Orbitz paves the way to enterprise open-source contributions
Company is demonstrating you don't have to be a vendor to write great open-source code. It makes good business sense whether you're buying or selling software.
On Monday, Orbitz Worldwide plans to announce the creation and release of two open-source projects, Extremely Reusable Monitoring API (ERMA) and Graphite. Though there were hints of these projects at JavaOne earlier this year, Monday's announcement will add significant context to the work Orbitz has done to create two highly compelling open-source projects, whose applicability extends far beyond the travel industry.
On Friday, Orbitz gave me a preview of the announcement and the opportunity to talk with its sponsors, Winthrop Short, senior director of Orbitz Worldwide, and Matt O'Keefe, senior architect of Orbitz Worldwide. In talking with Winthrop and Matt, it's clear to me that Red Hat CEO Jim Whitehurst's vision for enterprise collaboration through open-source communities is going to be led by companies like Orbitz, companies for whom technology is not necessary drudgery but rather competitive advantage.
Consider the following: Orbitz employs 1,600 full-time employees and has another 500 contractors. So, 2,100 people total. Half of this total number is made up of technologists. As Brian Hoyt, Orbitz Worldwide's vice president of corporate communications and government affairs told me, "We have always been a technology company, one that just happens to be really good at selling travel."
But why open source? What benefits does Orbitz derive from open-sourcing these projects? Why not keep ERMA and Graphite to themselves?
ERMA and Graphite are "part of a Complex Event Processing system designed to monitor large distributed applications, analyze the data that is gathered and display that data in real-time graphs."
What does this mean? It means that these two projects, used in combination with Streambase, ensure that an error in the system (perhaps a customer gets an error when attempting to purchase a plane ticket) can immediately be identified by Orbitz, with the location and cause of the problem displayed in a single, easy-to-understand message that saves valuable time and makes it easier to fix the underlying problem.
This has obvious uses in the travel industry, but it's really much bigger than that, what with so much business migrating to the Web. The more complex the Web becomes and the businesses that use it as a communication medium, the more ERMA and Graphite become critical components in delivering customer satisfaction for a wide variety of businesses.
This does not, however, address why Orbitz would open-source the projects. I asked Winthrop and Matt to provide the background for the decision:
[Winthrop] From a company perspective, we've been around since 2001, and started small and scrappy. We needed to do more with less. As a result, we've spent a lot of time with open source. As we've grown, we've kept that emphasis on open source.
But open source isn't solely or even primarily a cost decision for us. If you think about the technical problem that we're trying to solve--a big, complicated "spaghetti-web" technology problem--there are certainly some areas where proprietary systems make sense. For example, we run a lot of Oracle. But there are many others where open source makes sense because it returns control to Orbitz, allowing us to tailor software to our particular needs.
[Matt] Open source gives me a level of comfort in that if I need support, I don't need to rely on a third party. I can do some searches on Google and, if it's a mainstream project, I can quickly discover an answer to my questions. I don't need to rely on a third-party vendor.
As an example, adopting the Spring Framework has been a huge advantage for us. We now have a code base that is easier to test and develop. It has also been a pleasure to work with because application monitoring is much easier with Spring as it provides a generous number of "hooks" for extension and customization. Spring support is top-notch. We bet big on Spring early on and have been able to help the project mature.
How does this affect your hiring?
[Winthrop] It's personally rewarding for people to be able to work on an industry standard like Spring, much more so than working on some proprietary piece of software. Open source helps us to hire better people and to retain them.
[Matt] As a software engineer, I feel that a company that makes contributions to open source is doing something that will help to attract and retain high-quality engineers. The best engineers really "get it" when it comes to open source.
It's clear that you are a premier consumer of open source. But why are you open sourcing these technologies? What goal(s) do you have in mind?
[Matt] We have a history of contributing to other open-source projects. Brian Zimmer and others on the team have been very active in open-source projects.
But this isn't something we do purely out of the goodness of our hearts. If someone contributes to ERMA, for example, that helps us. If you think about other projects like log4j, we get the benefit of others contributing to it, just as they benefit from our contributions.
[Winthrop] But it's also part of our job. It is who we are. It makes logical sense for us, but it's also part of our culture. It's a public display of our values. We're proud of it and believe we should be doing this.
So, are you hoping that Travelocity, Expedia, and other competitors will contribute?
[Matt] ERMA is really a general purpose API for monitoring, and is not specific to the travel industry. It could help our competitors with their customer experience, sure, but this is much broader than travel.
[Winthrop] If this helped them do a better job with customers, great. Ideally, we'll see other groups beyond the travel industry pick it up. It makes it easier to monitor technical operations, broadly speaking, which is good for everyone.
[Matt] As the Internet continues to evolve with more and more interconnections, the complexity of the whole is increasing over time. This is why complex event processing is so suitable to this growing problem. It lets us take an enormous amount of data from our data center and boil it down to the essentials: One message that says a customer's attempt failed because of "X," and the source is "Y." Our operations people shouldn't have to learn each of the seven layers of our architecture. They should just receive an easy-to-understand message from the system that makes it easy to solve the problem.
What do you think is the primary inhibitor to organizations like Orbitz contributing code to open-source communities, or creating their own open-source projects/communities?
[Matt] It's not as easy as it looks to create an open-source project and build a community around it. It's hard to create a standalone project that doesn't pull in all sorts of internal code dependencies. It requires extra work, which means you really do have to think about it as an investment that may take time to generate significant returns. We hope that ERMA will grow and attract a vibrant community, but we're not kidding ourselves that it will be automatic.
[Winthrop] We're in this for the long haul. It's not something that we're going to see helping us in the short-term on the quarterly balance sheet. As a company, we have a flood of great business ideas that we need to sell to the market. It's actually a big deal to have developers shift focus and resources to contribute open-source code. We do it because it's important to our technology teams and to the company, but it's a longer-term effort.
[Matt] Yes, and perhaps it will get easier. Now that we've paved the way, we're hoping that this leads to more open-source projects emerging from Orbitz.
I do, too. Leon Chism used to fly mostly solo as the face of Orbitz's open-source approach to the market. It looks like that is set to change by widening the group within Orbitz that actively and publicly engages in open-source communities.
Twenty years from now, we may look back on this moment and pinpoint Orbitz's bold approach to open source as a turning point for enterprise IT, one that opened the door to greater involvement in industry-wide, collaborative development. If we get to that point, I hope Winthrop and Matt will take the time to call me up and say, "Told you so." It's very impressive what they're doing.