sense yeah so the technology really lines up but but the thing i told ursheet when i first met him was save all of your own marketing money and do whatever it is to have Microsoft spend money for you. Because at that point, they were just a SharePoint solution.

So Urshit, and I think he was getting this advice from other people too, he got as close to Microsoft as he could, which meant crawling into their undershorts and really driving sales. And Urshit figured out how to make Store Simple a product that would help Microsoft Azure salespeople retire quota. Right. Yeah. And that worked well enough that they bought you. Pretty much, Howard. Yeah. It was Urshad's relentless determination to drive sales with Azure that really made this company ultimately successful.

So there is no secret other than looking for a void. Yeah, or a new opportunity. I'm not sure what those new opportunities are. I actually have some ideas for Greenfields, but I think there's a big space for private cloud storage that doesn't care about performance. It's the idea that private cloud storage will be a lot faster than public cloud storage if it's local,

and you're going to have all your – just like the store simple. Once you eliminate the network latency, the disk latency becomes insignificant by comparison. Yeah. Buy other products for tier one, but there's a private cloud storage infrastructure that can be built for tier two. Like Greybeards for storage, private cloud storage service? Something like that, greybeards.com. Come on, Ray, let's do it. We do have greybeards.com.

Well, no, greybeards on storage.com. Oh, I think I screwed up, man. Uh-oh, wait. Register a domain now. Go for it. Mark, as a fellow graybeard, I'm interested in what your take is on vSAN and the other servers. Oh, it's really fascinating. I mean, it's so interesting. You know, 20 years ago, we went through this whole process of taking storage from inside servers and

putting them outside so they could be centralized. Right. But now the technology is there that you can logically, you know, virtualize this stuff that's inside servers and is much cheaper and manage it as if it is external, you know, to some degree, right? You can manage it as a pool. You manage it as a pool. And I think it's definitely a trend that is going to grow and continue. I think the virtualization technology is there to do it. This is why Nutanix is taking off.

SimpliVity is taking off. These companies. vSANs are, I think, very exciting for the industry. I'll just have to do my little pitch. Spaces in Microsoft Windows Server 2012 has got a lot of similar capabilities for managing volumes and clustered servers. So I think Spaces doesn't get enough credit. Yeah, I mean, the reason Spaces doesn't get enough credit is we haven't really done a very good job talking about it.

And then I'm going to say something that's really going to get me in trouble. I mean, when you have a product like Windows Server that has such an incredible breadth of technology, and you've got, you know, a relatively small marketing team trying to explain all of it, and it's just because it's clumped into Windows Server, it doesn't get the exposure that I think it deserves. It's just one of the other 7,000 features. Exactly. And it's a dynamite the other 7,000 features. Exactly. And it's a dynamite feature. Yeah.

We do this MVP summit every year at Microsoft, and they hit us over the head with spaces and some of the capabilities. But when you leave there, you never hear about it, quite frankly, other than maybe the tech ed stuff. Yeah, exactly. It's pretty hard to hear about that sort of capability. Yeah. You look at what VMware does marketing-wise, and they don't have to put all the wood

behind one arrow, but last year at VMworld, vSANs were a big part of what they do. Space just never gets that kind of attention. There are other features in Windows Server 2 that probably deserve a lot more attention. But we'll see with our new CEO if we find ways to become more competitive to help people understand the technology that we have here. Yeah.

All right, so we're about running out of time here. Is there any last questions for Mark Howard? No, I think that covers it for me. Yeah, yeah. Well, Mark, it's been a pleasure having you here. I thank you for being on our call. This was fun. It's always great to talk to you guys.

Love it. Next month, we'll talk to another startup storage technology person. Any questions between now and then, please let us know. Thank you now. Bye, Howard. So long, Ray. And thanks again, Mark. Good night, Mark.

Good night, guys. Until next time.