On this episode of The Six Five – Insiders Edition hosts Patrick Moorhead and Daniel Newman welcome Jeff Barr, Vice President and Chief Evangelist at AWS to discuss AWS’ response to COVID-19 and exciting new product announcements.
AWS’ Response to COVID-19
During this COVID-19 pandemic, AWS has been focused on helping customers continue to operate as efficiently as possible. One of the first things the company did was make it easier for businesses to get access to Amazon Workspaces, a virtual desktop in the cloud. They also had special offers for Amazon Chime and Amazon WorkDocs to help companies facilitate collaboration.
In addition to helping customers, AWS launched the Diagnostic Development Initiative (DDI) to provide support for innovation around both testing and development of different kinds of solutions. They’ve also made AWS compute power more available to researchers who are studying different possible solutions and vaccines.
Finally, AWS launched a public data lake with curated data sets so experiments could actually run queries on real data which will hopefully allow scientists and researchers to find a cure faster.
AWS Virtual Summit
Traditionally, AWS hosts Global Summits throughout the year, but given the current stay-at-home orders still in place, they’ve had to pivot to a virtual summit the first of which was Wednesday, May 13.
The summit is free, online and anyone interested can join. The event is designed to bring the cloud computing community together to connect, collaborate, and learn about AWS. Attendees will hear from CTO Werner Vogels, CEO Andy Jassy, and several other AWS employees who are subject matter experts in various AWS categories.
Breaking Down AWS Announcements
Jeff, Patrick and Daniel spent time discussing several of the new AWS announcements from the last few months that are making a difference for customers all over the world.
Macie is a security service that uses machine learning to automatically discover and protect sensitive data in AWS. This service has been around for about a year, but AWS has recently made some great additions including updating the machine learning models so customers can scan for even more types of private information. They’ve also added some customizability if customers have special data types that might have different kinds of proprietary or sensitive data inside.
The best part is AWS has lowered the price down to a fifth of what it previously cost so more customers are able to benefit from the different services.
Amazon Elasticsearch Service
Data is growing exponentially in quantity and size and Amazon Elasticsearch Service customers are needing new ways to store and access data as efficiently as possible, specifically data that was collected for the long-term. The current storage tier was called Hot for quickly accessed storage. AWS just introduced a new tier — Ultra Warm — that will hold the more historic data that customers don’t need as often and it will take slightly longer to access.
SaaS applications have been highly functional, but the data created and collected across these many applications has effectively been in a silo. Amazon AppFlow is a service that allows customers to securely transfer data between SaaS applications. Customers can unlock the access to that data, making it easier for data to flow from the SaaS app into AWS as well as the other way around. and the other way around as well. Customers can run data flows on a schedule, in response to an event, or on demand — basically whenever they need it.
AWS’ enterprise search tool Kendra gives customers powerful natural language search capabilities across websites and applications so users can easily find information in the data spread across the enterprise.
While most search tools use keyword queries, Kendra is able to use natural language questions to search through portals, wikis, databases, and document repositories to find whatever is needed. It not only captures the data, but also the access permissions for the data too.
Amazon Augmented AI (A2I)
Many current machine learning applications require humans to review predictions to ensure results are correct. Amazon Augmented AI or A2I makes it easy to build the workflows required for these reviews and provides a built-in review system for common machine learning use cases.
Customers are able to choose three different categories of human reviewers. They can use the 500,000 global workers that make use of Mechanical Turk. There’s a set of third party organizations that have a base of pre-authorized workers, or organizations can make use of a private pool or workers.
AWS Snow Family
The Snow Family devices are physical devices that have both storage and local compute power making it easy to migrate data into and out of AWS. Recently AWS launched an improved version of the Snowball Edge Storage Optimized devices. These devices provide both block storage and Amazon S3-compatible object storage.
AWS did a hardware refresh, added additional processing power, and added some additional SSD storage inside. Now if you launch EC2 instances on the Snowball Edge those instances have access to this SSD powered storage.
You can use these devices for data collection, machine learning and processing, and storage in environments with limited connectivity giving customers the ability to use them basically anywhere.
Finally, AWS also made it easier for our customers to set up and manage the Snowball Family devices with AWS OpsHub, a graphical user interface where customers can unlock, configure, copy data to and from the device via drag and drop even if they’re not connected to the Internet.
If you’d like to learn more about any of these announcements, be sure to visit the AWS website. Make sure to listen to the entire episode below and while you’re at it, be sure to hit subscribe so you never miss an episode.
Patrick Moorhead: Welcome to a special insider edition of the Six Five podcast. I’m Patrick Moorhead with Moor Insights and Strategy, and I’m joined with my very famous and ever present cohost Daniel Newman with Futurum Research. It’s definitely an interesting time to be a tech analyst these days. I’ve gone from being in airports 45 weeks out of the year to grounded since mid-March. Daniel, how are you, my friend?
Daniel Newman: Patrick, doing very well. Yeah, just like you, grounded, just like everyone that’s been checking into our show on a regular basis, probably hearing about this from us, but hey, we don’t know what else to do. Our life, we went from airport to airport, I happened to have a better relationship with certain flight crews than certain family members, not intentionally, just because that was our lives, the tech industry, tech events. And I’ll be honest, Pat, I miss it. I do.
But at the same time it’s been a really tremendous opportunity to get back home, spend a little extra time with my senior daughter, and of course, a senior in high school, not senior in age, and just get a little calm. Because I’m sure we’ll get back out there at some point soon, but you know, what else has been great, Pat, has been the opportunity to have some great discussions on our podcast with some really smart executives from the industry, no different than the one I think we’re going to have today here on the podcast.
Patrick Moorhead: That’s right, Daniel. And we have a very special guest joining us today, who in the cloud world needs no introduction, and that’s Jeff Barr, AWS’s VP and chief evangelist. How are you doing, Jeff?
Jeff Barr: You know, I’m doing great and I’m really happy to be here, looking forward to chatting with both of you.
Patrick Moorhead: Excellent. So while many people know what you do, not everybody knows who you are and what you do. So for those who don’t know you, Jeff, can you talk about what you do for AWS and how long you’ve been doing it? I think you were the original blogger at Amazon, or maybe the original blogger out of any bloggers.
Jeff Barr: Okay. So let’s see, I joined Amazon in the summer of 2002. So that means I’m actually coming up on my 18 year mark, which I find totally unbelievable. My official title right now is VP and chief evangelist. And early on I took my job description as simply spreading the word about our technology and really being the bridge between all of our teams building really cool technology and our customers, and being able to take what those teams built and explain it in ways that would work for and resonate with customers. In recent years I’ve also added sharing our culture to my repertoire. And I often find myself speaking about Amazon leadership’s goals and our working backward model and two pizza teams and the like.
I speak to live audiences on occasion, but I mainly focus online. I’ve been writing the AWS News blog since the fall of 2004. I’ve written well over 3,000 posts for that, all of them in depth, all technical. I write them myself. I make a point of always using the service myself. I listen to and respect what the teams tell me about their services, but I insist on actually using them myself and getting direct hands-on experience. So that’s a little bit about who I am and what I do.
Daniel Newman: That’s a big job. I started thinking about, I was listening back to everything you just said, Jeff, and first of all, 3,000 blogs, there’s a couple of books in there in terms of experience. The second thing is, as we’ve gone quarter by quarter, I always do a deep dive into the earnings of Amazon each quarter and I look at the list of new products, services, zones, regions that come out from AWS, which has been a tremendous growth every quarter. And I can’t even imagine, just from when you started to now, how much more you probably have to cover. It’s got to be tremendous.
Jeff Barr: It absolutely is. And it resonates to me with one of Amazon’s leadership principles, which is, learn and be curious. And there’s not a day that goes by that I don’t get to just dive into something new that a team has built and is excited to share with me. Everything that I learn and then write about, it’s something brand new for me. I’d say no two days, maybe even no two hours are the same. And I constantly find myself at the absolute edge of what I know.
And sometimes maybe I’m stepping over that edge looking down into the abyss. But I think that helps me identify with our customers who are really trying, just like I am, to understand and make sense of and to organize and position all these different services that we are building and making available to them.
Daniel Newman: Yeah. It’s interesting, there’s so much terminology. I do sometimes as an analyst wonder, when people use the words, how many of these words they intimately know versus just utilizing them like marketecture. I have an MBA with a technical background, so I bounce on those lines between technical and just using really big words that sound impressive in meetings. And sometimes we have to really balance. But no, the company has been doing a lot of wonderful things. And I want to ask you a question, about speaking of wonderful things, you know, we’ve had a number of executives come on this Six Five insider podcast in recent weeks, and we’ve heard these incredible stories about tech companies doing some really exciting, interesting, important work around COVID-19. We’ve also written a lot about this, but we’d love to hear through your lens about the stuff that AWS has been doing and give some perspective around that.
Jeff Barr: Well, let’s see, a lot of different things have been happening. And to me one of the most important at first was, I was actually in mid business trip. I was in the Nordics actually, I think I was in Denmark. And the word came through in very early March that we were asked to do our absolute best to stay out of the office and to work from home. So it was really good to hear that the company had our best interests and our health at heart. Very quickly the teams put their heads together and said, “What can we do to make sure that people that are newly working from home, that are remote workers, what can we do to make their job easier and more efficient?”
And so we looked at things like making it easier for them to get access to Amazon Workspaces, which is a virtual desktop in the cloud. We gave them special offers for Amazon Chime, which is communication and video conferencing. And we gave them access to Amazon Work Docs, which is collaboration. So all these services were already available, but the teams very quickly said, “Let’s see if we can put some special offers into place. Let’s see what we can do to really be extra clear about communicating what these services are and the value that we think that they might have for remote workers and for home workers.” So that was on the service side, but I think probably more importantly we also said, “What can we do to really help out in actually finding cures and finding solutions and mitigation?” So we launched something called the Diagnostic Development Initiative or DDI, with the idea that we were going to help with the aid of both AWS credits and the technical expertise of my colleagues to really support some innovation around both testing and development of different kinds of solutions.
We made available things like our AWS compute power so that researchers wanting to do things like protein folding and analysis of different possible solutions and vaccines, that they had compute power at their fingertips, storage, ways to facilitate communication and cooperation. We set up a public data lake that had a set of curated data sets so that experimenters could actually run queries against all these data sets. So I actually, once I got back from my trip, got settled in a little bit, I spent several weeks just really overemphasizing all of the ways that these newly at home workers and learners could get value out of AWS as far as, could they learn, could they connect? What could they do in this new operating mode that they were in?
And was really pleased by the response, the number of people that said, “Okay, thank you for, in these difficult times, pointing out something new I can do, something new I can learn.” I had people say, “Here’s something really neat. I’m going to share it with my technically minded teenagers,” for example. Just was really interesting just to see the amount of just quick and positive feedback that we got from doing all of that.
Patrick Moorhead: Yeah, Jeff, some really impressive stuff. And I’m glad to say, and we didn’t plan it like this, Daniel and I wrote up pretty much everything you listed off in our analyst notes and even on forbes.com. Super, super impressive. The data lake that everybody can get access to with one snapshot I think is incredibly powerful. AWS today has a big event that’s starting. Can you tell us a little bit about it?
Jeff Barr: Sure. So traditionally we’ve run AWS summit events in tens of cities around the world. And because we are not in a position to do that right now due to travel restrictions and lockdowns in various places, we have set up an AWS virtual summit as a substitute for the physical summits that we traditionally run. We’re setting up the summit to be free and online with the end our customers and anybody interested can join. They can hear from Warner, our CTO, they can hear from Andy, our CEO, and they can also hear from my colleagues, the actual subject matter experts in different parts of AWS.
We’re definitely experimenting with this new format. We’ve for sure hosted many events online in the past, but we always view everything that we try at Amazon as some kind of experiment. We put something out there, we test, we measure, we look at the results, we talk to customers, try to figure out, what did we do right, what did we not do as well? What can we do better? So we see this virtual summit as an experiment in this new format. We think it’s going to be a great way to connect with and educate our customers online. So going to be an awesome event.
Daniel Newman: It’s exciting. It’s been really a fun and interesting and sometimes exhausting experience to watch different companies trying to bring events online. I am confident that AWS will get it on the side of right, because every company is wading these waters and figuring it out. But what you’re doing sounds really interesting. So I’m definitely excited to check it out. It’s funny to think that December of this year there may not be a ReInvent, or likely, just because companies aren’t doing events, but there are going to be announcements. Like I said, I alluded to earlier, reading through the earnings, every month, every quarter, AWS is doing a lot of things. And the company’s done a lot of things since AWS ReInvent, since December. Can you talk about some of the key announcements that you’ve made recently?
Jeff Barr: Sure. So one thing I love is that we’re always doing something new, and the tech itself is always interesting and remarkable. But to me, what I find rewarding over time is that I know that what we do is actually driven by real requests from real customers. Because I participate in a lot of customer meetings, and when customers show up and they’d like a full day or even two days’ worth of time with our senior leadership and with the service teams, we always start with what we call the voice of the customer. And the customer gets to go first, and they tell us who they are, what they do, what their specific wish list items are for different, either for additions to existing services or brand new services, and sometimes even brand new business areas. So with that as a backdrop it’s great to think about just the customer driven nature of this.
One thing we could talk about would be Amazon Macie. So this service has been around for a year or so. We’re making some great additions to it. So the idea of Macie, it helps our customers to discover and then protect sensitive information that they might store in their S3 buckets. So this could include various kinds of PII, personally identifying information, it could be names, mailing addresses, credit card numbers, and so forth. So Macie works by simply, the customer points it at an S3 bucket and then says, “Go ahead and scan.” And using some custom trained machine learning models Macie looks for those various kinds of information and then raises alerts. The news that we have is that we’ve reduced the price for Macie down to 1/5 of the previous price. It used to be $5 per gigabyte and we’ve reduced that down to $1 per gigabyte.
We have updated the machine learning models in there so they can look for even more kinds and formats of PII. We’ve made it available in additional regions, and we’ve given the customers some additional customizability if they have special data types that might have different kinds of proprietary or sensitive data inside. Macie now has the ability to crack those open to access those different pieces of data.
Patrick Moorhead: Jeff, I have to ask you, how do you do an 80% reduction in price on such a sophisticated service?
Jeff Barr: Okay, so I don’t always get to be privy to the details, but one thing that I do know is that after we build the first iteration of a service and we put it online, that’s just the beginning of the engineering effort. That’s by no means the end. And so I do get to see reports from the teams as they continue to monitor and to tune and optimize these different services. They’re always looking for ways that they can add efficiency. And some of that’s brought about by new technology, sometimes the scale itself allows us to work more efficiently. You can imagine ways that we can process these machine learning models and do inferencing more quickly. And as we gain experience with these different services there’s a continuous effort. And this is one of those things that’s both driven by scale and is a necessity at scale, this idea that when you’re running something that the entire world can access, every possible improvement is mathematically worth doing in order to make it more efficient and to get those prices down so that more customers are able to benefit from these different services.
Patrick Moorhead: Yeah. I appreciate the deeper dive on there.
Jeff Barr: Yes. Okay, so the next one is, we have an existing search service called Amazon Elastic Search Service, and our customers use this to collect and then analyze the logs from their sites, their devices, their sensors. As you probably know, this kind of data is actually growing exponentially in quantity and size. The customers are now saying that they want to take this data, all these logs, all the sensor data, they want to store it for the long term, for months or for years. But they need to do that as efficiently as possible. We introduced this new service called ultra warm for Amazon Elastic Search Service. So the Ultra Warm prefix basically says that there’s a new storage tier. The current storage tier was called hot and the new one is ultra warm, so a little bit less than hot.
And so this is a distributed cache that sits on top of Amazon S3. And in typical use what’ll happen is that the current data will be in the hot, very, very quickly access storage. And the more historic data is going to be in the cache, and then in storage it takes slightly longer to access. Again, big price difference between the two, the ultra warm storage is priced at 1/10 of the cost of the hot storage. So this is another way that things are getting more cost effective for customers. The cool thing is, this all happens under the covers, the existing code and queries and tools the customers have. All those are run without any changes on the customer side.
Daniel Newman: Just interesting, I’m just scanning through this massive list, Pat and Jeff, of different things, and what catches my eyes, and we talk about applications too, I’d love to hear a little more about app flow as well.
Because you got this, it looks like a new integration service. And I think that this is huge with what’s going on with data and just the massive opportunity there. And we’ve heard so much growth in terms of how companies are trying to enable more rapid development of data in platforms. And this app flow looks really interesting. Can you share a little more about that?
Jeff Barr: Sure. So the idea is that traditionally a lot of these SaaS applications have been highly functional, but they’ve effectively been something of a silo. So there’s a lot of really awesome data that’s somewhat captive within the different SaaS applications, within Salesforce or Marketo or Service Now or Slack or ZenDesk. So the idea of app flow is we want to really unlock that access to that data, make it easier for data to flow from the SaaS app into AWS and the other way around as well. So this is a bi-directional integration service where we unlock this access to this data.
And because these apps are often used by enterprises at scale, this integration can go at petabyte scale or even exabyte scale if necessary, and giving our customers the ability to take data from these apps, process it within AWS, maybe even do some integration between separate SaaS apps, while doing it both safely and securely.
Daniel Newman: Yeah, that’s a big thing right now, though, is going to be the data flowing through multiple applications, being able to work towards more of an open source of all these different applications. You’ve seen different initiatives. I know AWS actually announced a fairly significant one, Jeff, if I recall, this wasn’t part of these new wave of announcements, but last year, I believe it was with maybe Genesis and Salesforce, there was a partnership to actually streamline data in the schema of data and such, so that all the data could flow. I imagine this is building on that or it’s an expansion? Or is there any interrelation there?
Jeff Barr: That’s right. I do believe you’re referring to event bridge. So event bridge basically couples actions that happen inside these applications with AWS. So when a customer is in a SaaS application and one of their users adds a new contact or adds a new sales record or puts a row into a database event, bridge basically has triggers that are going to flow that action into AWS, that can then be used to initiate something else happening. You can imagine those triggers actually initiating the use of app flow. So app flow and event bridge are going to ultimately be able to work together to build this nice integration between multiple SaaS apps or SaaS apps and AWS.
Patrick Moorhead: It’s exciting stuff. So Jeff, let’s move to enterprise search if we can. I got briefed on some of the updates to Kendra, and also sat there in December in Las Vegas and got a whole lot of information out of it. What’s new with Kendra these days?
Jeff Barr: Okay. So Kendra is all about enterprise search with natural language. It’s powered by machine learning. And the idea is, just like we talked about with app flow, effectively it’s running across multiple different data silos, this time data silos within enterprise. So portals, wikis, databases, document repositories. Kendra supports keyword queries, but is actually better served if you use natural language queries. One really cool thing it does is that when Kendra is scanning all of this different data, it actually captures not just the data, but it captures the access permissions for the data, and it then reflects those in the search results that are returned. And it always returns links back to the original data sources but it respects the various permissions for those pieces of data.
One recent cool development that we did is that the Allen Institute for AI that was founded by Paul Allen, which just happens to be just right up the street from me, just a little bit to the east, they put together something called the COVID-19 open research data set. And so Amazon Kendra actually powers a natural language search of that data set. So this is another thing that we’re doing to help out with efforts around COVID-19.
Daniel Newman: Yeah, that was a really interesting one, Jeff. Enterprise search as a whole has left a lot to be desired. So when I first heard about Kendra, I believe, gosh, I feel like I was in Hawaii when I was first reading about it. Sorry, those were the better days. I was really encouraged. And so I’m looking forward to actually playing with that a little bit and getting a little bit more customer feedback about how that’s starting to really work inside of organizations.
You do have one more big release that I’d like to hit since we’ve got a little bit of time left, and that’s something, a really interesting announcement was around the general availability of Amazon augmented artificial intelligence.
That’s a mouthful, but we’ve heard for a long time, Jeff, that the relationship between AIML and data is not all machine. There is really an important role that humans have to play in terms of optimizing and managing and improving. And it sounds like that’s something that’s really part of what AWS is doing here with this new offering, A2I is the abbreviation.
Jeff Barr: Exactly right. And so I think that when the general public hears about machine learning it sounds like magic, and that you’ve got this all-knowing oracle that you just ask it things and it gives you the right answers just by some kind of amazing technology magic. The reality, as we know, is it’s a lot more complex and a bit more involved in that. So as developers have to build models, train models, and then you need to test your models and actually say, “When I actually start asking this generation of a model to do some inferencing and to give me some results, how good are they? And as I continue to refine my model, am I actually getting improved accuracy or not?”
So testing that at scale, you need humans to look at the computer generated results and compare those to expectations. And so the A2I service, basically I look at it as putting a human in the loop and basically saying, “Here’s a set of predictions that the models have made, and how well are these models doing on these predictions?” So effectively a human review of the accuracy of the models. And you could think about this in terms of maybe identifying objects and images, analyzing documents, moderating various kinds of content for quality or for emotional content, positive, negative, upset, angry, happy. And our customers, when they start to use A2I, they can actually choose three different categories of human reviewers. They can use the 500,000 global workers that make use of Mechanical Turk. There’s a set of third party organizations that have a base of pre-authorized workers, or organizations can make use of a private pool, if you will, of workers. Lots of flexibility to get people in the loop in different ways to give you a judgment call on how well your machine learning models are performing.
Daniel Newman: That’s excellent. People are always looking for ways to put that human element in. And I know you covered this back in December and it’s great to see these things becoming more and more of a reality. So Jeff, in pure AWS fashion, multiple announcements even in between ReInvent. And I was curious, is there anything else that you will be hitting at big AWS summit online?
Jeff Barr: So one recent announcement I would like to review with you is what we call the Snow Family of devices. The Snow Family devices basically are physical devices that have both storage and local compute power. They can run connected but they’re generally run either mobile or disconnected. We recently updated the Snowball Edge and we launched an improved version of what’s called the Snowball Edge storage optimized. There also happens to be a compute optimized, but we updated the storage optimized variant. The current one and the previous one both have 80 terabytes of storage inside. We did a hardware refresh, we added additional processing power, we added some additional SSD storage inside, where if you launch EC2 instances on the Snowball Edge those instances have access to this SSD powered storage.
We also made it easier for our customers to set up and manage the Snowball Family devices with something called AWS OpsHub, which is a graphical user interface where they can unlock, configure, copy data to and from the device via drag and drop even if they’re not connected to the Internet. So perfect if your Snow Family device is on a factory floor or in a mine or in a ship or something like that where you’ve got either [inaudible] or no connectivity whatsoever. That’s another piece of coolness I’m definitely really excited about.
Daniel Newman: Yeah, that catches my attention, Jeff. I’ve been big about the edge for a long time. That’s probably one of the big next frontiers. I think there’s a lot of punditry around it, but it just makes so much sense. And AWS has continued to stretch its reach from data center, obviously from the cloud itself to the data center and out to the edge and expanding the family, and building the tools, which sometimes I think people don’t talk about as much as they should. That the tools, like you mentioned, the GUI based tool, it’s simplification, that control play, making it easy for people to use, easy in the cloud, easy in the data center, easy at the edge. And it seems that there’s just a lot of that being considered with each of these new deployments.
So Jeff, I just first of all want to say thank you. We realize it’s a lot to be the chief evangelist. It’s a lot to try to know more than a little bit about a lot of things. And it’s clear you’re definitely getting the firehose over there at AWS. So want to definitely just stop here and say it’s been awesome having you here on the Six Five insiders podcast. And thanks so much for giving us and of course our awesome community out there an update on all the things going on in AWS.
Jeff Barr: Well, thank you so much. It’s been a pleasure to speak with both of you, and really enjoyed my time.
Daniel Newman: Yeah. So for everybody out there, hit that subscribe button, keep on tuning into more and more of these editions of the Six Five Insider Editions, and Pat and I when we do our regular weekly rundowns, deep analysis, no news, just analysis, keeping you in the loop of what’s going on in the tech industry. For this episode of the Six Five insiders, though, we’ve got to go. But thanks again, AWS, for being part of the event, part of this podcast, part of this conversation, good luck with the AWS summit. We will be tuning in. For now we’ve got to go. See you later. Bye-bye now.
Disclaimer: The Six Five Insiders Podcast is for information and entertainment purposes only. Over the course of this podcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we do not ask that you treat us as such.