CTO Advisor 48 - ClearSky Data revisited

Machine-generated from the episode audio and not hand-corrected, so names and technical terms may be imperfect. The audio is authoritative.

All right, welcome to the first episode of a CTO advisor for the year believe it or not So technically I did a I did a podcast with Chris wall But it was so long in the making that I forgot to record the Chris wall side of the conversation So I didn't want to publish a one-way conversation with you guys. Just listening to me X questions air X question and air So today we have a returning guest last last go ahead and introduce yourself Okay.

Hi, I'm last Becky Rita's CTO and co-founder of clear sky data. So clear sky we had you guys on about a year ago. I look back It was back in December of 2015 and we'll get an update from what you guys have done in a minute but those conversations between me and you really go well because you really get the enterprise challenges and then Translate that into your storage offering and we'll get into the storage offering How that's morphed in a bit, but first for a high level Can you walk us through what the original offering we talked about way back when and you know almost a year Year and three months ago.

Yeah, it's been it's been a while. Sure so when when we went to market initially, we were focused on offering a a storage service for primary storage what we were addressing Was a need to provide sort of the the cloud scalability and cloud economics to the on-prem market and the way we did it is via building a network so Almost a CDN type of a concept where we run our service out of points of presence in major metro areas in order to control latency and then we connect into the customer's data center and provide them with a piece of Flash and edge cache and then give them the experience of you know flash based storage but without the footprint the limitless scalability and Capability of the cloud and we would do that through our network and through some Clever caching infrastructure that we built so this was You know first of its kind it was a very unique product And I think you know we we spent a lot of time talking to customers about that We found some interesting use cases I think you and I discussed some of the things that we were seeing in the marketplace especially for really large workloads in the analytics space medical imaging Where you have a lot of processing power needed and you do have a need for performance?

but you have these really really large footprints that are very very hard to deal with with respect to the footprint and the management of them and the protection of them and You know so that was that was the initial offering and You know we we brought that out. We were initially iSCSI a lot has changed I mean I could go through though if you want. I'll go through the whole history of 2016 for you, it's it's been a Kind of a you know an interesting set of you know customer acquisitions and new feature expansion You know first of all we we landed a very big health care provider Partners health care here in in the New England area.

They're the largest operator of hospital facilities in the region, and they're they're actually a very large Organization greater than 11 billion dollars, so it's it's a uncharacteristic of a startup to get a customer like that added things like fiber channel connectivity into our edge devices so that you could literally connect to the cloud through a fiber channel network which Is was also at the times of first you know and you know it allowed that level of flexibility to to enterprises And you know we've had you know a series of customer wins You know really trying to address that primary storage You know use case and you know the interesting Discovery that we had was that there's more to it than just primary storage, so we When we were going getting these conversations we started to realize that you know a lot of the data protection features that we had built into the Service the product and the software stack They actually provide additional value that You know that customers actually want in the enterprise especially around data protection You know for example.

You know being in the cloud is the same as being an off-site backup, so let's let's let's before we tease out the Additional use cases let's recap. I think a little bit of the primary storage idea for though Quick recap or or a quick crash course in the concept of Off-site primary storage a year a lot has happened in the industry in the past year AWS has continued to just you know expand like gangbusters Enterprises are becoming less less afraid of this concept of having primary storage Off-site as you've shared you know you have a 11 billion dollar market cap use case So we're more accepting of this.

I think one of the very first things that I wrote out when I was talking about clear sky was like these guys are crazy if they think that I'm gonna host My primary data off-site We've gotten over that hurdle now I think you you guys have gotten over that hurdle that we can take a storage array Put that storage array in the cloud offer some type of Local access point to that cloud based storage and have the type of support that we expect to have from a four-feature storage array, so you know we get to the point where Table stakes have moved you know in the past year.

I think what you've described in the past year table stakes have moved Customers have said you know what yeah access to storage in the cloud. Yeah, we kind of expect this This is this is cool, but what about all of this other stuff. You know one of the things that I conversations I've had with other Data services company because what you're talking about is more than just storage is the data problem Is now that I have my storage in the cloud. I have it in clear sky I still need to back it up Then we talked about we talked actually we talked about this scenario in which we have primary storage in the cloud now I have secondary storage on Site and that kind of you know once you actually deploy it you're like wait hold on I have mission-critical in the cloud and secondary local That kind of doesn't make sense so how did you guys address that?

Initials at the customers got it is that that's cool, and then went into practice in there You know they're doing level zero backups, and they're like wait hold on this Why wouldn't why am I moving this heavy data from the cloud back on prem when I can just have it all in the cloud? How did that initial set of conversations go well? It's really interesting there? It there were two vectors the the first one You know was around backing up some of these really large workloads right so they Some of these things are actually You know they're unwieldy, right?

So how do you back up several hundred terabytes of data or or something that? Generates 10 terabytes of new data a day or something like that I mean it. That's that's actually a monumental backup challenge and a lot of that data. You can't back it up So there was that question The the other one which was sort of a bit more mundane, but you probably hear this all the time Keith It's you know it's the cost and the complexity of all these these backup products and So you know the the pain point that we keep hearing is geez can you get us out of this this backup treadmill, right?

Because it's really complicated. It's expensive and It doesn't always work the way we want it to and gee would be really really cool since you're already in the cloud to offer us these use cases and And you know we sat there We you know we you know that that was what really motivated us was like wait a second You know this is this is actually you know something that we can do relatively easily You know that data is sitting in the cloud We have this super duper high-speed network And we have all this caching infrastructure that will make it super easy to do these recoveries And you never have to see tape and it will always have the the high performance of a disk And it's kind of you know infinite in size, so Literally you could back up and retain something for as long as your pocketbook will allow you to in our world And then on top of that you guys have the network so you know you have the various locations So you can do you can you know this data locality you can position data on?

Each coast if you want or you know from a DR perspective if customers want that Not just Backup, but DR capability because we'll talk about DR in a second They want their backup in DR capability you guys already have the network to provide that that's right. That's right And and you know it's a really You know but by itself the the network is you know we're able to support the cost of that network But a lot of these enterprises building something that big and that fast It's an extraneous cost that you can never justify in the context of oh I might need it for DR or I might need it for a restore from off-site or something like that It you can't just turn on a network on a dime, and so We we basically you know allow people to borrow the network when they need it and You know that's part of the service, so it's an immense value if you look at just the costs that we're eliminating From from the the alternatives that they would have to provide So it's very compelling so let's talk about some Industry-specific use cases because you know we've in the past me and you have talked about this You know basically you guys were providing block level storage, and you weren't providing many services on top as you guys have you know the Healthcare provider that you talked about have you guys moved to a point where you're providing?

Stored services on top of the block storage, or is it still mainly hey, here's the Here's the raw loan or block storage do as it as you may We've actually so the reason we got those guys was because of their they had two sites in the in the area and They needed a better disaster recovery solution, so so it segues very nicely into like a use case that they they really were You know you know impressed with the value proposition of just you know buying the the data once in other words you buy One durable copy of the data and then being able to access it from two locations At that same cost because the alternative is I buy a whole bunch of conventional storage in location a and another bunch of conventional storage in location B, and then I orchestrate copies from one place to another and you know you could you don't have to be too imaginative to think about all of the infrastructure that that entails and so We you know what we're doing is we're giving them business continuity as part of the the block storage And so that you know we're already at a higher level of service for those guys And you know that was part of what made the TCO you know really paddle palatable for that Yeah, that is really interesting the DR concept I know we talked a little bit about this in the past and then from a data services perspective What are some of the more interesting data services that your customers have?

later layered on top so obviously you guys are providing the let's call it the the layer one physical old zeros and ones and this lessening of geographic Restrictions, and you know we talked about initially data locking etc. Etc What are some of the more surprising use cases that you've seen the solution used for? so You know we actually had You know a number of POC's That went Last year in in the health care space actually in particular where you have data sharing requirements, so It kind of almost like an Internet of Things type Problem where you have data being generated in one place, but you want to crunch it in another place Having a network and being able to access data You know in multiple locations on that network means that you don't have to carry all your processing to where the data is being generated And so that that's been a really interesting use case So it's very much analytics based workloads.

We're seeing things in in the genomic space where You know there's there's genomic data that needs to be shared Amongst institutions and the best way to do that is in the cloud so you know we have this cloud edge Capability that allows you to to mount a LAN inside of Amazon and Do these remote distributed processing things you know in in a burst mode, and then shut everything down collect your your results and go home And so you see a lot of those types of use cases where the customers themselves are able to loan out their data or a copy of their data to somebody For purposes of research sharing and then you know they they come in and they share their data To somebody For purposes of research sharing and then you know they make it go away That is fairly.

I got to tell you that's fairly interesting because I run into this use case all the time of the data Just being too heavy this to borrow the term the data gravity problem whereas You know we always like to move the compute closest to the data But the temptation is always to move the data closest to the compute because in a lot of these cases It's not as simple as moving the compute if there's a set Set types of Products or services that you use in a cloud service like AWS You know AWS has infinite capacity that I can throw against some a data set I can't replicate that at the edge I can't you know in your genomics example if I'm collecting if I do some sequencing and let's say Washington State and Then I want to process that data somewhere else.

Let's say in the AWS data center It's quite hard to replicate that compute at Washington State I'd have to build out data center facilities or let's say that I'm collecting Genomic data or sequencing data at multiple locations I'm getting multiple data sets, and I want to essentially compute against that that is a example we're moving the compute to the edge While it may take less time it is Administratively heavy to do so So it's easier just to centralize that data set in some in some data center or some virtual data center This cloud scenario you give is the perfect thing if I can say okay all eight Sequencers have at it.

Here's the directory. Here's this global file system right against this global file system and separate directories my AWS analytics once all of that data has finished his replication and made its way into A cloud and completed writing go ahead and process against it. It's like Nirvana for a lot of the researchers are I've talked to so you're saying practically you guys have done that yeah I mean, it's it's it's straightforward to do with this architecture, and you know we're starting to see You know at first when we talked about it Last year I you know I actually made a note of that one I kept thinking like you know what are the hybrid use cases and that was notably the most interesting one I mean it was You know you you don't hear the other one that I hear a lot to be honest is is DR Everyone wants to get rid of their secondary data center and just do DR and Amazon it and we can get into why that That's not going to happen For a little while because you know the various you know idiosyncrasies of Amazon But here's a use case where you know you got genuinely you know a portable workload Where the compute can be generic and you don't have to do conversions And you can just you know it doesn't matter if you can spin up your compute physically in an edge data center Or you want to do it in the cloud?

It's that's a it's a great example. You know and you know I don't know how many others Of similar you know sort of complexion exist out there, but you know we're we're thinking that this is sort of the future of Computing it feels like we're entering a world right now where a lot of data gets generated where you and I sit Not in the cloud you know especially with things like you know genomic data where people are in Chicago, right? Or you know you have machines that actually generate data in in big cities And and you know you you have to have a superhighway to get that stuff to the cloud Yeah, I'm really excited that you guys have started to leverage your network because the primary I think the primary storage idea was a great idea You know I kept looking at this back-end network thinking man It'd be really cool if I could use this as a way to transport data take write the data and just the data in one location Process it somewhere else Shrinking that that that latency and distance Without actually giving it any thought so from a File system perspective, let's get maybe a little geeky So you guys are providing the block storage.

That's right from a Distributed NFS perspective let's say this is either in us in NFS or SMB How are customers accessing? That Global file system that metadata in between are they is this just generically there? You know they're attaching this storage to a server Creating a NFS share, and that's it or is there any other considerations? No, that's that's kind of it. I mean one of the things that I think I Always have to to point out on the on the geeky level is that we have not actually Solved the problem of the speed of light right worked around it, and so it's it's really easy to Confuse yourself and say oh well.

You know now now. I'm not constrained by by geography so You still are so what we do is we we actually take snapshots of the data And when when when you're ingesting data in in Boston, and you want to view it in in Ashburn, Virginia in Amazon What you what you would do is you would take a snapshot of what you have in Boston surface that that way you can protect the two so then it just becomes a question of mounting the You know the block device and then exporting.

You know the file system that way so that's that's how you know the implementation works So yeah, go ahead. I'm sorry. Let's talk. Let's finish. Let's wrap up on ecosystem What's the ecosystem that's growing around or that you guys are integrating into when it comes to this concept of distributed data any Party third parties or partners you guys have started to work with to say you know what we have this raw capability Now let's talk about how we layer services on obviously AWS is one with Cloud Wolves the term cloud storage or the where I can attach a long directly to AWS workload, which is pretty cool.

Yeah Yeah, so so ecosystem for us, I think you know we're looking at It's it's interesting. I think I think we're You know we are working with the cloud providers and You know so I definitely consider them ecosystem partners But the guys that are really You know doing doing though that what I would call the heavy lifting of building solutions are You know the set of MSP partners that we're starting to sign up in various Locations that are you know they're they're basically rolling out the technology with their customers And then building the the next level functionality, and I think that the most valuable Ecosystem is those guys because they understand the business problems You know in great detail and so to orchestrate some of this stuff is is you know?

they provide sort of the the glue the professional services that that take the sort of what we provide is a Technology component they provide the end-to-end solution Yeah, that's near and dear to my heart. I think That is probably going to be the model for most ecosystems moving forward that Technology companies are going to give us 80% of the solution, and then whether it's the MSP the VAR or our own staff where we're building custom IP on top of that we're gonna have to move the bar forward I think In the past we've relied on traditional storage companies Traditional IT vendors to get us probably maybe 90 95 percent of the weight where their way They're expecting when it comes to infrastructure solutions.

There was not a lot of IP that goes on top You know if I bought a Cisco router You know I kind of put in the router put a Cisco config on top of it, and that was Pretty much it whereas now. I'm buying the storage and Now I have to put some orchestration some Storage services Custom to my use case on top and then move that I think a little bit closer to the business value And I think you guys are I think more or less a core less correlating correlating on Where the industry is moving from just a ecosystem in general, so we've talked about a great deal From our last conversation that we had over a year ago to now What's the one thing you'd like to leave the audience to kind of I'll reference though the first podcast And they were like though most people are oh man this solutions come amazingly far But what's the thing that you want to call out the most?

Well, I think what we are now is we've completed The vision is much more complete and so we're now sort of a real cradle-to-grave data management data services solution and you know we We offer the business continuity features. We offer all of the data protection features and We now offer the connectivity into the cloud itself, so you know we don't just use the cloud We allow you to use the cloud which is you know the the end-to-end of it So when when when you look at you know the big picture?

We're a much more credible end-to-end solution and we provide much more of the Necessary componentry in order to get the customer there, so So it's been a really busy year on the development team because of that that that development agenda And I think we've got another one of those coming You know this year as well given the the things that I'm hearing from from the field these days, so you know it's We're trying to become a more of a solution, and I think that that's the big takeaway To you know to do that.

I hope everyone gets from looking at the before and after Okay, so if people want to find out more about you more about clear sky. What can they find you? com Is our website? We have lots of information there. Please sign up. We have a weekly webinar on Thursdays Thursday afternoons Eastern Time so Anyone who's interested in getting a walk through the system every week? We do something a little bit different demo different part of the system or talk about some aspect of a use case You know feel free to sign up You know it's myself and some of my team that are conducting that and we're pretty interactive So you get to talk to us one-on-one?

and you know Obviously you know if you if you'd want to visit from one of our you know sales reps I'm happy to accommodate that and and you know reach out. I'm always talking at conferences You see me all the time as well Various places, so you know reach out to me. I'm also Twitter handle at las back lazy be came so Always happy to field questions commentary that way as well All right, and that does it for this episode of the CTO advisor you can find me on the Twitter at CTO advisor On the web hopefully you found this via the CTO advisor comm subscribe to us in iTunes again even if you don't like technology podcasts Subscribe to us because I like I like seeing the numbers go up.

We'll talk to you guys next CTO advisor