Google Cloud User Tips

hi I'm James McGuire and our topic today is the Google Cloud how to get the most in the Google Cloud platform to talk about that I'm joined by Paul Nash senior product manager Google Cloud platform hi Paul how are you doing today great good morning how are you good good and I know you were up in uh Seattle this morning yeah I am uh we have quite a few people up here working on cloud it's kind of become known as the Cloud City abolutely cloudy out today too it is literally cloudy in addition to like you know business-wise cloudy it's not just a pun yes speaking of which before I give you my first official question I want to mention a tweet that you gave a while back this is actually a cloud joke and I I love this there's not there's not that many Cloud jokes out there the Tweet goes like this quote went away myself got a low battery error I have a scale ability problem yeah now you you definitely have to be a cloud geek to get that I think I I think so yeah it's it's the kind of thing you know um some people have brilliant ideas to revolutionize the world when they're in the shower and uh Wi-Fi battery powered scale and it doesn't work and I start thinking about um what my sres would say so excellent I'm gonna nominate that it's the best cloud joke in 2017 we'll see see what else comes in I'll keep an eye on the retweets so you know thinking about a business that wants to get on board with the cloud or or more likely wants to expand their Cloud deployment yeah you know there they're looking at you know what might be thought of as the big three you know Microsoft isure Amazon web services Google Cloud platform what what would you say to that business you know why should they choose Google Cloud to either get on board or expand as opposed to one of the other large infrastructure as a service vendors what's the advantage there yeah so I think the primary thing that I would say is you know Google we feel like we've been a cloud business for pretty much our entire existence and so the thing that's changed recently is we've entered the business of providing that excellent technology that we've been developing to power internet search and YouTube and some of these other giant scale Services uh we've been working on providing that power and flexibility to our customers so uh for us it's always been about Innovation and that started way before we had Cloud products but then continued with things like app engine uh which is one of the very early Cloud products out there um and also other things that you might not think about as originally cloud or or even originally Google like Hadoop which was really derived from map reduced which was a Google technology right there are others out there as well like Google file system so uh now that we have um all of that technology um being brought and made available to Consumers um we believe that really provides some of the best technology that's available in Cloud uh to deploy your workloads on so uh we're working hard every day to continue innovating there and bring the most performant uh price efficient and uh competitive interesting product to Market I think you know price is a key issue and I you know a lot of Industry observers point to this you know so-called race to the bottom pricewise yeah some some of the larger vendors are are cutting prices aggressively in these still fairly new days of the cloud and an effort to gain market share um what what would you say to a business in terms of the either the price advantage of Google or I I'd make this a complicated two-part question if I can either a what what what is the price advantage to Google or or B if a business is going to get on board with Google what what is the what what are The Insider tips about pricing like where where's the 59 cent off coupon so to speak or right what would you might say to them in terms of what how do you shave a few pennies if you're gonna get on board with gcp sure so I think the primary thing is um we've approached pricing from a sense of trying to keep it uh Innovative and also flexible and easy to use and simple right so um I actually gave a talk about this at at our conference Cloud next in in March um they really sort of underscored each of the the pieces that that go into that and so that starts from having a permanent billing option so we're not locking you in for an hour uh we automatically give you sustained use discounts so if you use a VM for a significant fraction of the month starting at about 50% of the month and going all the way up to 100% we automatically give you a discount back so uh that's more scalable and flexible you're not locked into saying in advance that you want to pay or that you're going to use it for the whole month if you don't know um now more recently we've introduced another product called committed use discounts and we've been talking about this a lot uh with customers there are some customers that know that they want to use a product for the whole month or for the whole year and so they want to be able to tell us in advance and and have us basically give that uh sustained use discount over that even longer time period so um on top of that uh we offer Uh custom machine types which is a flexible way of of dialing in the size of the machine so that you're not paying for more than you need uh and and a number of other Innovative features like that so um one of the things that we're doing is trying to make it so that they don't have to go searching for that 59 Cent coupon they just automatically get a great price right so I guess one other thing I'd mention in that respect is we have a feature called right sizing um and this has gotten some some interesting commentary like in in the Twitter universe um we have a feature that will automatically tell the customer you're spending too much money click this button and you'll save money s seriously so how how does how does someone access the right sizing button so it's automatically there in the developer console um what we'll do is from the outside without looking at at actually what's running inside your VM we can tell if your processor is effectively idle or if you're not using all of the memory that you asked for in the VM and so we can say hey you never used more than 30% of the memory on this machine you should scale that machine down to a smaller size click here to do so so so in theory actually Google would have less Revenue at that point but the customer would be happy happier is what you see is going on yeah that's basically I mean we're we're happy to help the customer uh spend less money on that particular uh resource because we think they're getting a better value that way and we know that um you know they'll they'll be happy to spend the money doing something else that's important to their business we we don't feel like we need to overcharge um because it's not just a vendor relationship to us like where you know we have a fixed product we're going to sell it to them they're going to buy it and go away this is kind of we see this as a partnership so uh we uh feel and a lot of our customers tell us um that this thing of give us all your computers and we'll run them for you and just trust us it'll be fine that's not really a vendor relationship anymore we hold their business in our hands and we take that responsibility very seriously and part of that is uh not overcharging them when when we think they're not looking so do you are a lot of customers taking advantage of the right sizing option is this still still a new idea yeah we see a lot of excitement about it it's a feature that I believe is still in beta but um we've seen a lot of interest in it and we have some great ideas about where to take it forward in the future um it also leverages custom machine types so they work together right we will tell you what the uh size of what the right custom machine size is um on the custom machine type side uh that feature's been out for a little bit over a year and we actually see about 20% of our total usage uh is from customers using custom machine types so uh that's been pretty amazing and and um pretty clearly answers the question of like is that a weird hard feature to understand how to use no we're trying to make it as simple as we can what about um the idea of Google being you know the open Cloud I sometimes hear that I mean it's that either AA is that true does Google want to be known as the open Cloud and does that mean focusing on open source because I know some of the some of Open Source developers Grumble about not all vendors give back equally and some companies give back a lot more of the open source community so is Google focused on on open source or or is it being open in a in a different interoperability in that sense that's a great question so uh definitely open in the open source way but I think it goes beyond that of building a community of Technologies and uh best practices for um how the industry works so on the open source side I would mention a couple things first of them uh is kubernetes uh which has really brought a lot of Google's uh previously internal knowledge and expertise about how to manage containers out into the public and invited other contributors in and and we see that there are other companies out there contributing at a level that is uh you know as serious as we are to that effort and it's it's become quite a strong Community efforts and there's foundations and it's really great to see that energy the other one I'd mentioned is tensorflow uh which is really sort of becoming the de facto standard uh as a platform for machine learning uh technology and we see it running in all clouds um so a conference open source just for backr it's an open source machine learning library correct yeah that was Google technology that was uh invented and then released open source about a year ago um there's another one also uh we have a product called data flow uh which has a programming model that we open source called a it's now called Apache beam um so this is sort of three examples and that is definitely a part of our key strategy um the thing I would say that goes beyond that where we've taken some efforts to engage the community is uh a a tool kit that we released a while ago called perf kit which is a open- source uh collaborative way of measuring performance of uh VMS in the cloud so um we're talking about price but uh at Google uh we like to talk more about the this combination of price and performance so really it's it's not just about what you're paying for a virtual machine that you know you don't not made me sure how much performance you're getting out of it because what you really want to do at the end of the day is um pay as little as you can to accomplish the work that you need to accomplish so really that combination of price and performance is is what's important so uh perf kit benchmarker is a set of tools that we developed in open source and we have uh collaboration from a number of different companies all contributing different benchmarks into that and then that can be run on any cloud or on any on premise workload and all the data is publicly available so you can see how those benchmarks are running across clouds um and it's not just us saying that cloud over there runs this way we run at this speed or we have the best you can actually go measure for yourself and and even monitor it if you want it it literally is going to compare performance among different clouds even in not non-google clouds with Google clouds that's right yeah so the benchmarks can be run on any Cloud um and then you can you can compare the performance of those results so um that's just one way and I I think in general you know what we see is that um most if not all customers when you survey them they express an interest that or or an intent to use more than one Cloud so I think the the future reality uh for the next you know number of years is that there customers are going to be using multiple clouds so that ability to um choose the right Cloud for the right workloads and move things around if they need to is definitely important so I think that that interop and and openness is is an interesting aspect that um so we don't just mean we want to put this Source package out there and then have everybody use it it's it's much bigger than that right interoperability I do hear that from a lot of businesses that very few businesses want to have a a sole relationship with one Cloud vendor I mean I think I read this survey recently where it's like it's some very large percentage of businesses have a multicloud setup going on yeah our research is that it's at least 75% of them right so I mean let's let's let's talk about the future I mean if if we if if we were suddenly sitting here in the year as you know 2020 which is actually not that long away scarily enough um you know what what are we going to be talking about in terms of of cloud where where is cloud going and I guess does Google clearly seems to see the multicloud world is going to be the dominant world but what do you see happening in the cloud in the next few years several years yeah so um you know it's it's a it's a broad spectrum of uh needs that we see from customers today so uh there are a lot of customers who are still getting started really with Cloud um almost everybody over 90% from what we can tell uh have something in the cloud now um but maybe it's an experiment or a digital marketing campaign or a mobile app or something and they're kind of still figuring out where they're going to go with their their core infrastructure and and line of business services so uh in the next few years I think we'll see that the vast majority of people have moved a lot of their work to cloud and that that sort of continues to accelerate and and that's either um a a lift and shift type operation or it's a re-engineer and a transformation um a lot of the reasons behind that is just it's pretty clear now that there are cost savings to be had in Cloud no matter what your use case is you can probably do it more affordably in cloud and you certainly get more flexibility so um we look at that and it's it's really um we see that the cloud model of engineering and Innovation and flexibility has led to us being able to build some incredible Technologies as Google and we think that that's only accelerating when you look at things like machine learning and AI um Internet of Things uh devices those Technologies really benefit from having the scale and flexibility of cloud and being able to use what you need when you need it so there's kind of a revolution going on um and so another technology would be like serverless it's a brand new way of programming that right doesn't really exist on Prem and as more and more companies move to the cloud and look towards how they want to build a solution uh 's more leverage in fact they can build Solutions faster and uh there's more flexibility there for them to do that so um it's interesting the serverless is worth taking a second on because it seems like it's the ultimate and headache free cloud computing could you talk a little bit about say how Google is using server lless or what that means for gcp in particular yeah and I think we have probably some more to say about that in other forums as well so I think we're we're kind of running out of time but uh Quick Way um so we have Google Cloud functions um we also have uh app engine technology which we continue to invest and innovate in that was one of our earliest Cloud products and both of these allow you to uh create Innovative applications very quickly that don't require you to get involved in the behind the scenes operation and provisioning of servers and and those bits that might not really be uh what you're after and so we see a lot of interest and excitement in that space and and we're uh that's an example of a place where we're innovating rapidly so Paul I think you said it I mean any any closing comments you want to add before we wrap up today no just uh really thank you for the time it's been great to chat with you and I really appreciate it and uh hope to see everybody in uh Google Cloud fantastic thank you very much all right bye bye

This transcript was generated automatically from the video's captions and may contain errors.