welcome everybody to this cube conversation my name is dave vellante and we're joined today by richard goodwin who's the group director of i.t at ultraleap and abhishek kumar who manages dell's power store product line just directs that product line along with several other lines for the company gentlemen welcome to the cube hey dave hi that's me so richard ultra leap very cool company tracks hand movements and so forth tell us about the company and the technology are really interested in how it's used yeah we have many uh product lines uh obviously we're very innovative uh innovative um and the organization was spun up from phd a number of phd students who were the co-founders for ultra leap and initially with mid-air haptics as many people may have seen but also hand tracking mid-air touch uh sense and feel uh so yeah it's it's it's quite impressive um what we have produced and the number of sectors and markets that we are in um and obviously to to push us to where we are we have relied upon lots of the dell technology both software and hardware and what's your role at the company i'm the group i t director and i'm responsible for the it and business platforms um all infrastructure network hardware software um and also the transition of those platforms to ensure that we're scalable and we are able to develop our software and hardware um as rapidly as possible awesome yeah a lot of data behind that too i bet um okay avashek you direct a number of products at dell across the portfolio unity extreme io the sc series and of course power vault it's quite the portfolio that you look after so let's get into the case study if we can a bit uh richard maybe you could paint a picture of of your environment uh some of the key applications that you're supporting and maybe what your infrastructure looks like give us a high level view sure uh so um pre uh powerschool we had um quite a a disparate uh architecture so um a fairly significant split and siding on the side of uh cloud uh not as hybrid as we would like and not uh not as much as on-prem as we would have liked and hey that has changed quite significantly um so we now have a number of servers and storage and storage arrays that we have on on-premise um and then we host ourselves so we are moving quite rapidly you know as a startup and then moving to a scale-up we needed that that scalability and that versatility and also the whole op-ex versus capex and also not being driven by lots of um sas products and architecture and infrastructure where we needed to be in control because of our development cycles and our products product development so wait okay so so too much cloud i'm hearing you run a little bit a dose of on-prem explain that a little bit more the cloud wasn't doing it for you in terms of your development cycle your control can you double click on that yeah some of the some of the control and you know there's always a balance because there's certain elements of uh our development cycles and our engineering uh software engineering where we need a very high parallelism uh for some of the work that we're doing which then you know the capex investment makes things very very challenging and not commercially the right thing to do however uh there are some of our information some of our ip um some of the secure things that we do we also do not want upgrades as an example or any outages or certain types of server and spec that we need to be quite bespoke and unique and that needs to be within our control got it okay thank you for that abhishek we're going to talk about powerstor today so set it up please tell us about powerstor what it is you know why it's important to this conversation sure so power store is a product that we launched may of 2020 roughly a little bit more than a year now and it's a brand new architecture that dell technologies released and at the end of the day i'll talk about a few unique aspects of the product but at the end of the day the where we start with it's a storage platform right so uh where we see similar to what richard is saying here uh uh in terms of being able to consolidate the customer's environment whether it is block file v-balls physical virtual environments uh and and it's as i said it's a brand new architecture where we leveraged pieces of existing products where it made sense uh and it's a it's we are using all the latest and greatest technologies delivering the best performance based data reduction uh and and where we see a lot of traction is the options that it brings to the table for our customers in terms of flexibility whether they want to add capacity compute uh whether in fact uh we have a apps on dev deployment model where customers can consolidate their compute as well on the storage platform if needed so a lot of innovation from a platform perspective itself and it's not just about the platform itself but what comes along with it right so we referred it as an ecosystem part of it where we work with ansible playbook csi plug-in you name it right and it's the storage platform by itself doesn't that doesn't stand by itself in a customer's environment there are other aspects of the infrastructure that it needs to integrate with as well right so if they're using ansible playbooks we want to make sure the integration is there got it and last perhaps uh not the least is uh the intelligence built into the platform right so as we are building these capabilities into the product uh there is intelligence built into the product as well as outside the product where things like cloud iq things like uh technologies built into power store itself makes it that much easier for the pro for the customers to manage the infrastructure and go from there thank you for that so richard what was the workload so actually you started with sort of a green field on-prem if i understand it correctly what was the workload that you were sort of building around or workloads sorry we had a um a number of different applications some of which we cannot really talk about too much um and then we engaged them regarding um the storage uh issue that we had and we engaged the our account lead accounting exec and a number of solution architects were working with us to ensure that we had the optimal solution dell were selected over the competitors there's a many reasons you know the new technology the deduplication the compression the data overall data reduction um and the guarantee uh that also came uh came with that the four to one data reduction guarantee which was significant to us because of the amount of data that we hold and we have you know as i mentioned we're pulling further further data of ours back um into our hosted environments which will end up on the power store especially with the deduplication that we're now getting we've now actually hit nine to one which is you know significant we were expecting four to one maybe five to one with some of the data types and what was excellent dell where that confident that they did not even review our data types prior and they were willing to stand by that guarantee of four to one and we've excelled that you know we've got significant different data types on on that array and we've hit nine to one and that's gradually grown over the last nine months you know we were kind of at six and we moved to seven and now we're hitting nine to one ratios that's great so you get a little free storage that's interesting what you're saying richard because i just assumed that a company that's guarantees four to one is going to say okay let us let us inspect your workload first and then and then we'll do the deal uh so avashek what's the tech behind that data reduction that you're able to with such confidence not have to and pre-inspect the workload in this case anyway yes sir so it goes back to the technologies that goes behind the product right so so we stand behind the technology and we want to make it simpler for our customers as well where again we don't want to spend weeks looking at all the data scanning all the data before giving the guarantee so we stand behind the technology where we understand that as the data is coming in we are always going to duplicate it we are always going to compress it there is technology within the product where we are offloading some of that to the uh outside the cpu so it is not impacting the performance that the applications are going to see so data reduction by itself is not going good enough performance by itself is not good enough both of them have to be together right so and that's what powerstroke brings to the table yeah thank you so richard i'm interested i mean i remember the power store announcement uh i sort of saw it leading up to it and one of the big thrusts from dell the way i phrase it is essentially trying to create a cloud-like experience on-prem so really focus on simplicity so my question to you is let's start with just the deployment you know how complicated was it to install what was that process like you know how many clicks not that you have to tell me how many clicks but you know what i'm asking is is how difficult was it to get from zero to you know up and running well we actually stepped down a very difficult challenge um we were in quite a difficult situation where we'd pretty much gone off of a cliff in terms of our iops performance um so the rfp was quite rapid and then we needed to get which whoever which vendor was successful we need to get that deployed rather rapidly and on the floor in our data center and server rooms uh which we did um and it was very very simplistic within three weeks of placing the order we had that array in our server rack and we had begun the migration it was very simple to set up um and the management of that array has been we we've seen a i think 40 reduction in terms of effort to be able to manage our storage because it is very self-contained um you know even from a reporting perspective the deployment the migration was all very very very simplistic and you know we we've done some works recently where we had to also um do some work on the array and some other migrations that we were doing and the resilience came came to came to the forefront of where you know the the dual architecture and no single point of failure enabled us to do some things that we needed to do quite rapidly because of the the dual nodes and the resilience within within the unit within the power store itself was considerable where we kept performance up it will also prioritize any disk rebuilds keeps the incoming ingest rates uh high and prioritizes the you know the workloads which is you know really impressive especially when we are moving so quickly with our technology we don't really have much time to you know micromanage the estate can you can you just repeat what you said on the percent reduction i think i heard you you cut out there a little bit a potent reduction on on management on on on the labor side yeah so our uh our lead storage engineer is estimated around 40 less management wow okay so that's that's good so actually i love this conversation because uh you know in the early days of automation people like ah that's my job provisioning luns i'm really good at it but i think people are realizing that it's actually you know not something that you want to be really good at it's something that you want to eliminate so now maybe it's a he that storage engineer got his or her nights and weekends back uh but but what do they do now when they get that extra time what do you what do you put them on you know more strategic initiatives or you know other other things in the to-do list what's that like the last thing uh you know any of my team whether it's the the storage leads or some of the infrastructure team that are also involved and engaged because you know the organization we have to be quite versatile as a team in our skill sets we don't want to be doing those bau uh mundane tasks even the storage engineer does not want to be you know allocating luns and allocating storage to physical servers vms etc we want all of that to be automated and that you know those engineers are now working on you know some of the cutting edge things that we're trying to do with machine learning is as an example um which is much more interesting it's what they want to be doing um you know that aids the obvious things like retention interest and personal development we don't want to be you know that base i.t infrastructure management is is not not where you know any of the engineers want to be in terms of the decision to go with dell power store i'm definitely hearing there was a relationship there was an existing relationship with dell i'm sure that played into it um there were many things so you know the relationship wasn't really part of this even though i mentioned the end user compute you know in any sector or anything that we're procuring we want best of breed and you know a best of set and that was done on you know cost is definitely a driver the technology you know is of interest to us we're a tech company new technology to us is also fascinating not only our own uh but also the storage guarantee the simplicity um the resilience with it within the uh unit also the ability which was key to us because of what we're trying to do with our hybrid model and bring bring back and repatriate some of the data as it were from the cloud we needed that ability to with ease to be able to scale up and scale out and the uh power stall gave us that when you say cost uh i want to dig into that price or you know the the the price tag or the the cost i mean when you do the business case and i wonder if we could add a little color to that yeah there's two elements to this so they're not either the cost and the price tag uh but then also cost of ownership and the comparisons that we were running against the other vendors but also the comparisons that we were running from a capex investment against opex and what we have in the cloud and also the performance you know the performance that we get from um the cloud and our cloud storage and the resilience within that and then also the initial price tag and then comparing the capex investment to the opex were all elements that were key to us making our decision and you know that there has to be some credit taken by the dow account team and that their relationship towards the final throws of that rfp you know were key initially not all we were just looking for the best possible storage uh solution for ultralite and to to determine that on your end was that like a feature because it's sometimes fuzzy what the business impact is going to be like that 40 you mentioned or the data reduction at nine to one when there's a promise of four to one did you what did you do did you kind of do a feature function analysis and sort of line that up and and say okay i'm gonna map that to our business pro our processes our i.t processes and try to predict what the impact would be is that how you did it or did you take a different approach we did so we did that obviously between vendors as you'd expect in an rfp but then also mapping to how that would impact the business and that that is not an easy and easy process to go through and we've seen more games even comparing one vendor to another some of that because of the the technology the terminology is very very different and sometimes you have to bring that upper level and also gain a much more detailed understanding which at times can be challenging but we did a very like for like comparison um and also lots of research but you're quite right the the the business analysis to what we needed um we had quite a good forecast uh and from some of our historical information and data and also our engineering and business and strategic roadmap we were able to map those two together not the easiest of experiences not one that i want to repeat but we got yeah a little bit of art and science involved avishak maybe you could talk about power store what you know give us the commercial what makes it different from other products in the market uh things like cloud iq uh maybe you could talk about that a little bit sure so uh so again from uh it's music to my ears when richard talks about the ease of deployment and the management because there is a lot of focus on that but even as i said earlier from a technology perspective a lot of goodness built in in terms of being able to consolidate uh customers environment into onto the platform so that's more from a storage point of view give the best performance give the best data reduction storage efficiencies uh the second part of course the the flexibility the options that power store gives to the customers in terms of sort of disaggregating the storage and the compute aspects of it so if as a customer i want to start with different points in terms of what our customer requirements are today but going forward as your requirements change from a compute capacity perspective you can use the scale up and scale out capabilities um and and then the intelligence built in right so as you scale out your cluster being able to move storage around right as needed uh being able to do that non-disruptively so instead of saying that mr customer your uh your storage is going to you're at 90 capacity being able to say that based on your historical trending uh we expect you run out of capacity in six months some small things like that right and of course uh if the uh the dial home the support assist capabilities are enabled cloud iq brings a lot of intelligence to the table as well in addition to that as i mentioned earlier there is apps on capability that gives another level of flexibility to the customers to integrate your storage infrastructure into a virtual environment if the customer chooses to do that and last but not the least it's not just about the product right so it's about the the programs that we have put around it anytime upgrade is a big differentiator for us where it's an investment protection program for customers where if they want to have the peace of mind in terms of three months nine months three years down the line if we come out with new technologies being able to be upgrade to that non-destructively is a big part of it as well so it's a peace of mind for the customers that yes i'm getting into the power store architecture today but going forward i am i'm protected from that point of view so anytime upgrade it's a new business program that we put around leveraging the architectural benefits of power store uh whether your computer requirement your storage requirements change your your you're covered from that point of view so again a very quick overview of uh of what power store is why it is different and again that's where that comes from thank you for that richard are you are you actively using cloud iq do you get what kind of value do you get from it not currently um however we have we have had plans to to do that the um uptake and how basically our our internal work node is not allowed us to to do that but one of the other key reasons for selecting powersupport was the the non-disruptive element you know with other sas products other providers and other issues that we have experienced that was one that was a a key decision for us from a um a power store perspective one of the other you know i i to go back to the conversation slightly with in terms of performance you know we are getting getting now you know there's a 400 percent speed of improvement of publishing uh we've got 80 percent faster code coverage and our firmware built 1 300 quicker than they were previously and the time savings of the storage engineer and you know as a director of it i often ask for certain reports from from the storage array when we're working out for um storage forecasts performance forecasts and you know when we're coming close to product releases and code drops um that we're trying to manage the the reporting on the power store is is impressive whereas previously my storage engineer would not be the uh the most happiest of people when i would be trying to pull you know month-end quarterly reports etc uh whereas now it's it's ease and we have live dashboards running we can easily extract that information i love that uh because you know so often we talk about the 40 reduction in it labor uh which okay that that's cool but then your cfo's gonna say yeah but it's not like we're getting rid of people we you know we're still spending that money and okay they're getting you're now into soft dollars but when you talk about 400 percent 18 1300 percent you're talking about business impact and that's telephone numbers to a cfo so i i love those metrics thank you for sharing yeah but when they obviously some of our dashboards when they're visualized that they are very hard-hitting you know the impact you know you're quite right to see if it does chase down you know the availability and the resource profile however we're on a huge upward trajectory so having the right resilience and infrastructure in places is exactly what we need and as i mentioned before those engineers are all reallocated to much more interesting work and you know the areas that will actually drive our business forward speaking of resilience are you doing any replication uh not currently however there uh we've actually got a meeting regarding this today with some of the dell's enterprise and some of their storage specialists in a couple of hours time actually because that is very high on the agenda for us to be able to replicate and have a high availability um cluster and another uh potentially power stone made because so i was going to ask you kind of where you want to take this thing i'm hearing you you're looking at cloud iq really try to exploit that so you got some headroom here in terms of the value that you can get out of this platform uh to to do replication faster recovery etc maybe protect against you know events guys thanks so much for your time really appreciate your insights and thank you for watching this cube conversation this is dave vellante and we'll see you next time you