API Economist: Cloud adoption is clearly accelerating in the enterprise. What are some of the biggest challenges that you're seeing with the typical enterprise wanting to move to the cloud?
Jared Wray: We see really two major problems that come up in the enterprise. One problem is, how are they are going to be able to move their resources and actually take advantage of the elasticity of cloud or even the services that come along with cloud? The second problem is, how do they architect legacy applications or even mold them into a better, fully distributed type of system? Those two problems are common in enterprise, and really there haven’t been great solutions to this day.
API Economist: Most enterprise web apps are very desktop or laptop centric. How do you see this changing in the enterprise with cloud computing?
Jared Wray: If it’s a new application, enterprises are definitely looking at mobile first. Bring your own device (BYOD) has turned mobile devices into a number one priority and laptops are becoming second priority out of the system. Tablet adoption is making a big difference. This is something that's really changed. This really draws on the next generation of APIs. It's not just, “How do I provide a great REST API?” It's “How do I provide a great mobile API?’ The data has to be a lot better; it has to be thin. It has to be able to make sure that it works inside a mobile device where the carrier could have a lower latency. This is something that we hear from the enterprises constantly. I was just talking to a large company that’s doing everything mobile now. That's their top priority.
API Economist: You mentioned APIs. What’s your sense on the state of APIs?
Jared Wray: APIs are a mess today. When you look at what companies are doing—even APIs or the REST standard—you really start seeing a lot of gaps. What we're seeing in this space is that the fundamentals of API life cycle management aren't even being built. Even the core fundamentals of an API, making sure it's secure and cacheable, aren’t always addressed. Other considerations include how to make sure that you can have a great identity management platform as well as application tiering. And let’s not forget good documentation. Many of the APIs today aren't even doing that. That's a real good problem for the industry in general, as APIs are flourishing at a huge rate.
This really gets into what we talked about before around mobile. How are you going to be able to document that? How are you able to have a mobile developer design to it, and get it up and running from core services? These web services are going to struggle because of that. I really think that there's going to have to be a framework you can easily install to give you a lot of those core fundamentals and manage that API life cycle even better than what a service does today.
API Economist: When you started Tier 3, what initial problem set were you guys looking at tackling?
Jared Wray: I started Tier 3 back in 2006. This was the new dawning of what cloud was becoming. Amazon’s EC2 had just started getting into the next generation of cloud computing. Our whole goal was then, and still is to this day, how do we enable the enterprise to get rid of 70 percent of the mundane tasks that they do today? What that means for them is many of the database administrators and IT operations people perform a lot of manual tasks, from backups to monitoring to being able to provision servers to even getting servers or troubleshooting them. Those tasks should be eliminated. That's what cloud really is about.
When you look at how to manage those resources and where you can scale them well, this is something that they struggle with. We wanted to be able to give them a management platform in a public cloud that could not only give them security, and the SLAs that they want, but also get rid of the mundane tasks that they're doing manually today.
API Economist: A lot of enterprises adopted the public cloud primarily by just taking existing workloads and apps, packaging them up as a virtual hard disk, and then deploying them. What about native cloud apps? Adoption here is not moving as fast.
Jared Wray: This is one of the biggest struggles that I've seen with the enterprises: the whole idea of building a distributed system. Developers today just aren't ready to do that, or haven't been trained on how to build a highly distributed system. When we built Tier 3, we started with the concept of always having multiple data centers. That really changes the game in the enterprise. Most applications are built for a single data center or even just a couple of servers.
When you come into the next era, which is this cloud era, you really have to think of being highly distributed. You have to look at how your database is going to be architected and how it's going to be fault tolerant. That's going to give you a lot of problems.
This is something that we've seen with our customers who are now moving outside of their comfort zone and into things like NoSQL, or looking at being able to do VMs [Virtual Machines] that are better for load balancing, or even throwing VMs away or scaling them at a certain time. Those are big things that they haven't even looked at before to keep things up and running.
API Economist: We're hearing a lot about the idea of the “Internet of Things” or the industrial Internet. How do you see that impacting how enterprises look at both cloud computing and APIs and the need to manage machine-to-machine interactions?
Jared Wray: We need better baselines of what needs to interact and how it needs to interact together. Machines being able to talk machine-to-machine makes complete sense. I'm continually for it in every way. What I think is very interesting is somehow we're going to have to be able to manage it in an ecosystem. Today nobody has really figured out how to do that well. This gets into how are we going to enable people to develop not only for the industrial age or the industrial service layer, but also being able to augment it with cloud services in a really easy, contractual way.
That still has not been figured out. There are a lot of one-offs. We need at least some fundamental standards to be able to move forward. That's what cloud did extremely well. We all understood what a VM was and how to get it up and running. We even understand how to deploy application code to a cloud platform. Now we’re getting into the industrial side of the house, and there's a lot of differentiation. That's going to be a big problem in the long term.
API Economist: As application marketplaces continue to gain adoption, who's going to manage that service level agreement (SLA) for the customer?
Jared Wray: When I look at the current marketplace today, I think it's going to be a very big problem that we have to start doing better integration. We talk about how to provision services or even manage services from third-party service providers. Now you're getting this downstream effect of, "OK, we can do that but what's the next step? How are we going to integrate better so that either the service provider that initially has the customer and is running the application can own that SLA or is the customer OK with that?"
I think one of the biggest issues that we're going to have is really deciding, "Is it worth it to go down the path of owning the SLA?" Many customers today are starting to understand the idea that when you buy into a marketplace, everybody has a different SLA for each service.
Now, that can be great, but really, is the enterprise going to adopt it? The enterprise today struggles with this concept, which means that we have to either do deeper integration with the third-party service providers, or be able to provide an overarching SLA, which then causes more grief for the service provider running the application.
API Economist: Do you think that a lot of the existing cloud platforms today are repeating the sins of the past, where they're basically creating roach motels where they can get in but they can't get out?
Jared Wray: With current platforms today there are two sides to this issue. One is, how do you drive innovation? I'm a huge proponent of innovation and being able to make things happen. I've been very against even declining APIs or base level APIs for that fact. When it comes to a roach motel and being able to lock in people, it's going to happen. Innovation is going at a very fast pace. You're going to see proprietary APIs and proprietary services being built and customers integrating with that and causing that roach motel scenario.
But at the same time, I think customers really need to be aware of that. If they're building services and they're buying into that, they need to understand how to proxy those services correctly so that they're not stuck on those providers forever. That's going to be up to the customer until tougher standards emerge.
API Economist: Do you see a cottage industry developing around inter-cloud integration?
Jared Wray: Absolutely. Or even the other side, which is service providers who come to market and really try to standardize. This gets into even what type of VMs you're running or how you're running them or what type of services you're doing. There are many companies today that offer very proprietary APIs, and developers are programming against those. There are also ones who will spin up what's common in the industry like RouterMQ or ZeroMQ for a messaging bus, instead of providing a whole proprietary messaging bus on top of it.
API Economist: What are your favorite mobile devices and some of you favorite apps?
Jared Wray: I'm going to include my laptop on this, just because it's something that people always forget about. I have a MacBook Air and I love it. It's one of the best systems out there. I love it because of the flexibility that it has and that I can do not only Windows development but also Mac development on it. I'm really able to do both extremely well. It's just great hardware overall.
On the other side of the device, I love my Mac Mini. One of the apps that I love the most is TweetBox. It's just an amazing, well-done Twitter system.
The other device that I also use is Kindle. Being able to have the same books on my laptop on my desktop and even on my mobile device and my iPad Mini is great. On the flip side, I also didn't like the iPad because it was too heavy. The iPad Mini is perfect.
I also use an iPhone. I'm not completely sold on the iPhone. It's great and it's good enough. It hasn't made me want to go somewhere else, but overall it's good enough. Some of my favorite apps include Zendesk and the mobile apps from the airlines, for example Delta and Alaska. And of course I love the lifestyle apps like Foursquare, Yelp, and Urbanspoon.
API Economist: Jared, thank you for your time!
Jared Wray: Thanks for having me!