Office Hours is back for the New Year, and we’re kicking it off with Rackspace CTO John Engates reviewing his 2017 Cloud Predictions. This is an annual tradition on our show - and this year we’ve got some great predictions to cover.
What a great way to start 2017. Rackspace CTO John Engates joined us in the studio to cover his 2017 Cloud Predictions in greater detail. John’s a frequent guest, and a long-time Racker. His unique view of the industry always yields great insights, and this week was no exception.
We took the opportunity to look at each of his predictions in greater detail. I was very interested in the way that all of the predictions work together and build on each other. These predictions aren’t just guesses, they’re observations based on a understanding of the tech ecosystem at a very high level.
Alan Bush: Hey everybody, we’re live at Rackspace, the number 1 managed cloud company. My name is Alan Bush and we’re here with our cohost, Drew Cox.
Drew Cox: Hello, hello.
AB: This is Office Hours, it’s our weekly live show about the cloud and cloud technology and today we are doing one of my favorite shows. We’ve done this for the past two or three years now. This is an opportunity to bring in our CTO from Rackspace, John Engates.
John Engates: Hey everybody. Look at that I hit the microphone, right off the bat.
AB: No worries.
DC: We should have been talking about it before we went live.
AB: I know.
JE: Break the ice, break the ice.
AB: Yeah, there we go, break the mic. I know, also, we’ve had you on the last two or three years. And you’re one of our frequent guests and we really appreciate that.
JE: Thank you.
AB: And especially, this time of year, because you write your annual cloud predictions.
JE: We do, every year we sorta look forward and try to figure out what’s gonna happen in the next year and we put it in a blog post and we share it with the world.
AB: Yeah, it’s really one of my favorite things. For the past few years, you’ve been pretty spot on, you know and I’ve looked back, we kinda look back at them, and taken a look at them and they look pretty good. But you know this year you got a bunch of new ones coming up and we’re gonna talk about them today.
JE: Sure, absolutely. Excited to get started.
AB: Yeah, I’m very excited about this. It’s again one of my favorite things. So there’s a lot of predictions. There’s a lot of stuff to talk about, so I wanna jump right into it.
AB: Let’s start off by going through 2016, see if there’s anything that really stuck out to you from 2016 before we jump into the ones for this year.
JE: Well you put me on the spot ‘cause I’m thinking about the next year already.
DC: It’s good to be looking forward.
JE: No, I think some of the things in 2016 that I was excited about, things like VR and augmented reality, some of the features of self-driving cars that started to emerge with Tesla and others. You know, the cloud, we talk a lot about clouds. So this is primarily what we do here at Rackspace but you know cloud competition, we started to see that continue to grow. Not so much focused on the price of cloud as much as the features, the capabilities. I think that was part of my prediction and that pretty much came true. You sorta stopped worrying who was a penny cheaper but you started thinking about what cloud or what technology set met the needs of my application or my business. You know I think security has been a topic that we’ve talked about for a long time, it continues to be one. I got a few new thoughts on it, but security got a lot rougher for us all. You know across the industry, personally and professionally. So we’ll dig into that as well.
AB: Yeah, well, you can definitely go back and check those out. They’re still on the Rackspace blog at blog.rackspace.com. Just search 2016 predictions and you’ll get them.
DC: You can actually find the link at the bottom of the 2017.
DC: Or there’s even a hyperlink within the blog itself.
AB: Yeah so.
DC: You find the one we’re gonna be covering today, you’ll be able to use that as a reference to jump to the rest of them.
AB: Absolutely. We’ll put a little link here on YouTube, make it out there. Let’s go ahead and jump right into the 2017 predictions. And they kinda followed your very first prediction, really followed on the heels of one that you made in 2016, which was a multi-cloud world.
JE: Yeah, and I think that’s continuing to be true. Actually, in terms of what I’m hearing from customers and people in the industry, it’s accelerating. The multi-cloud, when I say that, what I really mean is that consumers, businesses or developers, whoever’s consuming cloud is more and more likely to be consuming more than one cloud.
JE: We saw a survey last year, that said the average company was consuming 6 clouds average, average of 6. Which is crazy.
JE: But that cuts across public cloud and SAS and maybe platform as a service type clouds, possibly even private cloud. But that seems to be accelerating because of the nature of the competition between the big cloud providers. You know Google, Amazon, Microsoft, others in the industry, Rackspace with the private cloud offering and our own private cloud, Openstack really driving on the private cloud side. All of those are bringing so many options to the table that companies now have a variety of capabilities they can choose from. It’s like a palette. It’s an artist choosing a cloud capability from this palette of service offerings and weaving them together into a comprehensive application. And not being beholden to any one particular cloud or technology set. I think developers are curious people. I think engineers and architects love to try new things out. Often times, that’s how multiple clouds make their way into the organizations. Sometimes maybe it’s a business unit or a marketing team or somebody going off on their own and doing something innovative and you know by accident a company ends up with two, or three, or four cloud relationships.
JE: And you know that’s just a reality today, that we’re gonna have to learn to live with. But the challenge, the reason I put it in the blog post for the year, is it’s creating new challenges for companies as well. Being the complexity of managing all of that. Having the experience and expertise on staff on hand, and keeping those people around because they’ve all got really strong job options out there in the market.
AB: They do.
DC: Well there’s a governance component to that too.
DC: You’ve got the complexity of I didn’t even know we had that cloud and now we have to understand its implications to our business and either integrate it or deprecate it so that we can move forward making the most sense, but you’ve also got you mentioned this big palette of well we need to be in this specific geo, what offerings are there that can get us as close as possible to that geo.
DC: Or what technology’s out there and so to take the most value from those various relationships it takes understanding and having a clear view of where the value can come from is a really big piece of that conversation. Being able to govern it all well so that it fits within your business requirements.
DC: It’s gonna be one of the headaches that comes along with that multi-cloud world that I think solutions are gonna start cropping up more and more.
JE: I mean it’s great to have all these choices but it’s also…but what do they call it? The double edged sword or you know it’s the problem as well. I mean, you’ve got the benefits but you got all the negative aspects that go with it. Of keeping up with all of these cloud providers and what features they’re launching and which services are mature, which ones are ready for prime time. And if you build up a bench of experts on your team do you have somebody who knows these things in depth or are they just sorta tinkering around with them enough to be basic familiar level expert. And I think ultimately as you build your business on top of these services, you want really strong expertise. You want someone who’s done it before.
JE: And that’s really where we’re putting our focus for 2017. Is helping companies make their way into the cloud and make their way into multiple clouds and take advantage of these capabilities without the risk or the cost or the pain that goes along with it. We want to be that safety net behind the scenes and really help companies adopt these technologies as fast as possible.
DC: Yeah, I like that we’re able to facilitate getting all the right people at the table. All the right expertise at the table because you may have a couple of experts within the business that could do it but are they on the same page and do you have those experts with the depth of expertise on the things that they’re not currently using. So if we can bring our expertise in and have that collaborative discussion between developer groups and all the other people who are important to that decision I really can amplify the value that a customer can get-
DC: From all that stuff.
AB: Having one team of people that understands the entire business, you know. If you’re talking to five different vendors and you have five different points of contact, you know, they’re not gonna get the same picture as having one point of contact.
JE: The real challenge that I see a lot of folks struggling with is keeping up with the pace the innovation. I mean, we’re talking all about futuristic things here but every one of these companies have a road map that is aggressive. Everyone’s building as fast as they can. They’re building out new services and capabilities. They’re doing it across, like you said, lots of geographies, lots of different form factors of technology to be consumed. And it’s just hard to keep up with that pace if you’re not focused on it 100% everyday walking into work thinking what’s new what do I have to think about, your app can sort of stagnate. One of the things that we’ve seen some companies struggle with is they get migrated to a cloud. Pick your favorite one. They get migrated by a professional services organization. And then it becomes this sort of thing stuck in time. It’s like an artifact of two or three years ago, right? And then you see that they built it in a certain way because at the time that was the right way to do it, that was the right technology set or stack to take advantage of, but the world moves on and now your not really taking advantage of the new stuff and you’re stuck on an older sorta model or architecture. And that’s also what we’re trying to solve for companies. Is helping them stay current by bringing them along with us on this cloud journey and across, again, these multiple clouds. So that’s, you know, why I put it on the list. It’s something we’re spending a lot of time on. I know it’s a challenge for CIOs. They’re telling me this every day. I know the cloud providers are also thinking a lot about this, how can they enable this, but ultimately you know this is where the rubber meets the road in terms of cloud adoption.
DC: Yeah that threat for disruption is always kinda looming in the back of the mind of someone who’s established entity in an ecosystem. So when you have this two or three or four year old approach you are increasingly susceptible to that-
JE: Yeah that’s right.
AB: Well one thing, you know, with the multi-cloud that you mentioned that there are additional challenges that are brought on top of that. I think one of those things is security.
AB: And that was ironically enough the next item on your list for prediction.
AB: So let’s talk a little about security.
JE: Security-wise is a big one for me as well. Number one , personally, I got hit by some cyber criminals, whatever you wanna call them. Somebody opened a credit card in my name last year. Somebody filed taxes for me, so that they could hopefully get my refund last year.
AB: Oh, okay.
JE: So a couple of things hit me personally. I started thinking a lot about security last year on a personal level, but as a business I think every CIO every business leader is thinking about security. Because it’s impacting them and their bottom line. When they have a breach, when they have any kind of event that’s security related, it erodes confidence on the part of their customers, even their own employees in some cases when things get breached and all their data, HR data maybe, gets leaked out on the web. Your own employees may lose a little trust. And it cost companies a tremendous amount of money. And this last year we saw some very large scale security problems. You know the DDoS attack that came near the end of the year that was targeting-
DC: Target DNS.
JE: DNS took down lots of companies, because the backbone of your internet presence is DNS.
AB: If you don’t have DNS you’re memorizing
JE: And so Dyn, the company that was targeted had a hard time during those couple of days because they were flooded with traffic. And I think, you know, my prediction was that this problem is gonna get worse before it gets better ultimately. I think from a personal level and a business level we’re gonna see security events rise, the threats are gonna get more sophisticated, the bad guys are always persistent and are gonna get more emboldened because they’re succeeding, right? I mean, one guy is a copy cat of another guy, sees that you can make some money going out and starting up a botnet and hire out your services to the highest bidder and all of a sudden you know, we’ve got you know a real problem on our hands with security. One of the researchers said that they believe last year was the year where they were testing security to see where the boundaries were of when DNS falls over, or when a DDoS attack has the ability to really seriously take down. So if they’re just testing in 2016, what does it look like for us in 2017.
AB: Well, a lot of these are just a lack of importance of security at the time that some of these devices and products and things were created. And I think that if nothing else, last year taught us that security is very important and your company needs to build with security in mind. And we look at these devices that are created without the ability to change the firmware, which means they’re hardcoded passwords are in there and they’re known to the world.
AB: And so, I really hope it’s been a good lesson to people going forward, these physical devices, these internet connected machines are gonna become more secure and that it will be top of mind for everybody.
JE: One of the things we attached to the blog this year was a list of some of my security, my personal recommendations. I think it’s attached. But one of them was don’t buy a gear from companies that are fly by night, Chinese, you know nobody knows the name of the company that actually makes it. Buy your connected devices from somebody who you feel like may have a future and will update these things and will keep them current. And if you’re not able to keep them current, get rid of them, take them off the network, disconnect them, because all you’re doing is you’re putting your own network at risk you’re putting internet at risk. It’s hard to really think about that on a individual level but every individual device adds up to this DDoS attack that we saw last year that really has a big implications for really the internet as a whole.
JE: So I do think security is critical. Back to the topic we just talked about, multi-cloud. Multi-cloud security I think is gonna be something that you’re gonna start to hear a lot about. Because that is the new challenge when you think about applying security across multiple clouds. How do you do it consistently? How do you have a view of what’s going on in this different places. Everyone of these clouds on the market approaches security a little bit differently. Some of them, you know, have their own way of doing firewalls, or their own way of doing network architecture. And now the challenge on a company is how do I apply security consistently and effectively in multiple locations. How do I keep track of what’s going on where and do I have the expertise to do that as well.
DC: Yeah, some of the value of having multiple platforms is that you’re subdividing your network out a little bit. Your solution, so that if one piece gets compromised, it’s not the rest of it.
DC: Which is one of those concepts that’s really gained momentum over the last couple of years of focus on security. But it’s an arms race, there’s gonna always be bad actors looking to find a new exploit. So there’s always gonna be people discovering new exploits that are being used and trying to find ways to protect against them and so…
JE: Security is a full-time job, there’s no way to do security well, if it’s just a piece of your job. If it’s something you spend an hour on a day and only do it once or twice a week when you think about it, that’s a problem. Your security at this stage, due to the persistence of the bad guys and the level of sophistication, security is now something that’s risen to a level where you need to be thinking about it kinda on a full-time basis and have people watching your security on a full-time basis as well.
AB: I think the next item that you had up on the agenda was serverless architecture.
AB: This is something we’ve started talking about a little bit this last year actually in 2016 and I think it’s gonna be a really interesting way for people to build services and build little small applications that don’t need that server.
JE: Yeah, so serverless is interesting. I mean, the name is a little decieving because technically there are servers behind the scenes of every cloud service that you might consume. But the idea behind serverless is that you shouldn’t necessarily have to be a system administrator of a Linux machine to get a service up and running. If you just need something to check in every once in a while or a collect a piece of data every once in a while or you know, fire off a tweet in response to some sort of event, why do you need all this heavy infrastructure with load balancing and why do you need to put all that together just to get that one little bit of functionality. So what cloud providers have started to do, AWS is by far the leader here, because they’ve jumped out ahead, but everybody’s thinking about how to make these services available. But they’ve basically launched a service that they call Lamda, which allows companies to build little micro-applications on top and have these things spin up only when you need them. And that you don’t pay for them until you use them. They are very low resource requirements. They don’t cost a lot and in a way it really takes a lot of burden off of the administrator or the developer of the application. Some of our own teams here behind the scenes have really embraced Lamda and serverless architecture. Some of them say they never wanna go back to the old way of doing things. You know that’s crazy that it’s gone that far that fast but this sort of supplants some of the container conversations we might have had, last year. Containers was on my 2016, and we know that really blew up in a big way with Docker and Kubernetes and others. We’ll see that continue, but this just yet another tool in the arsenal for a developer and architect to build a really capable application with much lower overhead in terms of their time and their energy.
DC: Yeah, if you can achieve the same thing with few requirements that’s always gonna be valuable. It’s always gonna be more efficient and so I know that if I could get the same outcome and not have the maintenance and upkeep and management and patching and all that stuff to do with a piece of hardware or just a virtual machine I would be happy to offload that to something that could run when I need it, I’m billed only when it’s used and I get so granular with what I am doing that a lot of the day to day work that the operational groups have to do goes away. In those cases, where they can focus on really optimizing the other components.
JE: And you know I think this is the year to start paying attention to it. Last year, you know, we saw a companies get their feet wet with it. Amazon at their Reinvent Conference in late November put a big focus on it with Lamda and other technologies that sort of complement it. They’re even, you know, sorta pushing some of these capabilities out into physical devices that might reside on a customer premises. There might be something you wanna build. For example, I think they have, you know, a way to get a little bit of computing power in a product they call Snowball, which is their on premises storage capability. And serverless certainly plays a role in all of that, as well, because now you can push computing into places that might not otherwise hosted a full VM or full stack.
AB: Those are one of the things they mentioned at Reinvent was that you know now they can put those into devices.
JE: Devices, yeah, embed some of this technology. Yeah, we’re gonna see a lot of crossover when you think about this whole list. Everything here, it all touches one another. It’s hard to actually tease this apart because a lot of what we’re talking about is foundational level stuff that other things get built on top of or integrated into.
AB: If we slice this up as a little playlist, you can jump in anywhere you want.
JE: That’s right, rewind, fast-forward, whatever you want.
AB: And it’ll all be there. It’s really interesting the way your able to break these apart and not use servers anymore. Like you mentioned everything is all connected and one of the things we’ve seen a little bit is at least a slowdown if not the end of Moore’s Law.
JE: You know what Moore’s Law is? Everyone know what that is?
AB: It’s the doubling of compute… But let’s give the full…
JE: Yeah ultimately, you’re right, it’s the doubling of compute power or approximately every 18 months or two years. You know for the history of computing, we’ve seen that sort of trajectory, where you can draw a line from where the first 8086 processor all the way to today’s, you know, fastest server processors and what you saw is this sorta steady increase in the number of transistors on a die and the amount of processing power available on that machine. But that has slowed down, the curve has sort of flattened out. It’s not quite the same slope as it once had. You’re seeing new computers come out that have a lot of whiz bang stuff but the processing power isn’t that much more than the previous generation. I’m watching what’s coming out of CES and other events and your just not seeing the level of increase that we once did. And what that means is that the idea of Moore’s Law which isn’t really a law of physics. It’s just a prediction that Gordon Moore made years and years ago. It’s sorta flattening out and coming to an end. And what that means is that architects, designers, application builders are looking for ways to get performance back and lower the cost of building their application and running their application. How do we do that? How do we get creative about where we find new capabilities. And one of the things that we’ve seen lately, is GPUs. GPUs used to be graphics processors and now they’re now becoming more general purpose. They’re call them GPGPUs. They’re general purpose graphics processor units. And they’re basically being used in ways to enable very sophisticated computing and it’s bringing back some of the performance that we were losing on a general purpose processor. And that’s what I meant by the end of Moore’s Law. It’s not that we’re gonna fundamentally stop getting more performance or lower cost performance. It’s just that we’re gonna have to get creative about the ways that we do it. And you’re seeing the way the cloud providers really help us with that.
DC: Well having quickly accessible resources has traditionally over the last I dunno, five, six, seven years been really easy for a customer who’s saying I just need more performance I’m gonna throw more horsepower at it. It’s been really affordable to go about it with that bandaid approach of yeah, we’ll just give it more horsepower and it’ll be fine. And with those changing landscapes and with it being more and more difficult to just upgrade the processor or replace the machine with something, the lastest and greatest, you’re gonna have to look back and okay how do we more efficiently build this and how have we approached the code and how can we be better on that end? And so you can give us the solution more horsepower or you can ask it to do less. Those are the only two levers from a really high level that you have to get more our of your end result and so, I think the pressure is gonna start moving more towards developers to do a better and better job of efficiently approaching what they are building.
JE: Well I think you’re absolutely right. I think that’s been a part of the realization that developers come to when they start to move to the cloud is that they can no longer just get a bigger machine or get a bigger CPU. It used to be that was the way you scaled your database. You remember the early days of MySQL or SQL server, if you needed more performance just get a bigger box. You know, put more CPUs in it, get some more RAM, maybe some faster disks and that was the way to go. But the cloud has sort of forced us to start thinking horizontally about our scale. And that’s ultimately also another trick to get around this Moore’s Law problem. Is just figure out how to scale across machines rather than just expect a bigger one to come next year.
AB: Well in explaining that to people was the reason I was brought here to Rackspace, was to help you..
JE: This is an old…
AB: This is old hat for me, well four and a half years old for me but …
JE: But it still rings true.
AB: It absolutely does and it’s one of the things. You know, as long as we can… We’re finding new and different ways. So you’re mentioning that yeah Moore’s Law is changing, but yeah okay let’s distribute the load across multiple servers. Let’s also… Maybe some of these processors
JE: Use GPUs
AB: Don’t need to be on there so we can use serverless architecture and use the processes that way. Or we can use GPUs to offload some of those.
JE: That’s right. So there’s a lot of special purpose computing that’s coming into the picture now. And that means you now have more options more in your palette to build from but again it increases complexity. There’s always a tradeoff here. It’s not so easy as it once was. I remember that the stack, you know, lamp stack in the old days, it was easy to sorta keep it all in your head. And now…
AB: Or all on one server.
JE: Yeah, all in one machine, that’s how people build a lot of apps the first go around. But that’s a lot different today. The sophistication of the layers of the stack, every single one of them is a lot more complex. And then you throw in all these other things. You really got a matrix of things you have to track.
DC: And across muliple stacks.
JE: That’s right.
DC: I think that this end of Moore’s Law concept is really being seen on an open sourcing of hardware side of things. You’re seeing how people are looking at what they’ve been spending and using brand name stuff and saying you know if we’re gonna continue to get a cost savings we’re gonna have to start doing this a different way. And your seeing big groups of large companies get together and say how can we open source this technology, how can we build a unique and purpose built and continually push down some of our costs to get exactly what we need. And so there’s that additional fall out in the cloud world.
JE: Facebook did that a couple of years ago. Facebook did that with Open Compute. They reached out to me via Facebook by the way. They sent me a note.
AB: I would expect nothing less.
JE: Said do you want to get involved in this open source hardware thing. And that created Open Compute, which today we use. Some pretty cool technology and also IBM with Open Power, that’s another area where they’re innovating their processor architecture, opening it up, allowing people to drive innovation beyond just a product, it’s something you can sorta hack on and create something that’s really unique for your company.
AB: That’s great. Well then this kinda dove tails into one of the other things that I think is really exciting is machine learning. And that’s one of the things that you’re mentioning creative ways of cloud companies changing things up. We saw Amazon at Reinvent launch their machine learning platform, things like that. So there are more ways you can approach this idea and use machine learning to improve a lot of the process.
JE: Yeah machine learning is an exciting area. It’s an emerging technology set. It’s under the umbrella of artificial intelligence when you hear it, AI, and you hear machine learning they’re part of the same sort of academic research area. When you think back to where it all started. But what’s happening is that companies today are trying to take advantage of some of these machine learning capabilities to make decisions, to drive new innovation in their products, they’re trying to use the data that they’ve collected, you know, over many months or years. They’ve got tons of data. How do we make use of that data? How do we find sorta meaningful outcomes or pieces of wisdom out of that data? And computers are more and more likely going forward to be able to do that better than we are in some cases. Instead of having to write an algorithm to look for a very specific thing, with machine learning you can allow the computer to sort of learn as it goes. You give it a little bit of a clue, here and there, but ultimately a computer can start to recognize patterns and find things in some cases that it would be very difficult for a human to come up with. And I think that’s where you’re seeing a lot of innovation going on. If you look at CES this week, the show’s that’s going on there. They’re showing off self-driving cars or talking about you know relationships. I saw today Audi with Nvidia are working together to get you know self-driving cars on the road by 2020. And they’re just one of many companies that’s doing this. Tesla been using the same technology for a while. But machine learning is what allows the car to understand the world and see sorta this computer vision and machine learning that they are understanding, okay, that is something I need to avoid, that is something that’s normal, this is how I find my way down the road.
AB: Well there’s a great video a week or two ago, where the Tesla stops automatically because it noticed there was about to be a wreck. A couple of cars in front of it.
JE: That’s awesome.
AB: And that’s great. And that’s the thing that we as humans can do with fairly decent frequency provided that we’re always looking.
JE: Paying attention, right?
AB: Even if you put your phone in the trunk and have nobody else in the car to talk to and you got the radio off. Like, you can blink, you can sneeze, you can get distracted by a billboard or something like that, the equipment, the sensors on the car are looking where they’re supposed to.
JE: And they’re 360. THey’re not just looking forward, they’re looking on the side, in the back, in every which way. And they’re gonna tell you, they’re gonna tell the machine learning machines that are embedded in the trunk. This is why Nvidia such a cool company. They’re the ones that are really embedding the gear right into the car. Because you can’t do this in the cloud. You have to do this locally.
DC: You can’t deal with latency.
JE: You can’t deal with latency when you’re driving down the road. This is the one area where the cloud doesn’t actually solve the problem.
AB: Yeah but they’re also sending that information back to the cloud to reprocess.
JE: That’s right and to get better as a network. And Tesla talks about their fleet. Basically their fleet is every car, every Tesla that any owner drives. And they’re basically collecting data about the road the conditions, how these cars are performing in the field. And all of that is shaping the machine learning algorithms and training these things to be better over time.
AB: And that’s the thing that doesn’t have to be instantaneous no latency, that’s the thing that can be shipped back when it’s charging or when it’s connected to your home wifi whatever.
JE: The reason I put it on the predictions is because it’s now available to almost anyone. If you wanna try it out, you can do it in Google’s cloud, you could do it in Amazon’s cloud, Microsoft’s got some pretty capable machine learning technologies. IBM has their technology called Watson. That’s a lot of machine learning and AI built into that. And ultimately this is now in the hands of developers who want to take advantage of it. It is a bit of a leap from what we’re all used to because you got to learn some new technologies and sometimes you got to think like a mathematician or a statistician right? But ultimately it is available now. It’s starting to be, you know, really easy to access. In some cases as easy as writing a little piece of code and now you have the ability to recognize speech or recognize images. Image recognition is one of the most common forms of AI machine learning we see today.
AB: What’s interesting is that you mentioned Facebook earlier. So my buddy and I were talking the other day and he mentioned he had a problem with his internet connection at home and because that Facebook was degrading and not showing any of the images. But the metadata was.
AB: And it was saying this image might contain three people two of which are smiling.
JE: There is Chrome plugin that you can get that will expose exactly what tags have been associated with your Facebook images. And it will tell you, there’s two bald guys and a guy with a hat and a couple of microphones. And it’s almost that good. I mean beaches or ball parks or cars or kids or dogs all of that stuff is recognized. And it’s based again on machine learning where the algorithm has seen enough of these over time. It’s been trained to say that’s a dog, that’s not a dog. And from then on, it knows how to recognize an animal in the picture.
AB: Well Google Photos has been doing that for a few years. That’s the one I use to save all my photos. And the search on that, is just fantastic.
JE: Amazing. Yeah, Amazon’s got that too now in their photo library.
AB: And so you say hey do I have any photos around California and even if I don’t have the geolocation on it. It’ll say hey I recognize that building from San Francisco, so this is probably one from there. Oh by the way these were all taken at the same time so those are probably…
DC: Likely all…
JE: Really cool stuff. I think you’re gonna see that embedded in more and more applications. It’s gonna be in cameras that are recognizing people as they are walking through business, the front door of a business, they’re gonna be able to say look I recognize that person. He looks like so and so, or I know that person has a certain demographic, he’s 35 years old and male. You know it’s going to really shape the way we build applications, the way we interact with objects and devices in a way things start to do autonomous things that they’ve never been able to do before. Robotics or self-driving cars or devices that sort of think ahead. You know we were chatting right before this we were talking about refrigerators that have potential to tell you when something is taken out of the fridge and not replaced. And I’ve suggested maybe what kinda recipe I could make tonight. And I don’t think that’s so far off. It’ll be able to look in the fridge and tell you you’ve got enough of these ingredients to make recipe X or Y or Z.
DC: And you might wanna use the milk ‘cause it’s about to expire.
JE: It’s about to expire.
AB: I like that.
JE: That’s very real and very on the near term horizon for us all.
DC: Well the thing that really stood out to me about machine learning is it’s the logical continuation of a couple of the predictions from last year. You mentioned big data being a huge growing thing.
JE: Key differentiator for companies.
DC: Well last year that was kinda what we expect to see. You called that out and with all of that data machine learning is really useful but the other thing you mentioned was a lack of talent. There’s just so much new technology data scientists get paid ridiculous amounts of money to parse through all that data. But with machine learning, you’re able to take… You don’t have to pay somebody a lot of money to go and dig out for you and basically just sick a machine on a problem and say go, go solve that. And it can go and get that done and bring you back something usable. And that’s really exciting to see you know some of these snowball effects of a year or two ago. This is what’s happening. It’s gonna be interesting to see which of the predictions your making this year impact this trickle down effect, this domino effect of what gonna happen.
AB: We’re gonna be talking about 2018, 2019.
DC: Because these thing happen.
JE: We’re gonna all be replaced with robots.
DC: Very very intelligent machines.
AB: As long as two of the robots are bald and one of them is wearing a hat so.
DC: I can part with my hat if it’s important to the show.
AB: Can you part with the hair? That’s the big question.
DC: Mechanically yes.
AB: Okay so, we’ve gone through a bunch of them there. Let’s go ahead and jump to Google Cloud. That was one of the other ones.
JE: Yeah this was a good one. I put it on the list because last year at the beginning of the year a lot of people looked at Google’s cloud and said you know it’s there, it’s doing okay but it’s not really setting the world on fire. But in 2016 Google made a lot of investments. They brought on a new leader for their cloud division. Diane Greene, she was a former CEO of VMware. Knows how the enterprise market works. Knows how to sell to enterprise buyers. Knows what they need, really has a handle on that world. And she’s now a lot of new… You know, she’s driving the company. So she’s driving a lot of new ventures or developments within the cloud. They are making a lot more noise. People are starting to pay attention. You’re seeing more companies moving over to Google Cloud. That could be behind the scenes negotiations but it also could just be that they’ve got a pretty good product. And so my prediction is that 2017 is the year the Google Cloud becomes a cloud that you can’t ignore. You’re gonna start to not just put Amazon and Microsoft on your shortlist. And say that’s all I really need to pay attention to. You’re gonna put Amazon, Microsoft, Google and possibly other cloud technologies on there as well. Because again back to that array of services. You’re not gonna want to miss out on some of the capabilities. They’re very good at marketing. They’re built on, you know, search and marketing. And their cloud has a lot of capabilities. Machine learning is something Google really helped pioneer. A lot of these technologies that we’re talking about. And they’re gonna be exposed in their cloud. I just put them on my predictions because I feel like this year is the year to really pay attention to Google. Nothing more than that. It’s not, you know, I don’t have specific predictions but I think everybody aught to really take note and you know if you’re a technologist get your hands dirty, go learn it, see if it’s something that maybe has a role within your application.
DC: The work they’ve done on Kupernetes and the ability to build things that are so fault tolerant and robust that heal themselves, pretty natively is really incredible. And we always talk, you know, week in week out about stability and being able to have whatever it is that you do available all the time. Customers more than anything else just want to be available, they want uptime.
JE: Uptime, performance and uptime, right? Reliability.
DC: And all sorts of different cloud providers have different ways of approaching that. I mean there’s all sorts of tools on all sort of technologies being developed all the time. But Kupernetes has done so in a really specific way with containers that allows for real granularity and those sorts of efficiencies that give you granular control with containers and incredible stability and self-healing are really really compelling. And I think that this is the year.
JE: Yeah Kupernetes is a big part of this as well. I mean the mind share around that technology is just amazing. It’s grown like crazy everybody is paying attention to it. It’s iterating very rapidly, obviously because they’re pulling some of the very best technology out from behind the wall at Google and putting it out sorta on display in the form of Open Source and products that they’re launching. So I do think that, you know, combined with just their business focus on the cloud, their technology is really amazing and those two things are gonna be a great great year for Google, I think.
AB: Let’s see, the last thing on your list of predictions. One thing that I know I’m personally excited about. The internet of things in our home. Yeah, well this again dovetails into a lot of things we’ve talked about.
DC: Right wrap it up.
JE: The most visible thing that we can see today in this area is just really the Amazon Echo. The Amazon Echo, maybe the Google Home. Some of these devices that are making their way into our coffee table. And I think you’re gonna see a lot more of that. I think your gonna see a lot more audio interfaces where you’re talking to your devices. You’re asking questions.
DC: It makes sense.
JE: It does make sense. It simplifies, you know there’s a lot of things in the home that you shouldn’t have to pick up a keyboard or remote to take care of. And I think this is really really where we’re going to see a lot innovation. And you know it’s going very fast. I bet you many many people got an Echo for Christmas and it’s sitting on the coffee table right now. Your kids are probably asking it to tell it jokes. And I think that’s just the tip of the iceberg. The integration with the rest of your IOT, your light switches, your garage door, your alarm system, all that stuff. The integrations are gonna be the rich nature of where this goes next.
DC: And they’ve made that really acceptable.
DC: That’s one of the things that differentiated the Alexa platform no matter what device you’re using it on from the others. Is that it’s open and people are building all sorts of tools. You can teach your Echo skills that are really tailored to what you wanna do. You can build skills on top of it very easily and simply.
JE: Some one this week wrote a blog that really compared Alexa to an operating system. And they said that really this is the operating system for the home. You know, we’ve had a data center operating system. We have server operating system. We don’t have one for the home. And Alexa is starting to look like an operating system. It has interfaces, it has sorta services that are exposed and you can integrate with them. And I think that’s probably right. I think that Amazon sees this as a foundational technology. But they’re not alone. Google’s working on it. Apple’s certainly working on it. Competitors are out there. And they’re all trying to get into your home or into your office wherever that might be. So pay attention.
AB: It’s really interesting. You know, we picked up a Google Home this holiday season. I know Drew you mentioned you have a Dot I think. And they’re great. They’re neat little devices. Amazon has a year and a half, two year head start. So they’ve got a lot larger ecosystem but I’m already enjoying the Google ecosystem. And I think Google’s gonna be able to allow a lot of services to connect to build some really cool things.
JE: That’s the exciting part and you know all of this stuff is going so fast and companies are trying to take advantage of this new technology that all relies on cloud. A lot of this stuff that we talked about today you can’t do without cloud. And you can’t do cloud without expertise to use it. That’s again where Rackspace is focused. So a lot of my predictions come back to some of the things we’re doing behind the scenes here to enable companies to jump in with both feet and do these kinds of integrations and build these kind of services and do it without having to hire every single person you’d absolutely need. You can do it with a shoestring budget nowadays.
DC: And you’re gonna hire them and then your gotta keep them. And both of those things
JE: They’re hard.
DC: Are difficult propositions.
AB: It is and again it goes back to last year’s prediction of shortage of talent, you know the data. I think your 2016 predictions were pretty spot on and they’ve really fallen into this year.
JE: Yeah absolutely.
DC: Yeah, I think last year was kinda broad buckets. And you’re trying to be more granular with this one. And as things get more and more complex and more and more interesting you’re gonna have a list that’s too long to publish.
JE: It’s getting harder. This is getting harder and harder. That crystal ball is like little crystal marble size. I have to look really hard to find the predictions. ‘Cause some of these things are obvious. I mean we’re talking about things that are trends that are multi-year trends. So it’s hard to sometimes put a pinpoint on this. But again it’s exciting for me you know really as a technologist and as a leader at the company to think about how we shape our product offerings in response to some of the things that the customers are looking for and our rackers are excited for as well.
DC: Yeah, having culture that’s so interested in learning new things. I think you put a few really good markers out there. Places where people who want to learn the next thing get their feet wet on some specific new technology or where they think the industry’s going. You’ve really highlighted a few really really solid places.
JE: Well the truth is a lot of these predictions came from rackers. I talk to rackers everyday and they tell me what they’re working on and they tell me what they’re excited about. And I just sorta distill it into this once a year blogpost. It’s not just me coming up with this stuff. You guys, you know, out in the trenches really working with customers everyday, you know, those rackers are the ones that really are working on the future. And I’m excited to be a part of it.
DC: It’s a fun ride to be on.
AB: Well we’re definitely looking forward to observing these as the year progresses.
DC: Participating in them.
AB: Participating in them, building some of them and helping our customers do great things. We’ll tentatively pencil you in for January 2018.
JE: Absolutely, excited to be here.
AB: To go back through those and go through the next year’s predictions. Well on that note, we’ll go ahead an wrap it up. And thank you for coming out and being on the show again and sharing these predictions. And if you haven’t read them yet you can go to the Rackspace blog and check them out. They are all available there as well. In fact, all going back to 2014 or 13 or so.
JE: Yeah, you can find them. Just search Rackspace blog predictions and you’ll find it.
AB: They’re all there and I think you can see a nice story unfolding of them. Of how technology has been moving and so. That’s all there. So on that note, I guess we’ll go ahead and say goodbye. I guess we will see everybody next week. Take it easy everybody. Thanks.