 From theCUBE Studios in Palo Alto in Boston, connecting with thought leaders all around the world, this is a CUBE conversation. Hi, I'm Stu Miniman, and this is theCUBE's virtual coverage of AWS Summit Online. Happy to welcome back to the program to help give us some insight into what's happening in AWS last week. And today, that is Corey Quinn. He is the cloud economist at the Duckville Group. Corey, I know it's always a pleasure to see me. Thank you very much. Is it always a pleasure to see you? No, it is always a pleasure to see me. Thank you for once again exhibiting remarkably poor judgment and inviting me back onto your program. Yeah, Corey, you've been on the program a few times now, including at some of the AWS shows, AWS San Francisco, AWS New York City, re-invent we see you, but this is the first time we've had you on online. So it gives a little bit about what that the impact of the global pandemic has been meaning to you. And more importantly, what you've been seeing from our dear friends at AWS. Sure, the fact that I'm not traveling anymore and spending almost all of my time at home means that I'm a lot closer to the edge when it comes to the content I put out because I no longer have to worry about someone punching me in the face. But other than that, from a business perspective, things tend to be continuing on much as they have before with of course different customer concerns. The more interesting question from my side has been, what is the effect this is having on people? Because yeah, we're working from home remotely. It's not really a fair test of how well you can do remote because I've been doing it like this for years. And there's a sense of existential dread that's hanging over people's head. More so than usual, more so than right before the AWS re-invent, when you're wondering if you're about to have your entire business put out of business by an AWS competitor. Now it's just that sort of dread that never goes away because they won't deliver the keynote if you'll pardon me of using a metaphor. Yeah, it's been really interesting to watch. Corey of course, I mean, Amazon is a big player in the industry before this. Amazon one that gets talked about a lot in the news. Amazon overall, when this first hit, they announced that they were hiring 100,000 people. Then they went through that faster than anyone could believe. You think about having to hire everyone remotely. My joke was, Alexa, please screen everybody and hire everyone. But then they hired another 75,000 and it's not just the warehouse and the whole food people because I've seen a number of people that I know getting hired by AWS too. So, you know, you talk it's all about the people that the number one folks are in the pandemic people. How's Amazon doing? What feedback are you getting for how they're doing? Well, I don't have too many internal sources that will confirm or deny things of strategic import because it turns out that I'm generally not cleared for those things. Who knew? Something I'm picking up on across the industry has been that if you're building a hyperscale cloud provider, you're not looking to next order. The investments you make today aren't going to be realized three to five years. No one is currently predicting a dramatic economic impact that's going to be felt for a decade based on the current situation. So, yeah, AWS is still investing in people which is always going to be the limiting constraint. They're still launching regions. We had two launch within a week past month and we're still seeing a definite acceleration if anything of the pace of innovation as AWS like today. Now, from my perspective, that's both reassuring that some things never change and of course the usual level of depression where, oh good, there's still more services to learn what they do, learn how the names work, find ways to poke holes in their various presentational aspects and of course try and keep the content relatively fresh. There's only so many times you can make the same joke where people start to plan. Yeah, absolutely. Corey, you bring up a really good point. Amazon, they have a long strategic plan there. If they're building new data centers, they're building the power infrastructure for these things. It's not something that they're going to change on a dime. They plan these things out far in advance and AWS does of course have a global scope. I really wonder from an operational standpoint, are there any pressures on them? You wrote an article relatively recently talking about one of the other public cloud providers that has had to prioritize certain customers and even had some performance issues. AWS seems to be running through this, dealing with the capacity. I've had phone systems that have problems. Everybody has, when they're working from home, and tension internally, even if you've got a gig bandwidth, when the entire neighborhood has children on the classrooms online for video, there's pressures there. So Corey, it seems from what I've seen, AWS operationally is running well and keeping things all up and running. Am I missing anything? Not at all. AWS is fond of saying there's no compression algorithm for experience. As I'm fond of saying, that's why they charge per gigabyte. But what that means is that they've gone through a lot of these growing pains and largest disruptional stories in 2010 through 2012. EBS outages causing cascade failures as everyone saturates links as they roll from region to region or availability zone to availability zone. They understand what those workloads look like and what those behaviors are and they've put an incredible amount of engineering into solving these problems. I think that anyone who looks at this and doesn't see this happening is in a very fortunate place because we don't have to. It's approached utility level of reliability. You don't wonder every time you turn the faucet on whether water is going to, and we're now at a point of seeing that with AWS resources. Now there are still going to be recurring issues and there have been basically since this thing launched. A particular instance size and family in a particular availability zone of a particular region may be constrained for a period of weeks. And that is something that we've seen across the board but that has less to do with the fact that they didn't see this stuff coming and plan appropriately and more to do with the fact that there's a lot of different options and customer demand is never going to be an exact thing to predict. We are seeing some customers dramatically turn off capacity and others dramatically ramping capacity up. It comes down to what is the nature of this current pandemic on their business? Yeah, well, this absolutely does put some of those promises of the cloud test. I should be able to spin things down. Some things I should be able to turn off and if I have to, you know, shut down pieces by business, I should be able to do that. I'm curious what you've heard on changing demand out there, Corey. You know, on the one hand, you know, customers, they're pre-buying, they're getting reserved and they're making sure that they can, you know, optimize every dollar. But when something like this comes up and they need a major change, you know, are they stuck with a lot of capacity that they didn't necessarily want? Sometimes it comes down to a lot of interesting variables. For me, the more interesting expression of this is when companies see demand falling off a cliff as users are no longer using what they've built out, but their infrastructure spend doesn't change. That tells me that that's not a particularly elastic infrastructure. And in fact, when people are building elasticity into their applications, they always interpret that as scaling up rather than scaling down. Because the failure mode of not scaling up fast enough is you're dropping customer requests on the floor. The failure mode of not scaling down fast enough just means you're spending extra money. So when you see user demand for an environment cut by 80%, but the infrastructure cost remains constant or the infrastructure usage depending, that's a more interesting problem. And you're not going to have a lot of success asking any cloud provider for an adjustment when, well, okay, you're suddenly not seeing the demand but your bill remains the same. What is this based upon? You need to actually demonstrate a shortfall first of, wow, you know how we normally spend a million dollars a month? Well, and now we're spending 200 grand a month? Yeah, about that. And once you can do that, there are paths forward. I have not yet heard stories about, frankly, any of the big three cloud providers, absolutely hanging customers out to dry in a cloud infrastructure company. I have heard whispers about, for example, with G Suite where they're not willing to, and this feels like a very dark way to go, but I'm going for it, where, well, we just laid off a third of our staff and we get a break on the annual licensing for those seats on G Suite. And the answer is no. That feels like it stings and is more than a little capricious. Yeah, no, absolutely. You know, one of the things that the underbelly of Bass is, you know, oh, it should be elastic like cloud but oftentimes you're locked for one or two contracts. And if all of a sudden you find yourself with needing half the demand and you call them up, you know, are they going to give you that break? So, you know, price inquiry, you know better than most, so. All right, let me spoil it for you. Every provider is going to give you a break on this because this is a temporary aberration as far as the way the world works. We are not going to start seeing global pandemics every year, I hope. And when this crisis passes, people are going to remember how their vendors treated. And if it's, well, we held your feet to the fire and made you live up to that contract, that sticks with people. And it doesn't take too many stories like that or people pulling lawsuits out of the pacer to demonstrate that a company beat the crap out of them to say, huh, maybe that's not where I want to make my sizeable cloud investment. Yeah, absolutely. So Corey, how about, you know, that there's certain areas where I heard, you know, certain groups that maybe were slow rolling cloud and all of a sudden realized that when they're working from home, they can't go hug and adjust their servers that are saying, oh geez, maybe I need to hop on this. Then there's other services. You think, you know, VPN usage must be through the roof. Workspaces, you know, when first announced, you know, many years ago was a bit of a slow roll, had been a growth area for Amazon for the last couple of years. Are you hearing anything specific to new services or increased growth in certain services, like I mentioned? There are two patterns that we're seeing of all. One is the traditional company you just described, where they build out a VPN that assumes some people will occasionally be working from home at a 5% rate versus the entire workforce 40 hours a week. That model is straining everything. Whereas if you go back for the last 10 years or so and look at a bunch of small businesses that have started up or startups that have launched where everything they're using is a SaaS service or a cloud service, then there is no VPN. I don't have a VPN, for example, the fact that I have a wireless network here at my house and I'm at this location, this IP address isn't whitelisted anywhere. The only benefit that this network has over others is that there's a printer plugged in here and that's it. The identity model of I authenticate to these services via API credentials, via username and password, via, I don't know, enchanting something and they send an email that I click the link that winds up handling the authentication and identity and there's no bottleneck in the same VPN direction. I feel like this is going to be the death knell for a lot of VPN centric corporate IT architecture. All right, Corey, one of the other things about AWS is they don't stop. And what I mean is, you talked about them always being online, but every week there's new announcement. It keeps feeding your newsletter, it keeps feeding your feed, everything going on there. How is number one, the announcement train from AWS going and anything specific, John Furrier was interested in Amazon Appflow, something that was released relatively recently. The problem with a lot of these new services that get released relatively recently is that it requires time to vet out how it works, how it doesn't work, how it should have wound up being implemented to solve your particular use case or in my case, how they could have named it better. But you're not able to come up with those things off the top of your head the first time you see it because it's irresponsible at scale to deploy anything into production that you don't understand its failure cases. Right now with everyone scrambling, most companies are not making significant investments in new capability. They are desperately trying to get their workhorses online and stay afloat and adjust to be very rapidly changing economic climate. And oh, they built a new data store or something of that nature is not going to be this sort of thing that gets people super excited in most shops. That time will change, but I do feel a bit of pity right now for a lot of these product teams who've been working away on these things for months or years and now suddenly they're releasing something into a time when people don't care about it enough to invest the effort that it deserves. Yeah, you bring up a really good point, Corey. There are certain things, if I was working on a project that was going to help me be more agile and be more flexible, I needed that yesterday but I still need that today. Some other projects might take years to roll out. AI is a technology that has been growing and maturing over the last couple of years where IoT solutions are a little bit more nascent. So is what you're thinking, it's a little bit more stickier knitting and the solutions and the products that you're leveraging today and some of the more visionary and futuristic ones might be a little bit of a pause button for the next couple of months. Exactly, if you're looking at exploring something that isn't going to pay dividends for 18 months, right now the biggest question everyone has is what is the long-term repercussion of this going to be? What is the year, what is it we're going to look like in three years? And because that's where a lot of these planning horizons are stretching to. And the answer is, is look, when I wind up doing a prerecorded video or a podcast where I talk about this stuff and it's not going to release for four days, I'm worried about saying something that's going to be eclipsed by the news cycle. I worry on my podcast recording, for example, that I'm going to wind up saying something about the pandemic and by the time it airs in two months, it's, oh, look at this guy, he's talking about the pandemic. He doesn't even mention the meteor and that's the place right now where people are operating from. It becomes much more challenging to be able to adequately and intelligently address the long-term when you don't know what it's going to look like week to week. Yeah, absolutely. For our viewers, when you hear my segment on Corey's podcast and you wonder why we did talk about Murder Hornet, it's because we missed that one week window that we are in right now when we are talking about Murder Hornet, not when we recorded it, not when we released it. So really good point, Corey. Talk to, Corey, data is one of the most important things. You've done a lot about data portability, all the costs involved, cloud, Amazon's trying to help people with bringing data together. I said in one of the interviews with Andy Jassy a couple of years ago, while customers were really the flywheel for AWS for a number of years, I think it is data that is that next flywheel. So I'm curious your thought is how enterprises think about their data and AWS's role there. Incorrectly, if you want me to be blunt. There's an awful lot of movement, especially as we look at AI and machine learning to gather all of the data. I've been on cost optimization projects where, wow, that's an awful lot of data. Sitting there in that S3 bucket, do you need it all? And I'm assured that, yes, all of the sales transaction logs from 2012 are absolutely going to be a treasure trove of data just as soon as they figure out what to do with it. And they're spending large piles of money on this. But it's worse than that, because it's not just that you have this data that's costing you money, that's almost a byproduct. There's risk to an awful lot of forms of data with regulation that continues to expand. Data can become a toxic asset in many respects. But there's this belief of never throw anything away. That's not really ideal. Part of the value of a same data management strategy is making sure that you can remove all of the stuff that you don't absolutely need. Right now with AI and ML being where they are, there's this movement for, oh, keep everything because we don't know what that's going to be useful for down the road. It's a double-edged sword. And enterprises are, at this point, not looking at this through a lens of this thing could hurt me so much as they are, this thing could possibly benefit our business in the future. All right, so Corey, I've really noticed over the last few months, you've spent a bit more time talking publicly about some of the other clouds that aren't AWS. So while we are covering AWS Summit online, give us what you're hearing from Microsoft, Google, and others, any strategies that are in view, any customer movement that is worth mentioning. Sure. I think that we're seeing customers move in the way that they've always been moving. People made a bit of a kerfuffle about a blog post I put out with the extremely clickbait title of Zoom chose Oracle Cloud over AWS, maybe you should too. And there were a few conclusions people drew, understandably, from that particular headline, which was, for example, the idea that AWS had lost a workload that was being moved from AWS to Oracle, not true. It was net new, Zoom already has existing relationship with both Azure and AWS by their own admission. But the argument, what I took that particular change to be in my case was an illustration of something that's been bugging me for a while. If you look at AWS data transfer pricing, the publicly posted stock, which again, no one at this scale is going to pay, it is over 10 times more expensive than Oracle Cloud. And what that tells me is that I'm now sitting here in a position where I can make a good faith recommendation to choose Oracle for cost reasons, which sounds nuts, but that's the world in which we live. It's a storytelling problem, far more than it is a technical shortcoming. But that was interpreted to me in that, oh, Oracle's on the rise, AWS is in decline. Zoom is a very strong AWS customer and has made public commitments that they will remain so. Right now, this is what we're seeing across the board. You see Zoom doing super well. They're not building out a whole lot of net new either. What they're doing is just desperately trying to stay up under crushing unprecedented demand. That's where the value is coming from right now of Cloud's elasticity. And they're not doing, oh, you know, we're going to go ahead and figure out if we can build a new continuous deploy process or something that makes on call a little bit less brutal. That's not what anyone's focusing on. It's here, wow, this boat is sinking. If we don't stay up, grab a bucket, start bailing. And that is what they're doing. The fact that they're working with every Cloud provider, shouldn't come as a surprise. Yeah, well, it's interesting. I'm thinking about Zoom and one of the things that I've been watching over the last couple of weeks, everybody has is the daily updates that are happening related to security. Think back, six, seven years ago, Amazon had, this is our security model. We're not changing it for anyone. Now, Amazon has a much more flexible and nuanced security position. There are still violent principles that Amazon will not and cannot shift. So to be clear, they have different ways of interfacing with security and different ways of handling data classification, but the rules that you knew are not changing. It's not, well, surprise. Now suddenly every Amazonian who works there can look through your private data. None of that is happening. I just want to be very clear on that. Yeah, no, you're absolutely right. It's more security getting more attention even than ever. And it was already coming into 2020 before everything changed was one of the hot topics. Corey, I'm curious, we're looking at a virtual event for AWS. Have you been to some of these? Are you getting burnt out from all of the online content? I'm sure everybody's getting tired of you. So are you getting tired of everyone else? I don't accept that anyone would ever get tired of me. I'm a treasure and a delight. But as far as online events go, I think that people are getting an awful lot profoundly wrong about them. For example, I think that people focus on, well, I need to get the best video and the best microphone, and that's the thing that people are going to focus on rather than maybe I should come up with something that someone wants to listen to. People are also assuming that the same type of delivery and content that works super well on a stage for 45 minutes is not going to work when people can tab over to something else and stop paying attention. You've got to be more dynamic. You've got to be able to grab people's attention. And I think that people are missing the forest for the trees here. You're just trying to convert an existing format into something that will work online. In the immediate short-term, everyone is super sympathetic. It's not going to last. People are going to get very tired of the same tired format he tropes. And there's only so much content people are going to consume. You've got to stand out and you've got to make it compelling and interesting. I've been spending a lot of time trying to find ways to make that work. Yeah, I had a great conversation with John Troyer. He said, we can learn something by watching some of the late night talk those. I think there's a new opportunity for you to say there's a house band. You have a small child at home. Give her a tambourine. There's your house band. You can have a lot of fun with it. Oh, absolutely. Especially during a tantrum. That's going to go super well. I'm just going to watch one of her meltdowns about some various innocuous topic. And then I'm going to wind up having toddler meltdown, the Amazon S3 remix. And I'm sure we can wind up tying it back to something that is hilarious in the world of cloud. But I'm trying to pull off a little bit longer before I start actively exploiting her for internet points. I mean, I'm going to absolutely do it. I just wanted to get a little taller first. All right. Well, Corey, I want to give you the final word, AWS, the online events happening. Give our audience what they should be looking at when it comes to their AWS estate. Cool. As usual, pay attention to what's coming out. It's always good to have a low level awareness of what's coming out on stage. But don't feel you need to jump in and adopt any of it immediately. Focus on the things that matter to your business, just because something new and shiny is announced on stage does not mean it's a fit for you. Doesn't mean it's not, but remain critical. I tend not to be one of the early adopters in production of things that have potential to wind up causing challenges. And I'm not saying, oh, stay on the exact old stuff from 2010 and nothing newer. But there is a bit of a happy medium. Don't think that just because they released something that, A, you need to try it, or B, it's even for you. No AWS service is for everyone, but every AWS service is for someone. All right, well, Corey Quinn. Always a pleasure to catch up with you. Thanks so much for joining with you, joining us. Thank you, Stu, as always for suffering with LinkedIn arrows. I appreciate it. All right, thank you for watching, everyone. Lots of coverage of theCUBE at the AWS summit online. Check out theCUBE.net for all the offering. And thank you for watching.