Jeff Bezos at Startup School 08

so what the heck is this blue line this blue line is the amount of bandwidth over the last seven years that we've been consuming at Amazon in our retail websites globally and uh let me show you another line what's this red line this red line is the amount of bandwidth that we're consuming with AWS so it has now passed the amount of bandwidth that we're using in serving our retail customers which I think is kind of a dramatic indication uh of uh of of what's going on with AWS and and where it might be going in the future there are a bunch of services and I do think most of the people in this room know what they are they're all kind of infrastructure foundational Services I'll go through them very briefly because I think this crowd is uh pretty knowledgeable about these things our storage service basically give us uh large data objects or small data objects and we'll store them in the cloud for you the elastic compute Cloud which is a scalable we call it elastic because it's scalable up and scalable down uh you can buy compute in the cloud for 10 cents per server hour simple DB you can index and query data in real time simple Q service allows you to do interprocess Communications even on different computers in the cloud flexible p payment service is the first payment service designed explicitly for developers and there's a great deal of uh there's a rich feature set that allows you to build billing systems or any kind of system that you want to be able to collect money from your customers or to do peer-to-peer money transfers uh if your customer if your application requires that and then finally Mechanical Turk which allows bite-sized chunks of work to be distributed to workers globally Mechanical Turk is artificial artificial intelligence you can incorporate these little uh code Snippets in your even inside inter loops and they go out and they get serviced by real human beings all of these uh web services these infrastructural web services are built along a certain set of very simple principles that AWS ad hears to the first is that we want them all to be really easy to use so the whole idea behind web service is is to let companies focus on the parts of their business that's really important and differentiated about their business and we don't want the services to be hard to learn so the apis are designed to be simple we've worked hard in the documentation low latency uh just because you're doing it in the cloud we don't want it to be slower in fact in many cases depending on where your customers are it can even be faster if Amazon's data centers have better connectivity to the internet than your servers might in your own data center elastic scalable up and scalable down highly available very important and pay by the drink these are variable cost Services there's no contractual commitment you don't have to enter into a year-long agreement or a multi-year agreement you don't have to talk to a salesperson uh you come and use them as you see fit and stop using them as you see fit there's no upfront charges we uh going to announce a new number that I think is you know very significant we're very proud of in October of 2007 we had 10 billion objects in S3 and now we have 18 billion objects in S3 uh we had 240,000 registered AWS developers a year ago and I'm announcing today that we have 370,000 as of the end of q1 I loved this quote Amazon a Munday announced persistent storage for its ec2 service and what's notable is how quickly the eilers running ahead of the competition in fact Amazon's real business down the line will be its cloud services Amazon will be like a bookstore that sells cocaine out the back door books will just be a front to Sol storage and cloud computing so why are people excited why are people enthusiastic uh about these services and I think I uh I I I have a hint of what I think I know why people are excited about this and one way to look at the future of something is to find an analog from the past does anybody know what this is it might not be immediately obvious this is an electric power generator um an old one I was in Luxembourg uh recently and I took a tour of a 300y old Brewery their business of course is making beer and about 100 years ago they were one of the first uh factories in Luxembourg to start using electric power to make to help make beer to help in the manufacturing process and of course they couldn't buy the electric power off of the electric grid because there wasn't an electric grid so they started making their own electric power and a lot of companies did the did did that in that era if you could make your operations more efficient or do new things with electric power the only way to get power was to set up your own generator and become an expert in electric power generation the important thing to notice here is that the fact that they generated their own electric power did not make their beer taste better and what startup companies want actually companies of all sizes is they want to go from their idea or their Vision to a successful product as quickly as possible the problem always is that there's a lot of undifferentiated heavy lifting that gets in the middle between your idea vision and that successful product the undifferentiated heavy lifting by the way has to be done at World Class levels of Excellence or your vision will fail but it's totally undifferentiated and isn't actually making the beer taste better in the case of data centers uh which I think are a pretty obvious analog for electric power generation of a 100 years ago um this price of admission the kinds of things you have to do at a worldclass level are very very hard uh it's very complex to build a set build infrastructure build scalable infrastructure build sensible infrastructure it's always changing operating system versions always have to be upgraded uh you end up with heterogeneous environments inside your data centors which further complicates uh as you get any scale at all you just have to start doing some forecasting for capacity planning how many new servers am I going to need 3 months from now how many new servers am I going to need 6 months from now as soon as in that world uh things get very complex indeed and of course you don't get to go through this undifferentiated heavy lifting just once the successful companies are the ones who iterate as quickly as possible they deploy uh their Vision in in the form of a product and then they have to change it iterate and through each one of those iterations there often are infrastructural pieces that have to be upgraded or changed or scaled and this can really slow down the pace of invention and Innovation I'm give you a few few examples of companies that are using these services today uh and uh you know they're basically driving this Innovation Loop this success Loop faster because of these Services the New York Times took their uh entire Archive of Articles going back to 1851 was 4 terabytes of data they had it in a a format that wasn't it wasn't a format that they could easily use they wanted to convert all four terabytes into PDFs as a first step and uh it's a quite a computationally intensive process with much data 11 million different articles and uh it wouldn't have made any sense at all for the New York Times to set up a fleet of servers to do this task because it's a one-time task once the task is over they would want to undeploy that Fleet of servers very easy thing to do with the elastic compute Cloud um once they had converted all 11 million articles to PDFs then they made them accessible through this archive website uh using Amazon S3 so all the articles are stored in S3 and S3 serves these articles uh you do a search there's an interesting search you guys can try on the New York Times archive website search for computer and uh you'll find this um this is a May 2nd 1892 reference to a computer and uh as most of you in the audience probably know in that time frame a computer meant something quite different from what it means today a civil service examination will be held May 18th in Washington and if necessary in other cities to secure eligibles for the position of computer in the National Alman Almanac office where two vacancies exist one at $1,000 and the other at400 the examination will include the subject of algebra geometry trigonometry and astronomy SanDisk is using S3 in a very Innovative application here where they have the sandis cruiser uh it's a USB drive that also anything you put on the USB drive is automatically backed up onto Amazon S3 so even if you lose your uh USB drive you still have the data uh here's one that's happening in real time just over the last few days it's a a fantastic example of Amazon web service company called Animoto and uh this what basically what Animoto does is make it really easy for people to create videos with their own photos and their own music or the there's also royalty-free music that Animoto has on the site and the way it works is the they have some uh some Secret Sauce in their software that listens to the Cadence of the music and it basically kind of autoedits the photos and aligns them with the music so that it looks good um and uh you know the the the they they cut from one photo to the next at the right point in the rhythm of the music and so on if you ever hand edited something it's actually quite difficult to do and so after it's mixed the video for you with your photos if you don't like that particular mix you just hit a button and it does it again um and so you can kind of just you know iterate through a few iterations until you find a mix that you really like and it makes this incredibly simple well the system is built on top of of Amazon web services in fact they they use many of our services the simple Q service they use S3 and they use dc2 I'm going to talk mostly about uc2 let me show you a uh a graph um this is and this this is all happening over the last three days this is the Animoto ec2 instance usage this is the number of server instances they're using let's see if I get this laser pointer to work so they're kind of going along here this is about 50 ec2 instances down here and uh then there this their their Facebook app kind of broke through and so this is their Facebook app taking off this is just three days ago this is April 16 I'm sorry yeah April 16th um and you can see that they went they've gone from 50 instances of ec2 usage up to 3500 instances of ec2 usage I would have been it's completely impractical in your own data center over the course of 3 Days To scale from 50 servers to 3500 servers don't try this at home um uh and and by the way you if the other alternative might be to you know raise enough Capital to deploy 3500 servers that's sort of equally insane uh um because it's just way too big of a gamble um you don't know whether you're going to get this app to take off in that way and you shouldn't be deploying that kind of capital um the other thing that's interesting to know here is see these big drops well this is the elastic part of the elastic compute Cloud because you can programmatically all this is controlled with apis so you write computer programs to deploy and undeploy ec2 instances um in the middle of the night or week periods of demand they can actually give back ec2 instances to the cloud and they're not paying for them once they give them back um and then we can try to resell them to somebody else who perhaps doesn't have such time sensitive demand uh com you know Graphics rendering scientific compute uh applications and so on there are bunch of applications that are more flexible and we can use those more flexible kinds of uh uh applications uh to smooth out the demand curves and make better utilization of the underlying physical Hardware so that's one of the kind of systemwide structural advantages of this kind of pooled computing but you really see you know from a startup company's point of view this is a very dramatic three-day period by the way it's still growing I think uh today uh they're peing at 5,000 instances so uh pretty cool uh case study here's from Brad Jefferson the founder and CEO of Animoto says before AWS we couldn't have launched Animoto in our professional video rendering platform at our current scale without massive capex and a lot of VC funding the viral spike in Animoto video Creations we experienced this week would have been disastrous without AWS and you don't you do face this issue uh whenever you have a startup company you want to be prepared for lightning to strike but you don't want because if you're not that's really that's that generates a big regret um if lightning strikes and you weren't ready for it um that's kind of hard to live with at the same time you don't want to prepare your physical infrastructure uh to kind of hubris levels either uh in the case that lightning doesn't strike so this this kind of helps with that uh tough situation I'm going to take questions and just before I do I uh just want to remind you you know we make electricity so that you don't have to uh and the electricity if you generate it on your own it doesn't make the beer taste any better uh so thank you very much and I hope this is uh I hope a lot of you out there that this uh kind this way of thinking about things uh will be instructive to your uh startup efforts it's very exciting you know in fact when we see something like Animoto um it's very inspiring for us because it is a big part of why we built these services to make just this sort of thing possible um and now if there yeah I see there are a few questions I'm happy to take them thank you start over here how the idea of the web services come about and for those of you who couldn't hear the first part of the question um Amazon used to be in the business of selling books how did the idea of Amazon web services come about um well uh about four years ago we started working on this we we launched our first web service about two years ago uh but we worked on it for two years before launching it and so we really started this business about four years ago and at that time uh we were having enough U issues internally at Amazon I think I should back up a little to reduce the feedback loop is that uh between the speakers and me but we we we we had enough internal users of infrastructure that we found we really wanted an abstraction layer between our network engineering and you know uh deployment all those kinds of very important pieces of infrastructure and data center management and our applications programmers and um things were getting complicated enough at our scale that we realized we were going to have to build we basically the we started out building these services or designing these services for ourselves so that we could use them internally and the more we thought about it and the more we liked this this the idea of these services and of this abstraction layer that we could abstract away the physical Hardware um we realized if this would be helpful to a web skill application like amazon.com that's all Amazon is it is you know it itself is a big web skill application if it would be helpful to us it would probably be helpful to others and with a little extra work we could externalize these services and turn them into hopefully a profit Center and meaningful business for Amazon and so that's what we did um we did not expect this level of traction this early um we're delighted by it and U and you know and we're we're uh determined to continue to uh earn trust with our developer customer set becomes a Third customer set for us we have consumers which is our you know our biggest customer set and the one we're most well known for over the last seven years we've developed another thriving customer set which are sellers who PE third party sellers who sell on our Amazon sites and now this is a new customer set for us developers and we're super excited about it great thank you I wanted to say it's a really elegant solution but there's some concerns I deal with a lot of developers what were the changes in architecture that the outage after um what was it um what's it um what's the holiday it was a Friday after that what did that make you change in the way you did things and what were the causes of that problem and also what are the concerns and how do you address issues where some developers feel like in the apps they develop if they structure it for your network it kind of limits the way they want to grow if they want to go to other top polities after you how do you have you seen a big issue with that with other customers as they grow on your platform I'm not sure I understood the second question say if I'm built on AWS for part of my server architecture they're guys like these guys who couldn't have done any other way but say they have explosive growth and then they want to build their own platform all like yours are there issues with them coming off of that and being able to build their own platform as they grow beyond what you may offer them in speed of response because that yeah I I think the answer to the second question is there are there's very low friction for people who want to switch off of our platform you know we don't have um we don't ask for long-term contract you don't have to talk to a salesperson since you're paying since a variable cost basis you know you're paying by the drink if we ever stop earning your business you can leave today um and you know we just stop charging you um so there are uh and most of these apis are also pretty simple and structured in such a way that there's not you know a lot of switching cost on the API side you know if you look at something like S3 the apis that you need to you know store and retrieve data are very simple set of apis that you could map onto an alternative service so our view on this is that um this is going to be a business with you know multiple winners I don't think it's a winner take all business um for us we're really focused on making these Services as low latency as high availability as simple to use as possible and as low cost as possible and that's actually one of the places where I think think our corporate Heritage as a retailer which is you know a low margin business where you have to be extremely efficient in order to make money um and we do make money we we we do a good job in that retail business but it it is earned through even though it's very low margins it's earned through a lot of very careful U uh kind of frugality and focus on defect reduction cool so that is I think it's pretty easy easy for people to move around um and that's why I predict that there will be multiple winners and I think if we uh hopefully we can be the leader of those of that multitude of winners and and for us to be the leader given our position today that should be possible for us and it'll only be that we didn't execute well if we're not the leader you know five or 10 years from now it because we didn't execute well uh so we're going to work very hard on that no I was just asking though the thing the first one was what do you see about the changes in your in your structure way you the main change that we made and I think people have liked it um is we have try to improve the communication that we do when we have either an outage or a uh Brown out so there are U and you know the the we first goal first point is but we will never be satisfied until there are no outages we want 100% availability um and we're going to work on that we're going to work on it year after after year after year but what we've heard loud and clear from customers is that when there is an outage we want really you know detailed fine grain communication about what's going on um and we just launched a dashboard that'll make that a lot easier for people we're trying to communicate more frequently uh when we have any kind of situation like that cool thanks uh did you want to go are there just two sets of mics or should we just alternate then okay great um last week I was talking to a customer mine in Canada and their customers are hospitals and when I was explaining to them about uh Amazon web services they as something they should offer to hospitals they said they couldn't do so because of USA Patriot Act and so there one I know that the abstraction is very powerful but do you see yourself de abstracting across legal jurisdictions well this is actually one of there a whole bunch of regulatory issues around where data can reside yes uh it's certainly true I mean I yeah I don't even your particular question about the Patriot Act is probably way above my pay grade but but the you know I'm not going to be able to answer that one but but the um but there are all kinds of you know the European Union has customer data rules um uh any kind of uh Medical Data often you know can't cross National borders uh and so one of the things that we're doing with availability zones there are really two reasons that we launched availability zones with with ec2 one is so that you can have you can if you want a fault tolerant application fault tolerant against things like fires and floods you can deploy servers and be guaranteed that they're in different physical locations they're all in the USA though right today they are but that's not the plan okay and then what will happen over time is that availability zones will also be in different jurisdictions so that you can choose the jurisdiction that you want for your data thank you hi I have a friend who um they have a consultancy and they run all of their rails apps um through AWS and he tells me that uh apparently spammers have ruined all of the IP addresses uh and so they can't actually send email successfully using AWS can you comment on what you're doing to fix that and what what steps you're taking to to keep the Well from poisoning I can't it's a it's an interesting question and I and I don't know the answer and you know I can go back and look at it I don't know if anybody anybody know the answer do you know the answer part of the answer is another feature we launched recently was static IP address for dynamic uh Computing environments which we call elastic IPS and so uh using that uh you could basically you know have have a an end point that is yours you can hold on to a maposs multiple e two instances and using that you can much more ready communicate with your network providers and tell them that that you need are not a the point is that the outside world would see the elastic IP address as the same IP address so it wouldn't be wouldn't appear to be hopping around is that the thanks thank you go ahead or over there yeah um the ec2s elastic capability is fantastic if you need it um but for people whose hosting needs are more steady and stable and particularly if those needs involve something equivalent to your large or extra large instances um then right now ec2 pricing is not really competitive with traditional data center hosting um I was wondering if a you agree with that observation and B if so do you have any plans to go after that market well I think I mean actually I think it depends probably on your application and your data set I mean we we have gotten tons of feedback on the services um one of the most but I would say most of the feedback we've gotten on pricing has been this is really inexpensive thank you um and so I do agree that there are some uh there might be some use cases where traditional hosting is very cost competitive I you know I I can't guarantee you that that's not the case um but uh and we what I guess the only way I can really answer your question is to say that directionally our goal and this has been true for it's kind of a cultural trait at Amazon our focus is going to be to be the lowcost provider of these kinds of highquality services so as we are able to get smarter as we're able to uh you know uh figure out how to do things more efficiently we are going to be returning those cost efficiencies to customers in the form of lower prices on things like ec2 um there are other things that you know that it may be that we need some instance sizes is instance sizes is sizes that are smaller uh than what we have today uh uh so the smallest instance we have today is 10 cents per compute hour um and we're also looking you know always at uh you know returning some of the cost savings we get on things like bandwidth charges uh dis space charges and so on and so on the good news is that those things do get cheaper over time uh and as we get more efficient and organized we'll be able to return I think quite a bit of savings to customers okay thank you yes hi sir uh Mr Bezos Jeffrey sorry I'm a little nervous uh um first of all I want to say I'm a big fan and uh I wanted to thank you um for Gmail and also uh I wanted to know what it was like to be able to launch uh like rocket ships into space I've always wanted to do that since I was a kid so um well uh that's well thank you for that question um face is a passion of mine you know you don't choose your passions your passions Choose You And since I was 5 years old I've been uh obsessive about space travel and want you know wanting to be an astronaut and to go into space and and uh I have started a company called Blue origin that is building a vertical takeoff vertical Landing rocket it's a suborbital rocket um so you'll get up into space and then come back down um it's the the rocket is reusable and it re enters and lands on its tail like a Buck Rogers rocket under rocket power uh so that is and we've we've we finished our first development vehicle and we flew that and we're working on our second development vehicle now and there'll be at least one more development vehicle after that uh and then you know hopefully at some point we'll enter uh commercial operations so it is pretty exciting I can tell you also um just to kind of tie the two subjects together blue origin is a uh is a consumer of Amazon web services and um I'm not making this up and uh they do uh aerodynamic simulations they they do computational fluid Dynamic calculations on the ec2 cloud they used to have a bolf cluster it was a 16 node bolf cluster that they would do cftd calculations on and the uh uh but you know it's not their bit it's kind of like the brewery and electricity you know whether they are really great at managing huge bolf clusters is not going to make the rocket perform any better um and so and now they're able to do these these these aerodynamic calculations are very they massively computationally intensive uh very parallelizable calculations and so now they can deploy you know very large ec2 fleets but just for a few hours do the computations that they need and what this has done is sped up a cycle it used to be on the bolf cluster like to do one aerodynamic cycle would take I think like 70 some odd hours and now they're down to where they can do one of those cycles and I think it's less than 12 hours which means you know you can like the next day you can see the results and Tinker with the vehicle and do some more thing and then and then iterate more rapidly so um it's uh it's pretty cool thank you what do you think of Google app engine well I anticipated that question and uh we have a lot first of all let me say we have a long history that that I like of not talking about other companies so we make exceptions to it occasionally uh but it's a pretty good policy in general and and the main reason is that what we have found in our business is that we are more Innovative and do more interesting things if we stay customer focused instead of competitor focused so it doesn't mean we don't pay attention to other companies and try to be inspired by them and see what they do and so on but really you know if we had been a competitor focused company instead of a customer focused company we wouldn't have started working on Amazon web services four years ago the only way you get to start working on something for like this four years ago is if you're really kind of paying attention to what customers might want and need uh so that's a very good uh thing so let me talk only about AWS rather than about U you know anybody else's offering AWS is pretty unique offering at this point in time in the whole Marketplace um and the reason is that it's really based on very foundational a very Foundation it's deep in the stack um the kinds of uh apis that we're exposed are very fundamental things like you know storage that's remembering things very fundamental compute queuing and then you can stitch those things together in fact several companies have written um different kinds of application engines that run on top of ec2 so it's such a foundational layer that you can build a number of very sophisticated things on top of it and I think that's really you know sort of our angle at this um is to build something that has that dramatic scalability and flexibility and really give people direct control over the knobs that they need to build sophisticated applications with whatever tool set they need to be using um so I guess I just leave it there the only other thing I would add is this is not going to be a winner take all space I believe this is going to be um an area where there'll be lots of winners uh I think and you know maybe I'm very optimistic but I think this is going to be a meaningful industry one day I really do I think it's going to be a significant industry and uh because just because I don't think it makes sense for everybody to be generating their own electric power um and so and it's very rare for significant meaningful Industries to be built by single companies so that's why I think they're actually going to be a multitude of of companies pursuing slightly different strategies flavors and I think developers are going to find many of them useful um and I just hope that you know they continue to find Amazon web services useful and I think if we continue to work hard at it and and earn that uh that they will thank you hi AWS is a shared infrastructure and uh especially in the case of ec2 it's a shared hosting infrastructure and um as you just pointed out ultimately Amazon can't control predict when one of their AWS customers could go viral um so the concern with shared hosting is always you know AWS is a finite resource at the end of the day so if two of your customers go viral or 10 of your customers go viral at some point you're going to hit the limit and you know new instances aren't going to be able to be launched or you're going to have capacity limitations what can what can you share about how how you guys are going to manage that in the future when these events do happen you know I found it very interesting the chart you showed about Animoto and the success they've they've had going viral over the last few days uh a real world example is for instance last night I was unable to launch even a single small ec2 instance for for about an a period of an hour time I got insufficient capacity errors errors back so so how can you address this how can you give some assurances that people hosting in this environment it's actually going to be there for them as they want to scale and and these sorts of things for for factors that are ultimately beyond your ability to control control well I think we can't control anything that's beyond our ability to control right I want to get that straight right away how do you manage the risk but yeah but I but um the way to think about this is first of all it's better than the alternative because since there is a pool of servers uh you you're much more and the pool gets large then the fact of the matter is that somebody who needs when the pool is large somebody who needs 3,000 servers overnight right that's still a relatively small part of the pool and so you know the alternative would be you know I have 50 servers in my own Data Center and I need 3,000 overnight so at least on ec2 you have a very good chance of getting that serviced the alternative is you have no chance of getting that service so that's one answer to your question the second thing is um over we have to get we have to be and we're pretty good at um trying to maintain the right amount of excess capacity and we get to look at much smoother uh you know demand increases over time because it's the average of hundreds of thousands of AWS developers so if you're looking at any one AWS developer yeah trying to predict their usage is going to be impossible it would be like an insurance company who decides to provide health insurance to one person right they have no idea is that person going to get some dramatic illness that's going to cost them a million dollars in hospital bills or is that person going to be healthy for their entire life and then be killed in a car accident instantly right those are the two extremes for an insurance company it doesn't make any sense to insure a single individual but to ensure a big pool of people makes a lot more sense and so we have that structural advantage in terms of predicting uh our capacity and then probably the final answer I give you to that question is you know the onus is on us over time to earn earn credibility in that regard you know we need to be uh we need to earn that people say you know yeah I I realized Amazon can't um you know it's not magic and they can't materialize servers out of nowhere but you know what they do a pretty good job of maintaining enough capacity to serve me even when I grow by two orders of magnitude overnight so just as a quick followup is there some commitment Amazon's willing to make about the level of overc capacity or you know is there is there some way you can communicate some some sorts of assurances about what are typical shared infrastructure I don't know I mean it's an interesting it's an interesting question um and maybe one we should go away and think about because there might be a subset of customers um you know who'd be interested in that kind of insurance thank you thank you okay so my question is how is Amazon web Serv web services or could they help large ENT PR companies I'm talking Fortune 500 companies reduce their infrastructure costs for the deployment of software so for example if you've ever looked at a sap uh integration guide just a table of confidence is like 20 pages and all they have to do is spend years going out getting the hardware the storage the redundancy they're basically building their own electricity to do that these companies to integrate software how do you see Amazon web services playing into that I.E maybe you know sap offers and Ami with some middleware and boom not a year of deployment but two seconds of deploy well we are we're actually we are starting to see that right now so in fact some of the um I was talking earlier about uh computational fluid Dynamic calculations so on there are a bunch of scientific uh Computing well basically proprietary people who bundle up this you know CAD cam software and analytical software and who offer that uh and they license it it's licensed software it's expensive software yeah um they are are realizing that there's a real opportunity for them to bundle up an Ami with these big you know they all have sort of backend calculation engines that then attached to these front ends so that people can have their own bolf clusters or things like that that they can bundle up Amazon machine instances and let people have very big uh uh parallelized computations on these things and whether that would also make sense for you know some of these CRM companies perhaps it would yeah but I think that's the kind of thing that you can see developing over time I think it's an insightful observation and just quick follow up to that is do you think there's enough security behind Amazon because you often hear you know an open source company or that has software the company buying it wants to deploy it on their own server behind their own firewall with us if we offered that with our software do you think we could make a good enough guarantee value proposition those companies to trust us to have on Amazon web services do you think over time it'll be there I I think so um I think you know what what we would expect is that people would want to understand how we're doing it and and and be able to audit that to a certain degree I think that the other thing is that with some application changes which are never free if you wanted to be especially careful you could encrypt the data before you stored it um and then you know you'd have to have a Gateway some a trusted Gateway somewhere to do to do that but that's I think a practical solution for highly sensitive if people are really concerned they can always encrypt okay thank you and free plug if anybody wants to help build that find me last question okay hi um basically my question's not uh regarding AWS but more like your other services that you recently got into is especially the music business and the video business and um since you know a lot of people have been a lot of us have been seeing the mobile platform you know spark them the last years with the iPhone platform and the Android uh how does how do you guys see yourself integrating your new Services into the mobile platform especially with the new abilities very enhanced compared to the older devices absolutely and you can see that even in our usage even people just web browsing on some of the new mobile devices like the iPhone new versions of blackberries and so on these more sophisticated generation of devices we're seeing a lot of retail sales coming through those devices now in a way that we hadn't seen previously except really in Japan where they've been a little bit ahead of the United States uh in that regard so um I think you are going to see uh more and more Amazon offerings uh deployed in on those devices I you know I'll have to just ask you to stay tuned I can't share with you to the details but I think it makes a lot of sense uh for those for our digital product offerings in particular to be deployed on a a wide array of devices devices and it also makes sense for us to facilitate mobile shopping in every way that we can so thank you and I'll just have to ask I have a question with a one sentence answer about sdb when do you expect to lift the 100 domain limit on simple DB when do we when do we lift the 100 domain limit on simple DB we hav an announced we're always try to be appropriately conservative when we come out with new products and make sure that they really really work for exting customers we've let a lot of developers in we're working on a lot of new features as well as expanding that as rapidly as possible restur we're very highly motivated to ungate Service as quickly as possible but you know will we will only do so as we really know it's going to perform Excellence over time and if and if since he didn't really answer your question if if um if somebody wanted you know to get an exception to it could they call you and talk about it so maybe that a way to get started because with with with these Services we take availability um and so seriously that when we launch new ones okay we do like to go step by step but that's Adam slipski all right and you should contact him I will and I thank you you guys

Channel: startupschool

Views: 109,386

Keywords: jeff bezos, amazon.com, startup school

Id: 6nKfFHuouzA

Length: 43min 19sec (2599 seconds)

Published: Sat Jan 31 2009