NIH is one of the best known acronyms in our industry. However, for those unfamiliar with it, NIH stands for Not Invented Here, and refers to the tendency of engineering teams to devalue technology from outside sources. Unfortunately, NIH is alive and well in many parts of tech, and nowhere is it more prevalent than in Cloud Computing. Seems some folks just never learn that no single company can be all things to all people. Paul Miller wrote a great post on this after reading our AppStore announcement (so good that I borrowed his title) and he got me thinking.
We’ve lived through this cycle since the beginning of the computer business. IBM in the 70′s would deny service and support if you had any peripheral attached that wasn’t theirs, from a disk drive to a terminal (remember those)? Apple rocked the computer world in-part by building a completely open system that anyone could write code build peripherals for. Within a few years their were more applications available for the Apple II than any other computer in the world. When IBM introduced the original PC, the team learned from Apple and so the specs and OS were completely open and a whole industry grew to support and promote it. After th PC industry grew large, though, IBM tried to pull back with the introduction of Microchannel, but the industry turned it’s back on them and built an open platform. More recently we’ve seen digital rights management beaten down by consumers in favor of openly copyable MP3s. So, now we come to Cloud Computing and we once again see companies trying to close the system. It’s their language, their data center, their API. And, once again, I don’t think it’ll work.
Opening up is hard. Customers often do things with your system that you don’t expect. How will the system react? How can you test adequately? How can you offer support? How can you assure uptime? These are all valid concerns, but they don’t outweigh the advantages of being open. For customers, being open means they can add other vendors products to yours, offering more complete solution sets, and the ability to move solutions to meet business needs. For vendors, customers will show you new markets, introduce you to new partners, and also let you know where your product is deficient.
At 3tera, we never assumed we could build a worldwide cloud ourselves. That was part of the reason we chose to work with service providers, and as a result we have a user footprint that stretches around the globe. We don’t speak many of the languages of our customers, but our partners do. We also decided not to require APIs for accessing storage or networking and instead did the extra engineering work to properly virtualize IO. As a result customer constantly surprise us with what software they’re able to run on AppLogic. For instaqnce, we never envisioned someone using AppLogic as a phone switch, but a customer did.
With AppStore we want to make the cloud more accessible to a broader set of solutions for customers. Software developers will be able to publish their products, users will be able to drag and drop that technology into their applications, and we’ll connect the two. ISVs can tell who’s using their product, and users will be able to get support. It’s the quitessential town marketplace. Will there be challenges along the way, of course, but in the end customers will get the best system.