 Hi everybody we're back this is Dave Vellante wikibond.org I'm joined with David Floyer and Stu Miniman also both with wikibond this is it for us we're gonna wrap right now I love this segment guys because we've had the Kool-Aid injection you know two days of Kool-Aid right into the main line vein and now we get to sit back and say okay what did we learn what do we think how does it stack up so David let me start with you this is the first time in the cube in these two days which we've been packed with just fantastic guests you've been roaming the halls talking to people geeking out listening to the software defined you know meme what'd you learn here a number of things the one thing that I'd like to focus on first is the emphasis and the quiet confidence at reducing the cost of operational cost the the infrastructure cost of managing systems in general the tagline for it is the pure systems they're putting together an increasing number of these pure systems which are really looking at the the system from beginning to end what are the costs all the way through and for example trying to take out the costs of having to put in new sets of software for everything and tie them all together so putting in versions of operating systems versions of microcode on the server on the storage on the application itself tying the lot together so taking out more and more of that really unnecessary unproductive work from the whole of the operational and lifecycle of the product what's what's enabled that David I mean we know we've known this is a problem for a decade you know more why all of a sudden now is IBM and others and others to attack that problem what's what's changed to enable that I think that what's changed is that the the technologies themselves are becoming actually pretty similar when you look at the flash architectures out there they're not they're a little bit different but they're not that different and the spinning disk is going away obviously slowly but it's still the same and processes are processes are processes you know you can vary a little bit about type of memory you have and how many cores you have but it's there is a degree of similarity in systems going in the open the open systems going in so what you're left with is okay we've got to package them better we've got to make them relevant to a particular application so they're working more closely with the ISVs themselves and and I think that is the next big slice that can be taken off there isn't so much you can take off the storage itself or the the chips it's you're down to the cost of the chips so it's the cost of all the stuff around it and I think the because of that similarity of the technologies now that the lack of differentiation the differentiation has got to come in the service as they put around it okay so you like the system view no surprise that IBM's got got a knack for that Stu what what's the highlight for you of edge yeah so that's two things that I spent a lot of time looking at this week first of course if you look at the converged infrastructure pure systems you know Dave we've really seen that this is a complex solution it's not only just architecting the technology but it's getting the go-to-market right it's getting the customers on board and pure systems has made progress they still have some software work to do some channel work to do I've heard while they have lots of flexibility and they have the phenomenal ISV ecosystem putting together all of the different pieces I heard from some channel people that said well yeah the customer can get whatever they want but I can't necessarily sell this piece of it and getting credit for it is something they need to sort out so you know we talked to people here on the cube that were saying they're having you know lots of adoption in the emerging markets you know China and Africa and the like but you know I don't think they're necessarily you know beating the other convergence players out there right now head to head but IBM sorting that out and the second one is you know really just IBM software breadth when I talked to the you know software define you know group there they've got so many different software products from DB to through all the different apps and you know the full suite of compute network and storage so they're going to do some kind of you know SDN they've got you know some really smart people aren't going on it their software defined storage strategy is really emerging out here you know IBM just has such breadth and such you know brainpower and knowledge out there that they're they're forced to be reckoned with when they put their you know weight behind the arrow yeah I mean I think for me the whole software defined thing is was a highlight of this event but specifically the open source mojo that IBM was emphasizing I mean I would think they love this whole you know discussion about software defined when you think about them from a storage standpoint you know IBM HP EMC NetApp Hitachi you know who's got more software than anybody there it's IBM no question now of course EMC has VMware they know a little bit about software but IBM is a leader in software and so the world going into a software-defined approach is very good news for IBM and IBM has chosen to compete on the basis of open source what I liked about what I heard this week is it wasn't a me too to what the EMC did an awesome job of announcing Viper it was very impressive what they put forth and the vision that they put forth and the excitement that they created but IBM didn't just do a me too they said okay we're gonna we're gonna leverage it SVC fair enough we're gonna you know abstract the hardware we can do that check but we're gonna really focus on the notion of open source as a competitive differentiator we know that game we know that game probably better than anybody else in the enterprise trying to think of does anybody know open source better than IBM in the whales no IBM's number one there so we're gonna make that the competitive differentiator now does open source is that a defining attribute of software defined storage no you can have proprietary software defined storage but the fact that IBM's choosing to compete on open source has a couple of things to me one it's got a long-term vision to it wants to collaborate with the industry it realizes it can't do you know everything on its own and three it's done this before it's done it with Linux and you know as a good street cred there so David this the definition of software defined is continuing to evolve IBM actually says there's some dissonance between what rights do between what they see a software defined and what we are seeing with software defined they I think we're saying that they like the IDC taxonomy a little better but of course I'm gonna say we like the wiki bond tax on me better I'm not sure where that dissonance is stew and I don't know if you have an opinion on that but David what are your thoughts on the whole software defined piece and IBM's ability to participate there in earnest without it put forth a good vision you know but you know let's squint through that vision so if you look at the IBM's history with Linux they they put the wood behind the arrow with Linux they put really a lot of code into the Linux base they made that work they made effectively made Linux the premier web based operating system and they drove they drove out what Microsoft from that from that place so that's a that's a pretty good track record and if you're looking at from the server point of view at the largest server installations which is where they drove them out from then they've got a pretty good track record and putting the wood behind open stack in particular saying okay they're going to help contribute to that they're one of the major contributors already to open stack was it 150 people that they have contributing to that if they continue to contribute in that way then they will use the the very rich layers of software that they have worked from a point of view of development software from point of view of management software to good effect as supplemental to a fundamental architecture and I think that's a that's a good strategy I I think that the definition of how it's going to evolve is is is not the point the point is that you're going to have a set of software frameworks environments they're going to be standard apis that are going to be used by a lot of people to put in their systems put in their their orchestration put in their management of that and that's that's the way it's going to go I think if that'll be there'll be an be a there'll be an open stack there'll be a VMware there'll be a couple more but pick your flavor pick your flavor and and and and it'll be a success so we had stew and I had Eva Helen on she's the president of symbolic you know come you know symbolic they're doing some interesting stuff and they're claimed to famous they've got this robust storage management stack now all the big guys are out now they're masters the whales are freezing the market and I asked her and she wasn't committal because you know on live tv and she really maybe didn't know or couldn't comment or didn't want to comment but I'll ask you put you in the spot so I said okay you've been developing a stack for over a decade full storage management volume management you know data management stack somebody said to be the other day it's not going to take a decade Dave to build that out and and because today's modern tools will allow you to do that faster and some of these whales can throw more resources at it and then somebody else said to me yeah but you know if that's all you have to do is a blank sheet of paper you know maybe but you got to do all this other stuff all this other integration um what's your take when you take a company like symbolic other guys like them who have a full-blown software defined storage management stack built out and then you got emcee viper you got ibm with svc you got you know hp doing its thing vmware also hp is mainly with uh with rebar an open stack right yeah but not starting from scratch no but but not 10 years mature either um can they close that gap in a year or two or is it going to take longer it's people like symbolic you mean no people like you can see the viper ibm and others uh I I think that the closed system is going to be tough I think that viper is going to they're going to be part of open stack they'll get little pieces of it but trying to create a whole orchestration layer where they're leading it uh I think is a tough road for them to hold um obviously they they own a lot of soft of storage software so in that particular area they may have some things that they can take out of that but orchestration is beyond just storage the orchestration has to be from a system level in my view so it has to be software led from a system point of view from an application point of view so so I have to say I see we're talking about open stack before right um so I'm on the github looking at the top change set contributors by employer you know who number one is for for uh uh open stack grizzly and cinder red hat good guess no uh number one solid fire yeah now they only have one or two contributors but so interesting enough but then of course hp ibm intel red hat net app rack space citrix are all up there and then emc so it's going to be interesting to watch this leaderboard right I mean uh stew what's your take on the whole open stack thing yeah uh you you know it I really like the message that ambush has that you know ibm is going to look to out execute the market not to control the market um so you know open stack is interesting of course ibm just has a soft layer acquisition to kind of counterbalance that solution uh you know the uh you know ibm is uh real involved in open daylight uh you know they were probably cisco and ibm or the the two pushing that to help sdn mature the controllers going to have tractions to um it's really early so uh the the code is now available for the controller we saw just last week big switch kind of pulled back from that a little bit um and uh you know I have to say that uh I'm I have limited hope for how far open daylight can really go um I think having an industry standard controller is a good start but uh if the the uh the project really can't get those northbound api set so that we can put any application on top of it uh you know I'll see it as a failure um but hopefully they can drive that ibm is definitely committed to work on that and you see ibm you know having lots of people committed to code um we had you know the head of the stg group talking about you know how many linux centers they have around the world and how many contributors they have so you know this isn't just lip service ibm puts a lot of resources uh behind uh what they're doing in open source uh I think we heard it across lots of the guests uh the only thing we probably talked about more uh on the cube this week was flash um it was really strong commitment and really integrated into every piece of what they said it reminded me of the hps analyst event earlier this year where you heard about how convergence and that blurring between server and storage uh you know went there so when you hear kind of that common messaging throughout it kind of resonates so flash so let's talk about flash um ibm you know big move buying buying tms bigger move maybe uh combined we're investing a billion dollars so ibm's kind of pretty clear where they're going emc extreme i o we saw at emc world what their strategy is directed availability they're kind of slow rolling it getting that right but we all agree they're going to do some serious damage in that market space hp today announced an all flash three par right looking really good we're going to you know dig dig into that unpacked a little bit del did the same last week so all the major companies now i guess with the exception of oracle but even in the case of oracle then with the hybrid strategy oh net app has have laid out and then net app just has still well i mean they've announced sort of a direction yes right yeah flash right yeah right yeah so there's maybe some acquisitions that still could be made you know uh some of our guests here predicted that but let's talk specifically about what ibm's doing ibm's basically saying look we're either looking for the classic tms sale which is the application doing the storage management will drop in tms stack light or we're going to put it behind an svc now we got wind that as well ibm's going to extract that svc stack and ultimately layer it on top of you know tms or flash systems now that's good strategy yeah so you clearly ibm's going to do some damage there so so what's your take uh david we'll start with you and then what do you have stu chime in so um i was i uh uh uh moderated two panels with six uh six customers on and the one thing that all of them were enthusiastic about was flash and and even though i was giving a panel on store wise they all uh with with one exception out of the three they all were focusing on the importance of flash um how more flash was helping them how the middle range of storage was was really not not particularly important anymore so and they were putting these flashes behind the store wise seven v seven thousand and they were putting it behind svc um they very very comfortable with doing that the strategy out there 100 microseconds wasn't bugging them yeah sure 100 microseconds overhead was not a problem it wasn't an eric i'd burn and he shared with us that those stats and you know when ibm says that it's legit you know like making stuff up so so they were very comfortable indeed with that strategy it gave them a way of moving the stuff around easily and the sdc is probably one of the not but is is the best virtualization platform that there is out there at the the largest best as because it's most robust it's got the most uh customers customers it's got the most um uh support functionality and support for third party products or not so they were able to put take the uh texas memory system and just put it behind it and automatically it had that stack and they could automatically move stuff now obviously hp have done the same thing with three par and uh del have done the same thing with the uh propellant the fluid systems uh so that there are just same strategies and eight eight hitachi also have the same uh idea of putting a flash only module uh on top of their um uh vsp so all of these uh different strategies are coming out but if you look at the quality of the virtualization the quality of the stack the ability to manage it and all of the customers were using the v7000 to manage other storage and a lot of it um it's an it's a it's a an end-to-end strategy it's more than just a flash strategy it's an end-to-end strategy where they can take the pieces they want and and dial it up dial it down across that uh virtualized environment i i was very impressed and the customers were very enthusiastic it's two it's two yeah i agree i did a panel in in in april with bunch ibm customers and very straightforward yeah it wasn't like you know mind-blowing like when you talk to fusion io customers and you go oh my god that's an unreal right and the hyperscale guys but very straightforward and practical still you talked to a lot of flash companies and customers using flash and what's your what's your angle and all yeah so uh you know there's there's a good vibe about flash at this show i've talked to the customers i've talked to the partners and they have a good understanding of where it fits um you know i i think we can agree you know ibm wasn't first to market with this but they were very deliberate as to how they did it very thoughtful put good solutions together they've got a nice portfolio that brings it all together so you know ibm now has a good good story it doesn't look like they have any you know major gaps uh into what they're offering there uh and you know they're bringing those real solutions to market so you know tms has you know those point deployments you know all flash arrays are not for every environment today um but you know they know uh i forget which guest it was that said you know we can do an analysis and we'll know you know 99 guaranteed that you know you're going to get huge benefits when you put this so uh they do you know the tco one the roys really well um they had some data as to how they're solving uh you know some of the database uh challenges out there uh which you know matches a lot of david floyer's research there so um you know good information uh good solid deployments and uh you know ibm's definitely uh you know a stronger player in flash than maybe they are in the overall storage market so let's sort of up level it around around storage for a minute because um i'd say in the last five years we've been watching this maybe six years even now um we've pretty much seen the big whales pulling companies where they have to maintain their positions yeah you saw you know oracle made a big move when it bought sun that sort of changed the landscape a little bit but not dramatically i mean at the end of the day you still get these guys you know ibm sort of hanging on to a chair maybe losing a little share maybe bumping up in a in a quarter you know hp sort of same thing the old stuff's declining the new stuff's coming on both of those companies in transition emcee kind of hanging on to its its leadership position so it's kind of been you know the rich get richer sort of thing picking up companies along the way um is software defined storage this whole open source theme you know ambuche goyal's vision is that a is that truly a tipping point are we going to see a change in the storage landscape who wins who loses i in my view the biggest impact of software led storage software defined storage is going to be the commoditization of products there will be far fewer alternatives out there it's just i think a inevitable result of that that their servers are getting you know this intel and unless they do something really stupid and and lose out to arm by by not covering the lower but lower part of the market they've made a number of moves recently to block that off and i think that's apart from the price that's that's a good strategy they put in place there it's going to be an intel environment that's that's who owns the market and in servers the storage is is becoming part of that overall software led infrastructure much more closely defined with the servers a lot of the flash is going to go up into the servers so you're you're going to see a set of functionality in moving that data backwards and forwards between the servers and the storage put itself into these platforms and the the there's the opportunity for true innovation is is i think having having gone through a period of time where there's been a lot of investment is actually going to come down and the emphasis will go away from the the platform to the isis these people who will take those platforms rewrite applications taking advantage of the phenomenal improvements in in in systems so that says guys with robust apis and open source mojo win and the guys who know how to market to isv's absolutely so that says emcee largely has to reinvent itself well it's got the vmware style of things so it's true use that okay so it can leverage vmware so vmware basically becomes you know the flagship yes okay uh so all right great and and ibm clearly has has put forth that strategy you know only four percent of hps 120 billion is software so we got to see you know what what what they're going to come up with netapp actually has inherently a good storage services excellent story absolutely you know maybe the best from an architectural standpoint because it's simple you know and now you know with new announcements today on on tap 8.2 but bot does it have the software juice right yes does it does it have the ability to scale that scale out architecture okay does does it have the capability of really making sure that things can be large things can be put on to it safely right and oracle probably doesn't really care to participate in the industry that way it just wants to sort of dominate the red stack and we miss and tell you know good solution for small mid-sized business it's right they've got a good solution at the bottom okay and okay and hitachi you know okay they'll continue to plot along do their thing and they'll make more money on selling the the ssd the the component the components there they make some really good components there as usual yeah okay so that's you know they really sort of we'll see but but they're one of the big four or five or five or six anyway but um so what nothing changes five years from now same sort of landscape i i i i have a feeling this at the moment you've got a lot of other software flash only vendors out there you've got solid fire you've got pure systems you've got pure storage i mean violin you've got you've got fusion io you've just got into the mix right you've got a half a dozen of those calamaro we've got half more than that 10 of those storage vendors with a lot of actually they've gained market share in that area no doubt about it yeah all doing well what's 20 million here it's 30 million here some of them are doing 100 million adds up it all adds up so the question is what's going to happen to to those players uh as things well they're going to consolidate right and they're going to get bought up some of them are going to die some of them with ipo and maybe they'll get bought up eventually eventually and and maybe there's another generation of flash to come out but uh but and if you look at the server market and from the server point of view the numbers have been flat etc but when you look at the winner the winner has been the people who are making the white boxes uh for google and everybody else their market share has gone through the roof and that's my view of what's going to happen so but IBM bought basically you know acquired its way into the analytics business and now it's a powerhouse yeah in analytics it's going to do the same thing in storage well will IBM get its storage mojo back by as ambush goyael says getting rid of the notion of storage as a box storage as a container and changing it to data as a platform software on top of it does IBM regain its storage mojo yes or no i i bm gains in terms of the orchestration layer and the layers above storage and the layers connecting the applications to the storage so yes they do effectively win that by learning it's kind of not a coincidence to that that there's only one pure play large storage company left and it's not it's funny you go back and you say you know what is kind of convergence in this move to software and it's turning into systems again and who knows systems we go back to the old days of the mainframe i mean you know this is it's all coming full circle right right towards IBM's wheelhouse all right good stew david thank you very much uh i think that's a wrap uh it's second year of edge you know good good job i think um major progress right doubled in size going from orlando to vegas good move uh next year uh edges in may market calendars moving over to the venetian another good move expecting further growth uh love the fact that it was not just a pure storage show you had uh we had tom ross amelia on who was fantastic we had you know plenty of folks from the the division or business unit formerly known as tivoli and uh to steal a phrase from prince and um expecting good things i i love the fact that ibm is shaken up the conversation i think it's uh clearly relevant in that conversation so hey that's a big takeaway from this event uh this is the cube we really appreciate you guys watching uh thanks so much for the crew uh mech did a great job kenny andre you know folks back in dallas uh all the guys blogging burt latimore was walking around today diana here what he uh what he learned at this event so uh i've been brought him here as a as a blogger so he's been writing ferociously so thanks for watching everybody uh this is the cube uh we'll be back uh tomorrow we'll be live at uh at hp discover so please watch it silicon angle dot com thanks everybody for watching been a great audience appreciate all the tweets we'll see you next time