I'm Andy Glover, and this is the Java Technical Series of the developerworks podcast. My guest this time is Judah Johns. He is the full-time evangelist for Jelastic. Welcome this fine afternoon, Judah. Thanks it's good to be chatting with you. So tell us. What is Jelastic? We are a PaaS, a Platform as a Service. Started off as just for Java, although we're expanding that, but our focus is Java. And so there's certainly more Platform as a Service focusing in on Java. Tell me about this market. Clearly there's a need, what differentiates let's say Jelastic from A, B, C or D. Yes, there's a big market updated the stats from Oracle, there's something like 10 million Java developers out there. Ten million, wow. Yes, 10 million. So you're talking about freelancers, guy who do it full-time for their work, enterprise guys, SMBs, small shops. So there is a demand and obviously we weren't the first ones to realize that. First you had I think it was GAE, Google App Engine. And then you had Amazon come in with their Elastic Beanstalk. You had Heroku. We were working on a different project.
In the process of building our startup, we realized that none of the solutions out there actually worked for us when it came to vertical scaling. And then we realized actually vertical scaling was kind of difficult to do, especially without code changes. And that's when we pivoted and that's where Jelastic came from. That was about two years ago. So I'm curious. You say vertical scaling, and then there's also the term horizontal. What does that mean for Jelastic? What does vertical scaling mean? Basically there's the one adding more horsepower to whatever unit you're using. And then you have the other, where it's basically adding more instances that way. So the first one is the one that you have issues with. Most people can...they can add you more of the same, essentially kind of cloning. But adding resources like that, especially as you start talking about the different software stacks out there, different databases, that's quite difficult to do. We re the first to do that. Something we're pretty proud of. So tell me about the Jelastic platform then. You brought up Google App Engine. We ve certainly talked with the good folks over at Google their platform. And as listeners might know app engine is very specific. So if you
code something to app engine, you end up building a very specific, let's say, app engine app. Tell us, so if I...how do I code for Jelastic? So that is one of the issues that we had. So we actually were trying out the different options out there and we realized if we wanted to go from Amazon to Google, we had to recode our app. And that was a lot of work. One is the vertical scaling and two is the recoding. Neither of those really worked for us because we wanted something that was standards based, using standard software stacks like Tomcat or GlassFish or Jetty, whatever anyone wants, with whatever database or balancer. So that is what the difference was for us. Eliminating the code change that comes from switching providers and the lock-in. It was part of our architecture eliminating those time wasters as we saw them. So then you did mention stacks. Is there a particular stack that you all support? Can we draw this back to the whole 10 million Java developers out there? So I'm presuming that whatever stack you support there's obviously a high percentage of the 10 million people in the Java development world are using these particular things. Yes. So for us obviously there was Tomcat. When you're learning Java you typically start with Tomcat. A lot of books start
with Tomcat. We realized it wasn't the only thing out there. There's a huge GlassFish community, a lot of people really like Jetty. We realized if we wanted to offer these, we needed to make it easy, just as we wanted it for ours. So we have Tomcat 6 and 7 and we have GlassFish and Jetty. So those are the app servers that we have that make the first part of the software stack. Now, what about a back end? So if you're talking about databases we have MySQL, we have CouchDB, MongoDB, MariaDB, and a few others. We realized as we started that these are big communities out there and there's a big percentage of people out there who are. In fact, I think we did a...we regularly put like our market share within Jelastic and 56 percent of our users use MySQL, 15 percent use MariaDB, MongoDB is at 16 percent. These are things that we would have never known unless we had made them available. And it's pretty cool to note that actually MySQL's stayed pretty much at over 50 percent the majority of the time. It s interesting to watch that. So I'm curious also early you mentioned Elastic Beanstalk. And I haven't looked at Elastic Beanstalk probably in a few months. But the last time I played with it, you were deploying I
believe to Tomcat. It was WAR file based. And from a back end standpoint you didn't have a lot of options as to what they offered you, you either interface with RDS or SimpleDB or you spun up a database on some other AMI and you talked that way For the folks that have looked at Elastic Beanstalk, what are the big differences between Jelastic? I think one of the biggest differences is it seems like Amazon's kind of quit developing it. They're not providing updates as they used to. Their CEO said not that long ago that that wasn't going to be their focus anymore. And something we've kind of seen with Google App Engine as well. Not sure why, to be honest. But Amazon definitely has some big plusses. It has one of the world's best back en hardware. The infrastructure is amazing. Although we've seen a few issues in the US recently. It took down other PaaS providers like Heroku and dotcloud. So there's definitely some things there that Amazon has going for it. But I think that the downside is that you're very limited. And then the things that we see as time-wasters, like having to set up app servers, having to set up database servers, tying them together on the back end, when you want to do things like scaling, you have to do it yourself. You have to go back in and be a sys admin. As a Java developer, I see that as a waste of time. I don't want to mess with that. I just want to worry about application
functionality. And that gets in the way for me. You mentioned that Amazon had some outages, and I think everybody in the entire world now knows it because when Amazon goes down you can't get into Netflix and whatnot, it takes down half the Web. You also mention Heroku. When Amazon went down, it also affected Heroku. And the obvious correlation there is Heroku is running on top of Amazon, right? Is Jelastic also running on top of Amazon? Or are you using other providers? How do you help avoid the issue ; if Amazon goes down does my app then also go down? Yes. Just a few days ago I actual got to talk with James Gosling, the father of Java. He s actual running his current project, some of it, on Jelastic. His stuff over at Liquid Robotics with the ocean going robots, mapping the oceans He said his biggest thing, the thing that he focuses on is reliability because he works with very important data and though there is not a lot of it, if you go down just a short amount of time gathering data, it makes a big difference. So what we did that was different than every other provider out there, actually, is instead of deciding that we're going to host it ourselves, we realized the back end running a data center is actually very time consuming and difficult. It's not easy. And there's a reason that PaaPs do that and just that.
So what we decided to do is we would partner with reputable hosting service providers, and we could then scale our platform internationally that much faster. So we launched in the US with a partner there Servint. We launched in Russia with our partner the Rusonyx. Then Japan, the U.K.; we have Brazil coming online. We realized that that back end wasn t our strong suit and that we should focus on what we're good at, which is our platform. So then we partner with the people who are good at that back end and who have higher up times than Amazon. In fact our partner in the U.S., Servint, has the highest uptime of any host service provider in the world. That s something that we realized early on. And we ve a lot of new things like offering our platform in Japanese, we didn't initially think it, but this made us different from, for example, Heroku. When you are with Heroku, you can only deploy in the U.S. If you're with dotcloud, you only deploy in the U.S. What if you're a U.K. based company that is working with customer data and that data demands that you keep it hosted locally, in the U.K.? Well, you couldn't do that. You bring up an interesting point; I think this is worth exploring the data locality issue and data privacy. Give us some more details there. And for listeners, Judah is in London, so obviously he has a lot of experience and knowledge of European privacy laws. And plus, Jelastic is an international company based on the folks
that I know that are working with you Please explain some of those issues in detail? For example is this a law in the UK that you have to have data physical servers that are also on. It depends on what kind of data you're talking about, right. Okay. For the most part it's preference. When you are working with industries like banks and oil and gas, they're old school. They want to be able to see where their data is. When you say cloud, they get scared. The first thing they ask is where it is really they want to know that their data, their data center, is somewhere they can go and actually see it. They want to go talk to the guys that are running it. There are countries, though, like Australia, for example, where it is mandatory to have things like government data, personnel files, those kinds of thing, have to stay within that country. So you run mainly into the preference. Okay. A developer in Japan, accessing his data in the U.S., is actually going to be kind of slow. So distance makes a huge difference. Running locally there's performance benefits and the preference of having a local provider that you know and trust that's
been there for a while. You run into in some countries where you actually have to do that. In the U.S., for example, you don't have to, but with the current situation we see that it's probably better if you're working with customer data to keep it local. And so when you go to the Jelastic platform you have the ability to run this in the US, or I'd like to run this in Japan. Do you have the ability to scale it? I know we use the term horizontally and vertically I want to say now intergalactically, I guess. Yes. That's actually something we're working on and that's something that actually James Gosling mentioned in the talk. He wants to be able to replicate his environment across data centers, across continents. And that's something we're actually working on actively for our next release. It's obviously a really big deal. There's a lot of work there that has to happen. Not just for our platform but for our partners. I think once it happens, people will be really excited about it. Because as of now, there is no way to do that, other than using a CDN. That doesn't work when you're talking about an application that needs to be running a database unless you in fact own the CDN provider. I see. So we've been talking about Jelastic, but you
brought up two interesting things, and I want to see if there's more. In general when you're talking about moving to the cloud and looking at cloud providers, I think you definitely have to do your reliability research. And another thing is data locality. Do you want to know where your data is, is there a preference to have your data in a data center where you can go knock on it and say hi? What else should Java developers be looking at? What differentiates the various providers out there? The cloud is...you just take that for granted now. That's the way things are going to be. So at some point we could probably just stop talking about cloud and just start talking about its benefits. I think as a Java developer, most Java developers... everyone's looking for Java developers. So most of these guys already have jobs. So then if they are working on a side project, they have very limited time to work on the side project, whatever it is. So if you're that guy that's working on a side project, you want to eliminate your time wasters. You want to eliminate things that create roadblocks to you deploying application. So you go in and look for things like no vendor lock-in, you want to look for things like not having to recode if you want to move, if you don't like the provider and going somewhere else.
You want to look into, how fast can you get up and running? So from the time that you create your account to when you deploy an application, how long does that take? Up until now, most providers like Amazon, Google, Microsoft, you're talking like hours, sometimes weeks. That doesn't work very well for someone who's coding an iphone app. You just don't have time for that. You want to be able to upload, deploy and enjoy as soon as you can. So that's very important. And then you look at things like price. Do you want to pay usage based pricing? For smarter people, though sometimes they're smart, they get lost and they're like oh, well, if I'm paying for usage, and there s a spike on me and I won't know what my bill price is. But usage pricing is typically the best way to go, because you only pay for what you use. You have the option you want to pay for like a monthly, a set amount, more like a dedicated server type. And, I want to be able to scale. I don't want to pay for what I'm not using. So those are definitely things to take into consideration. Of course you want to talk about location. If you're a Russian developer hosting your data in the U.K., that is not...it's better than the U.S. as far as speed, but it's not as good as being in Russia. If you are a Japanese developer is the platform in your language?
When you start talking about your end users, you're using your platform in Japanese, end user is Japanese, there's just small things to consider that in the long run actually become a pretty big deal. Sure. I'm curious because you did bring up pricing. And cloud computing in general, the pricing is at face value very inexpensive, right? If I fire up, let's say, a small instance, an AWS, Amazon, I'm paying pennies, right? There are so many providers out there, regardless of platform, there's plenty of let's say Node and Ruby and I've had this conversation now with a number of cloud provider, Because there are so many cloud providers out there the cloud in general becoming a commodity, and what differentiates different providers is data locality and reliability and support, et cetera. From the pricing standpoint, how is Jelastic priced? Is it ultimately cheaper to run my app on Jelastic? Am I paying a little bit more and here's why? We've done the comparisons for the most part. You're paying about the same as if you would use like AWS. I think someone like [INAUDIBLE] is slightly cheaper. And we have Google, where they just give it away. That model works great until you realize that for Google a failed business model has only 10 million users and they're going to close it. On pricing, for the most part, everyone's on the same page. Hosting, Web hosting and, in turn, cloud, commoditized, price went all the way to the floor pretty quick. People just give away hosting now. So where do you make your money?
It's on the things that you offer that someone else can't. Let's say it's the crazy awesome support like Rackspace talks about or, you know, the incredible API and back end that SoftLayer has. For us, it's the time saving bit and the fact that we're standards based. Some of the things we did a year and a half ago Heroku just announced yesterday. For example, having a pretty cool UI that will drop down menus. So these are things that the price really doesn't matter as much because it's all about the same. But in the end, what is it you're looking for? Do you still want to use a command line? All right. Cool. But the truth is you don't want to waste your time with it, if it takes you typing 15 commands to spool up an app server and a database server when you can just do like three clicks and click go. For me, the choice is easy. And I think that's where most developers are at. They're realizing I have my application, I'm worried about the functionality. I don't really have time to be a system administrator. If I want to save money and hire someone else, then I'm going to choose this platform or that platform. So you've talked about the time-saver and how easy it is to get started from Jelastic. I'm curious; can you get started for free? How does one get started?
Yes, so like all our providers, you know, all of our partners, we have five right now that have launched. When you sign up with us, you get a free two-week trial. And in that time, you'll know pretty quick. We have a standard hello world that you could deploy there that's pretty cool. When you realize, a) the hello world takes you a total of two minutes to deploy from the time that you create your account, I don't think anyone else does that. I love to show that to people at conferences: Let me show you how this works.... and... We're done. People like that they're like oh, cool. They have questions like how does it scale? What about custom domains? What if I want to work with the plate framework or something like that? That s extra stuff that doesn't take very long to do either. The getting started part, when you're talking about less than five minutes on the high end, that's nice. And you'll see that with the trial period Got it. Earlier you mentioned Tomcat and GlassFish and Jetty. Obviously, you know a core kind of background in Java and you mentioned Mongo and MySQL. When it comes to Java and the JVM, right, there's a whole lot of excitement around alternate JVM languages such as Scala, Clojure, Groovy. It's safe to say you guys support them across the gamut? Yes. If it's JVM based we run it. There s a lot of debate there about languages. I think it's proof of the fact that
Java I has staying power. It's not going anywhere. These are fantastic language you're talking about, like Scala. And they have their benefits. But the fact is that they're essentially Java derived I think speaks to what Java can do and the fact that it's going to be here for a long time. I love reading these stories out there that Java is dead, it's no longer cool, everyone's going to Ruby on Rails. But trying scaling Ruby on Rails. It doesn't work that well. That said, we realize that not everyone works with Java. Soon we ll be adding PHP. We re working towards that slowly so that we can make sure that all the benefits that we're trying to offer in our platform we can do them with every language we support. Got it. Where can we find out more information on Jelastic? It's jelastic.com, so it's elastic with a J at the front, dot-com. And I presume that there's a blog and whatnot I can follow and? Yes. Our blog gets a lot of reads. We did something like 25,000 reads yesterday. Clearly you guys are releasing something interesting. There's some interesting content there.
Yes. Every Friday we try to interview one of our users who has something cool. A few weeks back we did interview one of our users who made a game for his wife, a back end for a game that she plays. And in one week he realized he had 150,000 downloads. He didn't realize it until he went and looked at his account and his account had scaled up with it. I'm paying like 50 bucks a month. And I just made like $6,000 in the app store. So that really worked out well for him. Neat. So that's at Jelastic.com. Are you guys on Twitter? We are. @Jelastic. All right. Well, I know I speak for all our listeners when I say thank you very much for taking time out of your busy afternoon in London to chat with us about Jelastic. Yes, no, thank you. This was great. I'm Andy Glover, and this is the Java Technical Series of the developerworks podcast. Again, my guest was Judah Johns. He is a full-time socializer for Jelastic. So check out Jelastic at Jelastic.com, and thank you for listening. [END OF SEGMENT]