HPE Synergy Coffee Talk - HPE Discover

53:29 · Watch on YouTube ↗

Transcript 6,323 words · about 42 min to read

Auto-generated captions from YouTube, not hand-corrected, so names and technical terms may be imperfect. The video is authoritative.

and I've been involved in close ability and some of these kinds of topics many many years and you know I think let me see if I could frame is because I think you see that the show forest is all about these for transformation areas and all those kinds of things why right well i think the several thesis isn't it it's really all about the absent dated Doug it's all about the acid a pretty obvious kind of the statement and beg talks a lot about the idea cottage the notion that in this digital world if you have an idea in the past you had any kind of brick and mortar and an dad would call and set it up and configure it bring it in and you all kinds of backflips ela to actually move that forward from an information processing perspective in today's world no idea directly to a digital invitation in a very fast mover leg and part of what we need to do is work that idea come that notion digitizing our planes or trains or automobiles we're putting sensors and intelligence not just in the data center but out into the world is we have to build products software services to enable that idea economy to enable the applications in the data to flow and to really allow for business to drive to the speed they need to driving eat in this part so as we think about that new economy that new style of business it does require us to kind of think outside the box not so much about the technology and the fiscal bikes but about the outcomes about none of the outcomes but how do we ensure that those outcomes still open up the new forms of risk because again we're digitizing everything one of my latest projects across you look back at enterprise was working on the Internet of Things anybody notice the digit overlooked the windmills and stuff like that Brad was one of the solutions center we worked on in this notion that we're putting sensors in our car rentals they're putting in telemetry if we're connecting it to big data so that we can get actionable insights and control things is as part of that idea economy it's also part of the challenge that we have proactively managing risk it's so it's also part of the challenge we have in not just understanding how to connect the things when how to connect the things contextual right I know your location do something about you I and deliver a better experience I can deliver more content more value because of that context and that does require that we start to think about digitizing everything connecting everything and moving this thing forward so you know we can call this this new style business and that requires a new style of ideas a stout almighty bed requires some transformation if transformation as we all know is a stepwise process it's a process that you start and you sort of move forward in what I would call evolutionary model fits and starts it makes a big progress then for a while appleton you leap forward again if you look at what we've been doing in infrastructure for instance we've been talking about convergence from quite loud we make some big leaps in the way that our computing power storage our networking work together as a system work together to provide a more agile environment and then we started in refinement I'm starting to refine it in a variety of big in small ways now we're taking another lead with composability okay and that leap will again continue to get refined continue to get richer and more focus and the for transformation areas we talked about and you've heard a lot about the whole show is really worried about the ground or this notion of transforming to a hybrid infrastructure hybrid is Lao hybrid infrastructure thinking the ability to do service delivery at your site and managed VPN site or in public site to be able to protect those digital assets no matter where they are whether they're in the data center behind your firewall whether they're in a public or managed cloud environment whether out the wild in an Internet of Things title capacity naval in the workplace which is all about this notion no matter where I go in what I'm doing expect to be connected expect to be able to actually connected would be contextually connected and two if I opted in to have that location and that information use in service of the things that I want you need as I proud from my business to a consumer line and then this notion of empowered of the data-driven organization the notion that the analog the digital the machine to machine the human data that is being collected as we're connected and interacting with all of these digital environments is not only useful but critical to business outcomes and business decisions to better experience of our consumer in a business perspective you can see that the Internet of Things with the enormous interest that is fueling particularly within the industrial environment we talk about the Internet of Things this being the Internet of very interesting things if your wristwatch goes down not a big problem if a general electric turban goes down big problem right big cost in terms of Wisconsin excellent so predictive maintenance and context-aware capabilities are becoming not just interesting but a core part of what our business customers are looking for as they digitize their environment and the transformation area that we're really focused on with synergy is really this notion of transforming to a hybrid infrastructure environment now when we talk about hybrid infrastructure as we talk about that it's a journey right many of your readers many of the folks that you engage with then on that journey for a few years right starting with consolidation started with virtualization starting with converging different aspects of physical machines and our logical abstract components that can be more automated things as we move forward that next step is then to automate and orchestra to introduce more and work automation faster and faster just before the session I did a session not in the theater about some of the work you were doing within our own days we're on the light rail infrastructure as I say I manage for 50 thousands of data centers Houston for Collins Colorado to come back many of our rd developers said and their expectations are include of a catalog I can pick out the resource that I need where that's a bare metal machine whether that's a virtual machine whether that's a container us machine I can get that machine set up the way that I needed Sakura compartment and I can start to do my words and not only can I do my work and get that resource very quickly provision but as a developer you're going to help me automate my entire continuous integration continuous delivery so my build times go down I'm more productive and I can get more innovative products out the door faster so automation orchestration cloud-like behaviors are becoming more and more not just something that some people want but something that most people expect as the develop new products as the interactive digital environments and fundamentally we're on a path with hybrid infrastructure not just to converge and personalized such as the bottom eight orchestrate but also to transform the way in which applications and infrastructure and services are delivered in a very fast fluid way now we have about 300 plus it's given three thousand plus private cloud installations around we've learned an awful lot of what it takes to do some of this automation this orchestration and we a lot of scars quite frankly both internally with our own developers to try to do this and sort of externally with a lot of our customers today I'd say we're probably the leader in private cloud but our strategy is in private cloud our strategist hybrid cloud and notion disorders delivery is a choice is an opportunity to all organizations to say mmm I have set of applications that really require the data promising the security Mission Creek County exciting to run a bio data center you manage in a managed hosting environment or I've got the needs for very fast development systems have engaged mobile applications web scale applications if I want to take advantage of Microsoft Azure AWS the public cloud environments but this is a question of this or that it's a question of both it's a question how do I use all of those loads of service delivery that have different price points cost factors security factors for clients backers all the agility and abilities that I to deals with every day how can I use all of those things to really optimize my service delivery and optimize efficiency enable faster ability to respond to specific competitive threats or opportunities and fundamentally disrupt your change the game in my particular industry and so hybrid cloud is an enabled hybrid infrastructures and when we look at synergy is a platform it's really the culmination of many years of work experience of building virtualization convergence automation orchestration into our products into our services and then saying oh what's the next step what's that next step beyond virtualization that it's going to allow for even greater even greater support of traditional and some of the new more containerized applications that are now in favor in most development shuttles get back to our own R&D organization we've got to maintain a variety of different software platform we've got tests to build those platforms and make sure that you're doing secure development and delivering secure products at the market we also have to race toward these new our microservices architectures that are more componentized and that are delivered and more of a container or musician and so when we look at a platform like synergy it's a platform that really allows us to compose infrastructure computing stores and fabric in a very fast fluid leg bring that up install the OS install the containers install the virtualization pulse and get it ready for application development application in different kinds of departments and when you look at synergy you might say well it kind of looks like some of the other products that you've seen your past and you know it's got blinking likes lots of candles few cables in the stand when you really double a triple click on some of the capabilities what you see is some pretty big differences you see some pretty big differences in the way that that one infrastructure of traditional these more cloud native applications as pools or functional resources that I can provision that I could release that I can glue refactor that I can deliver to developers choices of mine in a bare metal machine with linux OS oh no idea a bare metal machine with windows no I need a hyper-v machine no I need a hyper-v machine with a container and all of those different choices are now available in one platform called synergy that allows us to fundamentally carry forward this notion of hybrid infrastructure into the next day we're very excited about this this isn't the end of the journey of course this is the continuation of that journey the continuation of convergence now possibility and we don't do that alone right the other thing you see on the show floor and then I really hope that you take a look at is the fact that because synergy is based on these open a Piazza is open integrations we can work with companies like dr.

like chef against the black puppet the cloud providers like the zoo or the OpenStack community to marry up this notion of sure with orchestration of automation all the way up into the application so that if i'm working in chakra puppet i can use what's called the one view driver i can insert that into my chef recipe i can configure the bare metal machine the OS and the applications in a very fast augmented play particularly coming to DevOps irrelevant that's a very significant opportunity I can do that with helium and so I can insert that one view driver that represents the one view composable api's right into the Ilia flavor stack and I can bring up a variety of different machines very very quickly and a lot of different options again her mental virtualization container decision one size doesn't fit all either from a infrastructure perspective from a development OS applications perspective so what we're trying to do is take that next leap forward into composability of the infrastructure relationship with the application a very automated way and then carry that forward into that hybrid infrastructure environment so our customers have the choices of public and private be an advantage depending on cost security compliance availability other aspects right and to be able to do that in a way that supports not just new applications for retail environments but supports also some of their existing environments so that they have one infrastructure of that vastly simplifies and Vaseline accelerates their drive or disney digital me questions yeah rable questions and tell more of a strategy question back I OT where does HP come from machine learning sector we see a lot of part we're going out I talked about McDonald's great Hardware let our leverage the data that's on it yeah great question Machinery is a really hot coffee and years and years ago ten years vishal telekis may I was then it became doctor and that's cool okay and back back in those days fees to call it elegant operatives and over the last couple weeks 10 years we've made some enormous enormous prize in what's called deep learning machine learning and that's based on elegant operators and what you will see HP do is work on our elegant algorithms really from top to bottom by that is from the machine from basic from the fpga level and where you can really start to do embedded machine intelligences of our products but then through the layers of big data analytics payment and things like that HP Labs is working pretty heavily on machine intelligence machine learning kinds of algorithms in the enterprise group today we're working on that from a perspective of acceleration and being able to optimize and accelerate things like our spark and some of those capabilities so that you can analyze data much quicker and much faster from an IOT perspective we see a awful lot of opportunity there and i have t space one of the problems that you have is that you can almost think about it is a is a couple tier architecture writing some applications allow you to instrument via sensors a particular set of things aggregate that and then send to the cloud right for pantaloons I can send it back but a lot of applications require much faster response time so one of the things we're announcing actually during the show is a new product line called edge line that brings some of processing and the analytics right to the edge so I've got a set of things that I'm monitoring I aggregate those things instead of sending it directly to the track cloud I actually do the processing and the analytics at the edge and then go back to the thing and that's a time / wishing so what we will see the edge like Pride line is based on our moonshot capabilities and so moonshot gaze is the platform that quite frankly is best positioned for some of that machine learning in elegant algorithm kinds of things because I can use GPUs I confused fpga types of environments and I can do programming of that processing machine to do sense of respondent ripple area so this is an area of great interest to HP as it is to the industry we have an HP labs or in leadership capabilities at both machine intelligence and next generation when I come neuromorphic chips right just a fancy word that just means more kind of chips that our model synapses and things like that that's based on the memristor technologies that we have and we think that there's a lot of opportunity but anything else that has to ensure testing for standardized and those types of things we offer Haven as a service right now in fact what's interesting about that is you expose all those empty eyes as a service all those web services and so it's fabulous autonomy is great example of that because i can go basically consume a set of autonomy web services to do all types of analytics on human machine readable data okay that's very very helpful very very interesting and we're really on that path of opening up all of our api's can be as a service if you live and able to integrate into the broader environment what is the example of value in the nation what you get with open one view maybe is that ability meant to integrate into your service delivery or into your skin TV or into other parts of the day Center in a much easier way fastest very very important piece to be able to think of analytics is a service and other things help the kind of value problem synergy over other data center orchestration solutions you know like a mazal sphere or open source solutions that are a little bit more agnostic versus synergy which it was seems to be HP focus from a hardware perspective and tied into open source tools like of its yet what's the value problem if I'm not if I'm not primary to HP shop well that's good question if the embodiment of composability in a synergy platform is about an integrated confers system right that allows you to burn quickly discover and bring up automate orchestrate infrastructure relationship applications you mentioned this is here mestres bog and it's great stuff use messes here with dr.

with synergy wonderful combination so what's energy is is a next-generation intelligent infrastructure that has this set of opening the eyes that connect to then the rest of the to live within the data center to allow even more levels of auditions mesosphere doesn't bring up very metal art and doesn't provision no essence in a instant right it doesn't autodiscover all that hot where it doesn't have template-based profiles that line to secure it by simply update your firmware on those kinds of things so when people get a little bit confused about is the fact that what we're building with things like synergy is the infrastructure laughing for that fast fluid automation that is problems now can you do it without synergy sure to take more time more effort you're going to have a couple more pulls together and you're going to have to do that in a number of different steps with synergy okay I bring it up from the time that I install the datacenter the time that I start to provision all of those bare metal machines I do that a very fast fluid unified way first time and then after first time throughout its life cycle okay in a complementary way the bestest fears to the shaft puppet to the server to the Helion and to the other pool the basin and that's the focus of synergy it's in that smart intelligent ruled infrastructure and complicated the compliments a higher level truly in applications so from a hardware perspective department of brownfield environment you know I have you know HP foundation from computer effective but I've gone to another vendor storage so why would you want to do good questions in their questions guess what but so we're the I might you know the city is still offer a value for customers who don't accept that pole HP stand yeah yeah absolutely i think that you know what we're really talking about is we live in an energy news work most customers that we talked to they want dual sources of vendors day they're looking for you know some level of safety of HP tipped over then you know what was their second done second choice in those kinds of things so we have to be open and we have to be practical with regard to tools like sitting so synergy integrates very well into an existing data center environment and there's a couple of different scenarios scenario number one is an HP shop that's using you know maybe track based systems maybe blade systems and now that you want to move into the synergy drive but I can manage all that infrastructure we manage all that immature and they can integrate into some of the higher level to the top stereo number two is then down and cats and dogs and all kinds of different things in that scenario you know what what the IT guys are going to basically realize pretty soon is that Wow shares a lot faster a lot easier with synergy there it is with my doll or by Lenovo stuff right and then I think that over time as we move forward with composability and is open API concepts will embrace more energy an auntie within the data center so I'll give you an example a couple of years ago we decided that we were going to open up our our infrastructure and make our infrastructure program with rest-based interfaces wax a rest-based programmability and by awesome other aspects and certainty and he said well that's really cool and that's very important for the journey be wrong with one new possibility but we said you know this needs to be standardized within energy so we then got together with our great friends del Microsoft and towel and others and said let's go ahead and standardize on set of breasts based api's for server redfish yeah red rock is something else jeez it's left specification which is that we took that in the game tita but we started them we originated them and says thank of course of course they did that's a that's an interesting that he said she says with Danzo I was there so i can i can attest it when we start writing very fun it doesn't seem that I would say it differently I've been saying that I used to work in basically the president of standards organization so what I would say is that standards take time to ensure they take time to get adopted right so the progression of redfish came from internal project that we call rich infrastructure services that then moved into rest redfish 1 point 0 and then got submitted to the dmt they got ratified by the MTA and is now in its adoption fees right and so it's going to take a while to adopt because people are going to throw away the old ways right as they embrace these new technologies everybody the present what no I hey Stan children yeah we've actually discussed well I don't want to go in too much of a topic into a standards body you know going to be this offline if you book but let me just say just a couple of things about about standards standards in general are very important to the dictionary computer and in the next generation of veteran structure applications and apply but it's very unpredictable when and how a particular scheme is going to be enough okay and let me give you an example of that the standards body that that I was president of was the open great form you were in the early stages in distributed computing and trying to standardize and some of the web services associated with did and those web services were wsf web services which were fairly complex and based on staff and some things like this fairly heavy weight and align Oasis yoga good for them and a few other of the organization are all working together on these web services but what happened the internet basically decided that all that stuff is way too complicated and they decided to do what to do to move to a de facto standard called rest that's because the center it's not a standard it's not the standards bodies that are it's the fact that the different vendors of the industries have competing interests and intellectual property associated with the protocols the interface is not example of that something processes as well as on top visible so glad to say OpenStack would be realized you would be amazed we're not the meta manager there the meta manager would be sort of the Underland well it becomes let's take a real example of both of them so you know pin' stack I have this concept of flavors right and flavors allow me to basically say okay I work for a specific virtual machine on a specific boxing and provision if you're using synergy with its open api's its ability to then put a bloody driver into that flavor recipe then I can configure the entire compute storage fabric bare metal but the OS on there and make it ready for Nevada conjugator lon that's right to exercise your right track I'm not sure about the end of the heart i think i read i read i read before big data centers we set up equipment all the time I have taken synergy through its paces I guarantee you that synergy is a big big big improvement and a big ass identity we're doing within our own data centers today and when you say it's just a hardware orchestrator very useful addition losing very same problems I would say those are all those are clip clips above beyond those are all from with issues but i would say it sits alongside of mr.

s. Neymar jerk pretty good estimate that's some tennis shoe Jesus Frank a good idea it seems pretty good right and you know people want to be like life you know they can look up to my current position great basketball player device detection people look up to what Amazon and what Obama and what some of these people are able to do in scale with you know five applications 10 applications hundreds and hundreds of thousands of crisis in the life and look up to that as a model for both agility for moving forward with regard to my tree service oriented applications that can deliver by mistake that's what people say is I want that to our major accounts or medium accounts or small accounts for life that's why they're embracing cloud virtualization containerization is because they aspire to have very low cost very standard IT capabilities that are highly automated in this game next look we're all talking what we're doing with the post mill team is trying to bring that to a broader spectrum people just two people that have patience we're trying to bring that people who have 50 100 1000 5000 applications who have our admission critical applications that absolutely positively can't go down that require different types of capabilities and one you mentioned the notion of algorithm right paypal looked at our car portfolio is it I got to have booty shot I am moon shot because i can put security algorithms into that moonshot platform to be able to secure my payment systems that was a very specialized require now does paypal geez moonshot for everything yes but there are a variety of requirements they're very workload and customer specific and what we do is a HP g vendor if we provide people a spectrum of capabilities so if you're a large second tier service provider with five applications our polilight or plot line are perfect for you and go off example yes their edge that goes up to but pattern the pattern there's a pattern here and one view unified API right I can calculate that and what we call internally a1g drive that one of you driver can be inserted into an OpenStack flavor chef recipe and dessert flavor that's dessert can call that dryer and compose physical infrastructure lay down the operating system the life right okay and that pattern that will be instantiated with dessert and with the variety of different dubstep software configuration of standing we already did where in the city question as long side don't content is facing like my decision yeah so the answer to the question is this gentleman back here in just a minute Paul Miller you know it works at our CD I organization it does our own diverse and synergy is going to make it a wonderful convert system integrated system or whatever you call that it is a platform that you know our Organa guys and our new guys and and a whole bunch of folks are dying in a hold of that has a diverse system have I experienced with the convert systems in our in our internal data centers when we did a separation I ordered one of our convert sisters OCS 704 cabos data center because we were doing separation of our IP and dressing between 80 and HPE a big freaking job quite frankly the additional capacity I heard when his boxes ordered it got it in about 23 weeks to a month and drop it off for Collins my team had never seen one of those boxes before had it installed in less than a certain virtual machines by developers that night and had an enemy sees throw rocks man selection anything about Parkinson's oh yeah good that we're actually hopefully during a lot of secretory merges the first with our cloud system with the double one private cloud black so we actually one of the biggest players are dead with cypionate awesome we're going to take synergy a grenade customer I like to publix by then I put us on down the template that gives it the personality or single fatality so we're going to have