 The U.S. Summit, we're here in Moscone, myself with Jeff Frick, my co-host today, John Furrier is down at Palo Alto, at Stanford, at the Excel big data event with Jeff Kelly. We've been here all day just checking out the innovations of Amazon, talking to executives, partners, customers. We're here with Noam Shendar, who's with Zidara Storage, the vice president at Zidara. Noam, welcome to the queue. Thanks. Thanks a lot. It's good to see you again. I've got our sort of old friends, Nelson Nahum, who's one of your chief technical guys, and good to see you guys off and running. I talked to you, it was a while ago now, when you were just getting started, Zidara Storage was sort of a seed, and you couldn't tell me too much about it. Right. The gloves are off, you guys are out, you're a shipping product. So tell our audience, a lot of people don't know who Zidara is. Tell us about Zidara and what you guys have been up to. Thanks, Dave. Yeah, Zidara Storage is enterprise storage as a service, and we make our service available at major public clouds, including, of course, Amazon Web Services. And the way to think about it is as a storage array as a service for mission critical or enterprise-grade applications. And at Amazon, we're positioned as the premium alternative to Amazon's native storage services. When you need something that's bigger, faster, better, that's when you turn to us to run your really important stuff at Amazon. So break it down for us. So my choice is with Amazon, I can get S3, I can get EBS. That's right. And that's it. Yeah, so let's walk across the menu of options. There's S3, that's object storage, S3 is probably the most popular storage service out there. They, Amazon announced a trillion objects, it's an amazing figure. Two trillion. Two trillion. Yes, they did it. They did it. They were at a trillion in 11, and they doubled in 2012. Two doubled. I mean, it's astounding. Two per six. Seven years to get to one trillion and one to get to the second. Unbelievable. Yeah, that is unbelievable. Absolutely. Many, many, many petabytes. Exactly. So that's object, and object is great for some things, but absolutely inapplicable to enterprise applications. It's truly archived in a low-performance storage, right? So it's not good for databases, and it's not good even for file shares where you're modifying files, because in object storage you can only upload or download files, but if you want to change a file, even just one bit of a big file, you have to delete the file and re-upload it. So then comes... It's good for pictures. It's very good for static pictures, movies also. Netflix runs S3. So then comes EBS. So Amazon recognized the need for block storage, because you need that for things like databases and analytics and so forth. So EBS provides block storage, it's quite limited in terms of performance, and it's also highly inconsistent in terms of performance, which then drove Amazon to launch EBS Provision Diops. So Provision Diops EBS provides more consistent performance, you pay for that, and it's a quality of service play. It's a quality of service play. It's still quite limited. It'll top out at 2,000 IOPS per volume, and it'll top out at 2 terabytes per volume. And just like standard EBS, it won't attach to more than one server at a time. And when you look at clustered applications in the enterprise, those require multiple server attach. So that's where we come in. We break the 2,000 IOPS barrier. In fact, our smallest, we call Zodara Engine as a 5,000 IOPS engine, and we go up from there to tens and tens of thousands of IOPS. There's no volume size limit, and we do clustered, so we do allow volume sharing. And then we also do NAS as a service, we do file, and that's not available elsewhere. So there is no Amazon service or other service for NAS as a service. Okay, so you guys saw the need for mission critical apps in the cloud. Talk about the tech behind there, the team behind there. How did you guys get started, and where did this all come from? Yeah, we're a team that's been together for 13 years since 1999. Live storage gurus have been building really complicated, highly reliable things like fiber channel storage virtualization appliances that would be in mission critical situations standing in front of very expensive sand devices. And very successful in that field, acquired by LSI in 2006, stayed as a team together at LSI, and then founded Zodara in 2011 with the same team, so we've been working together forever, we really know how to work with each other, and set our sights on how do we reinvent storage for the cloud? How do we maintain all the things that are good about those expensive bits of storage, but eliminate the negative, not make them as expensive, make them way more flexible, and multi-tenant? So one of the challenges obviously in cloud is getting the data into the cloud or moving it off the cloud. So what are you seeing customers do to sort of alleviate that challenge? Are they putting all their data and all their apps in the cloud and sort of leaving it there, or are they trying to get approximate to the cloud? They tend to put new projects in the cloud, so it is hard to migrate existing applications into the cloud, but large customers, and we have quite a few very, very large customers, Fortune 50 customers, they're using us for new projects, because then there's no data movement. So they start a new thing in Amazon or another cloud, they use us as the back-end storage for that, and then the storage grows in place. In some places we've seen migrations, multi-terabyte migrations, we allow those customers in a public port, and they can use that in order to copy the data, and then they close the public port, and now they're inside the Amazon cloud again. Can you talk about the customers, at least the type of customers, specifically what they're doing, kind of applications they're running, and how it's affecting their business? Yeah, our customers tend to be either running databases for submission-critical apps, it could be business intelligence, it could be actually running whatever their service is, like software as a service companies, and they can be running file servers, or file shares for things like CAD, or media, or for sharing, either internal within the company or outside the company, and the type of customer ranges from the very large to the very small. I mentioned we have several, we have three Fortune 50 companies, we have a few Fortune 500 companies beyond those, and then they get as small as software as a service startups that are looking for us as the way to scale. They are small today, but one day they'll be very big, and they need to get from 1,000 subscribers to a million subscribers without breaking them. And we're talking about Amazon's ability to pin quality of service. Talk about your management stack and your ability to do similar pinning, if you will, to guarantee quality of service. Yeah, absolutely. So first of all, when a customer sets up what we call the virtual private storage array, the private piece of that is that the customer gets dedicated resources. So unlike all other cloud-based services, the drives are not shared, the customer gets actual physical drives that already helps with quality of service. And then on top of that, we have the other resources, the compute and the cache memory resources that go along with that. Those also are not shared. So a customer gets a dedicated experience, even though they're in a shared environment. So that allows them to already have a predictable deterministic experience. And then we also have QoS controls inside the virtual private storage array, again, to pin the performance within that. And where do those non-shared disks live? They live at a co-location facility that is adjacent to Amazon. Amazon has a very amazing service called Amazon Direct Connect. And Direct Connect allows us to put our equipment at a very, very short physical distance away from the physical location of Amazon Clouds. And it provides us a lot of bandwidth into the Amazon Cloud. And it's been incredibly reliable and consistent. And it's a public service. Anybody can use Amazon Direct Connect. And we're one of the larger users of that service. So you work with Equinex, for example. That's right. And so they have data centers everywhere, facilities that are local to Amazon facilities, and you take advantage of that. So you've got a synchronous connection to the Amazon data center. Exactly. And actually, last week, we announced that because we're already at Equinex anyway, we have made our service available to Equinex customers, who are not necessarily Amazon customers, who wish to use our storage as a service rather than traditional storage. Now, where does Flash fit into this whole equation of missing critical apps in the cloud? Great question. We do provide Flash as one of the options. We provide both rotating drives and SSDs. And the reason this is important is there are some who are advocating pure Flash architectures, and that forces one into a specific cost performance point on the graph, which could be quite expensive on a per dollar per gigabyte basis. By mixing and matching drives, our customers can pick where they want to be. They want all SSD. Great. We provide that. But if they want to mix and match with rotating drives, or even have a pure rotating setup, that's available as well. Cache is absolutely necessary. It's really good for random workloads. Databases love Cache in particular. Sorry, love SSD and Flash in particular. So that's why we have it, but we don't force that into customers. And real quick, you had an announcement this week, right? Talk about that a little bit. Yes, we're now available at Amazon West and additional to East. And because we're now available on a biocostal basis, we have replication turned on at the same time to provide disaster recovery features between East and West or vice versa. And last question is, as we get in the hook here, talk about your relationship with Amazon and, you know, how that came about and where it's going. Oh, it's been great working with Amazon. They have no defensiveness about what we do versus what they do. When it makes sense, they send customers to us. And because it's better, it's good for both of us. They get to preserve those customers, the customers don't leave. And then we get business out of that because customers use us. And we though we had concerns before we worked with Amazon, maybe they'll see us as some sort of threat. Not at all. It's been, it's been really a joy to work with an open minded company that is business oriented. Big market. And I guess the name says it all, doesn't it? All right, listen, thanks very much. Appreciate you letting squeeze you in here. It was really a pleasure. I enjoyed it too. Thank you. All right, good. Thanks, everybody. We'll keep it right there. We're right back with our next guest. We know that analytics are important for your business. But building a data warehouse has always been way too complicated and expensive. Most data warehouses are complex, can often take months to set up and can cost you millions of dollars in software and hardware expenses. And that's before you even start using it.