a16z Podcast - a16z Podcast: Dumb Storage Gets Smart

Episode Date: December 19, 2014

Storage as a set of technologies in the datacenter has a conservative reputation when it comes to innovation. Reliability, capacity, speed, and cost -- those have long been the only levers to pull in storage technology. Until Paula Long had the idea to add intelligence to enterprise-grade storage. Long, the CEO and founder of DataGravity joins a16z's Peter Levine for a discussion about storage. Why (and how) intelligence is a fit for storage technology, and how this smarter approach to handling data fits in with the datacenter of the future. The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation. This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments and certain publicly traded cryptocurrencies/ digital assets for which the issuer has not provided permission for a16z to disclose publicly) is available at https://a16z.com/investments/. Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

The content here is for informational purposes only, should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund. For more details, please see A16Z.com slash disclosures. Welcome to the A16Z podcast. I'm Michael Copeland, and we are here today with Peter Levine, general partner here at Andrewson Horowitz. And on the phone, Paula Long, CEO, and founder of Data Gravity. Paula, welcome. Thank you. Peter, welcome to you too, I guess. Yeah, thank you. Thank you.

So, Paula, you guys have recently come out of stealth mode, and I want to dig into sort of, we want to focus on the future, but I want to get a little bit about the past because I know you and Peter have talked about the data center and storage in particular for a long time. So give us just a short synopsis of what Data Gravity does, and then I want to backtrack and talk about

how you guys got to where you are? Sure. So, first of all, storage is an incredibly conservative technology area. So what you see is, you know, other places where they're innovating, you know, 10 or 20 years before, storage will catch up, but it's not usually the leader. But storage is growing at such a rate. It's going to have to start to take a leadership position in sort of the evolution of how the technology works.

We first sort of noticed this, you know, as the founder, of Equalogic or co-founder of Equilogic in that, you know, storage hadn't started automating. Everything was done by hand and kind of an erector set back in 2000-2001. But if you thought about it at that point, the network had started to get smarter,

you know, not brilliant, but smarter. Your servers and operating systems were getting smarter and had been since the 90s. But storage hadn't got there. So, you know, Equalogic was all about automation and being able to grow out. Then dial forward to like 2000, 12 when data gravity was founded, and you started to see infrastructure was getting smarter and

smarter, and things were getting intelligent in the data center. The storage was still staying as a container and ignorant, and it couldn't stay that way, because it is sort of the center where everything touches in the data center. It's where the assets actually live, and it needed to start to get smarter and be able to actually provide value and mitigate risk for the things it was holding. And that's when Peter and I started talking about, you know, this is the next thing that storage is going to have to do. Yeah, Peter, from your perspective, you saw or you see, you know, plenty of companies who are kind of poking around in the space. What did you see at the time as a problem?

And then how did Paula's kind of approach start to address that? Well, first of all, I got to know Paula not through data gravity. I got to know Paula after I had, of course, everyone knows. Paula, she's the co-founder of Equalogic, which she just referenced, but, you know, Paula is the, you know, is the goddess of storage in my mind. And, you know, I got to know Paula through me coming here to Andresen Horowitz, and I saw, yeah, I would see storage deals all the time. And the first person who I would call as a sounding board was Paula. Right.

Paula, what do you think about this and what do you think about that? and the clarity and depth of her answers was, you know, second to none. It was just, you know, kind of awesome. And I'm like, wow, if I actually could have the opportunity to work with Paula, I would do that in a second. And that's really where our relationship started. When the opportunity came up to work with Paula and make an investment in data gravity, our belief was that, you know, to exactly what Paula is saying, storage had to get more intelligent. You know, storage has been about speeds and feeds, and this whole notion of providing data awareness and data intelligence into the storage array

appeared to be kind of the next generation of where storage was going. No one else was doing this. No one else is doing it. And so that was the kind of, that was the aha moment for me. I mean, to be clear, let's describe this sort of state of storage. I mean, maybe dumb is maybe the wrong word, but maybe it's not the wrong word. I mean, it was all about capacity and reliability and nothing else. I can start with this.

Paula can jump on. Look, storage has been relatively dumb. The smarts has always happened above the storage layer. And probably the biggest innovation to happen, and in storage, you know, storage has gone through a couple of transformations. It started really dumb in what's called a block device, and a block sounds pretty dumb. It's about as dumb as what block sounds like. And then, dumb as a block.

And then the next evolution in storage was the notion of files, right? And so that's when network-attached storage is typically correlated with files. and that made things slightly more intelligent. You know, a file's pretty dumb in and of itself. And since those days, the format of storage hasn't really changed. It's just been then about speeds and feeds and capacity. And this whole notion of kind of solid state disk and solid state storage, that's really been the exciting part of storage,

the perceived exciting part of storage is about, you know, going from spinning disk to solid state disk. That's been the big focus of storage. But in terms of the intelligence and the structure of what actually lives out on the storage array, whether it's spinning disk or SSD, it's still either files or dumb blocks. And what data gravity does is it takes this to a whole new level and really provides an entirely new way of thinking about storage and the intelligence moves from files to data intelligence. So, Paula, tell me then why, you know, if the intelligence in the system typically lied above the storage, or at least wasn't in storage, you know, why storage? Like, why is, does that sort of the logical next step?

Why not, you know, keep the intelligence someplace else? Yeah, that's a really good question. I hear that a lot. So what I tell people is, you know, there's sort of been this pride in storage that storage is a container, it doesn't try to poke at what's actually its container. it tries to stay neutral. So what you do is if you think about it, sort of an analogy,

so instead you've got somebody with a high-speed telescope trying to figure out what's going on in your house instead of just being in the house and looking around and telling people. So there's been this distance, right? And so when you look at a distance, you might not know who. You might know it's, you know, a person six feet tall and blonde,

but, you know, you might not know it was Robert Redford or whatever. Right. But if you were in the house and you were looking at it, you could say, well, that's, you know, that's a shelf, and that's a book. And with a telescope on sort of a blurred view, you can make some assumptions and some assertions, and then you can start to do all these really complicated, you know, multi-camera things. Or you could just stand there and tell them. Right. So it felt like just standing there in the house, we can tell you what's going on, we can see all the rooms.

We don't have to have that complexity or that guessing. So it just felt like it was time to say, you know, the data is not hidden. it's just being viewed from a very complicated, expensive, planted way. Why not just be inside and just surface it, right? If you want to know about it, just tell them. And that's kind of where the idea came from, if that makes sense. Yeah, and tell me then why, well, how long it took you to sort of see this problem and then solve it? And, you know, was it solvable years ago, or does it sort of take this continuum of,

of things moving to the cloud to make it possible? Actually, I don't think the cloud really has anything to do with making it possible. Moore's Laws helped a little bit. So really what happened is when I first thought about this, it wasn't possible. And that was when I was thinking about it in 2011 and 2012. Because if you looked at the storage architecture where it was today, as Peter talked about kind of block and file, and you tried to fit into that existing architecture, this isn't possible.

But if you think about a new architecture, and then you take all those great storage innovations around Flash, you know, multi-core processors, multi-socket, more memory, and you'd start with the new architecture, and then you have a new set of tools, you can just build a better house, right? right um and so we need a bunch of things to come together but first you had to get over the fact that the storage architecture couldn't stay the same it couldn't get smart and stay the same um and once you do that it's kind of like you can't raise the roof on a ranch house that the foundation isn't solid right it's going to stay on one floor building um until you fix the foundation or until you pull it out and put in a new one so in our case we had to pull out the foundation and put in a new one and once you thought about how that was going to look then building became much easier but you had to make that quantum leap.

So, Peter, for you and then, Paula, I want you to add to this, but, you know, we talked about looking around a house and seeing what's on the shelf, here's these actual specific books, but if I'm a company, like, who needs this and what does this start to allow me to do? Oh, great. So one of the things allow you to do is you come much more, so people talk about having terabytes to petabytes of storage. They save it forever because they don't know what's in it, and they don't know how to clean

it up. So the first thing you want is some file analytics. So you can actually, at a pretty fine grain say, you know, these are this kinds of data, this is this kind of data. This has value to me because, you know, it's got these kind of, it's got IP in it, it's got contracts in it, knows about customers, but it doesn't have value to me. It's the videos that we took 10 years ago that we're never going to use. So you want to get a way to map what's in the house and be able to use that space efficiently. Right. The next thing you want to do is you want to get the upside from the data. So you want to make sure you can find and reuse and quickly. You want to make sure that you can correlate data. Like, tell me about

what's going on, you know, if this was insurance, between, you know, what are the claims coming in and my unstructured data that would tell me that I've got a mold issue or in some states or what would tell me about how POs are tracking to part numbers and sales are up as I look at my unstructured data. So how do I get upside on the data? But as importantly is how do I cap my downside? So as you hear about all these data privacy issues, small companies have the same, and mid-sized companies have the same issues with data privacy as larger companies. It's just not as well publicized, and the fines are the same. The fines are the same. The state and the federal government don't care whether you're a Fortune

1,000 company or you're a 50-person company when it comes to allowing, like, personally identifiable information to be accessed inappropriately. And so you really need to make sure that you're managing your unstructured data properly because ignorance of compliance is not compliance, and the fines are probably more than the cost of the storage. So understanding and using your assets, well, making sure you leverage the assets you have in a positive way, and making sure you tap your downside. And your storage can help you do that in a way that doesn't require you to spend any extra time managing your storage,

and that's what sort of data gravity brings to the table. So, Peter, when you were looking at data gravity and this problem, again, like our customers, they're asking for all those things you just described, Paula, you know, security, privacy, et cetera, but they don't look forward from their storage. Peter, how did you kind of view the marketplace and its readiness for something like this? Right now, the way people solve this is through a host of ancillary products, you know, as Paula describes it, they buy telescopes to look into that. house and binoculars, telescopes, monoculars, whatever, right? And that becomes the, those are all expensive outside of the box products that people typically buy. It might be data analytics products. It might be, you know, security products. And my view was, is that with Paula behind this, inventing, really what I see is the next wave of storage innovation, I think there's really

the opportunity to provide real insight and value from inside the house and eliminate all the complexity and all these additional products that one needs to buy to get half the capability that data gravity is providing. So look, there is an education, but the value proposition and the elegance by which data gravity is solving the problem, I really believe that this, what data gravity is doing, becomes a property of all storage into the future. That is, everyone is going to build some form of intelligence into storage. Data Gravity, of course, has a, you know, hugely, they're the only company doing this.

It was deemed as being impossible to do. And, you know, I really think this becomes one of the waves of future innovation in storage. So let's say that it does become this foundational thing. thing, and that storage has intelligence, you know, by and large, how does that build on the other intelligence in the system? And, like, what changes? I mean, do we start to see, you know, our ability to do things change and, you know, or do we just get to do what we've always wanted to do and do it better? I think you're going to see intelligence at every layer. I mean, Peter talks about, you know,

the internet of things, and I saw sort of the amount of data that's going to be coming in. So I think your networks are going to have to get smarter. Your storage will get smarter. You know, there'll be some intelligence in the applications as well. But the storage is sort of the last bastion, if you

will. And it's going to give you a more holistic view than some of the other ways in. But I think you're going to have to see the entire infrastructure get smarter and it's going to have to define the metrics by which it provides better levels of protection and better level of information and it will be about how it touches the data.

I mean, and I want to ask this of both of you, Paul, I mean, it doesn't sound like you've built a storage company, honestly. I mean, storage obviously is a part of it, but it's writing on top of storage. Like, how do you, will people be building storage companies, you know, and I'm doing air quotes around that word storage now in the future? So it really depends on how you define a storage company. I think, you know, if you define the storage company, you know, if you define the storage company, more broadly and say people are going to be creating companies that not only hold but inform on the data and protect it,

then, yeah, if you broaden the view of a storage company, but nobody's going to be building a traditional storage box company anymore. Storage box companies are going to become very much a commodity. It's going to be an open source stack, and it's not going to provide a container, much like you can go up any highway and find a container you can rent for $100 a month to put junk and you don't want. But other companies are going to be buying, you know, things that are adding value and containing data you care about and data you can look at. Peter, are you going to be seeing more storage deals, do you think, going forward?

We always look at new storage deals. Storage is a very, very large market, right? There's all kinds of different aspects of storage. So what Data Gravity does is one part of the storage market. There's others back up and design. disaster recovery and all kinds of other pieces. And so, yes, we continue to look at storage. And anything we look at, I continue to ask Paula about.

So that part I definitely know. Right. What are you asking Paula about these days? Well, most of the questions to Paula have nothing to do with any other company. It's got to do like, how are sales going, how's the engineering going, where's the product, how come you're late, blah, blah, blah, all that stuff, all the typical, all the questions that are, you know, sort of the practical questions about running a company. Well, Paul, as you...

Go ahead, sorry. You're going to see other companies, like, if you look at, like, a co-host data who are doing things at web scale and removing some of the network problems as you get to web scale kinds of storage and big end the storage, you're going to see guys like that do really well. And you'll see guys like Maxinta, who are going to give you a software-defined storage for the places where converge work. So it's not going to, you know, if you're a primary storage endpoint,

for small and medium business and you're focused on structured data data gravity is going to be your play but there'll be some other really high-flying very important storage innovations are going on right now and if Peter agrees or not but well one that you mentioned coho or also an investor

he's going to have to read that okay I can be careful there you're right that's fine so you know Paula from your vantage point and as you are building out data gravity but as you see things other things built around up around you in the data center. And Peter talks a lot about this, too.

I want you guys both to describe sort of the data center of the future as you see it evolving. Well, I think the problem is the data center of the future are going to be trying to solve are very different than the data center of today. So I think what you're going to see is the data center of today is trying to get to be more bifurcated into a, you know, sort of an on-prem, private, and public cloud, depending on the type of data, the value of the data. It's kind of like Numa used to be when you were doing computers, where you'd want some close stuff near, some stuff, sort of close and stuff,

some stuff could be far. And that's how your compute and memory used to work. And you're going to see storage do the same thing, but it's going to be more seamless. And the next generation of that is going to be not based on, you know, algorithms that have, you know, how many times, you know, was this touched or was this hot, but it's going to be done based on the value of the content and who was reading you when they last read it. Interesting. There's a lot of false positives you get when you just look at it as a block level.

You know, Peter talked about you can't get dumber than block. You can't get dumber than tiering based on a block. You really want a tier based on the content and the value of the content. And that tiering can include sort of location, et cetera, speed. Yeah, all of that. So you're not religious, it sounds like, about, you know, on-prem versus, you know, remote or cloud? Nope, I'm not religious at all. In fact, I think Metro clouds for the small and medium businesses are going to be,

those tend to be local cloud, I mean, a service provider is providing on and off-prem services for their customers. I think that's the way most sites will go, or some will be on-prem and some will be in a Metro Cloud, and then some will probably make it to the public cloud. It will really depend on the value. and the nearness of the data has to have and the access rates for the data. Right. So you don't want to have one map of it so you can find it.

Right. Peter, how does data gravity then fit into your sort of view of this data center of the future? Well, I think Paula's view that this idea that storage, you know, the tiering of storage is going to be tiered based on the importance and value of the information as opposed to what's been read and written most recently is a really interesting, compelling kind of difference for the data center of the future, whereby, let's say, things that are less important to the organization,

maybe those things find their way out in the cloud, information that's really important, maybe things that need to be highly secure, find their way on-prem or in a secure data center. So the idea of flipping storage sort of upside down and not even calling it storage, calling it data information. And that information living in the most appropriate place for where the user wants to consume it

is very much along the lines of the smart data center of the future. And it requires innovation like what data gravity is doing in order to be able to go and do this. So I think it very much fits into the notion of the smart data center and really kind of having data live where it's most appropriate for it to live. Well, Peter, thank you so much.

And Paula, thank you as well. This has been really interesting, and I won't look at my dumb storage and my laptop quite the same way anymore, I think. All right. Thank you, guys. Thanks, thanks. Thank you.

Bye.