Dan's Take

It's Godzilla vs. Rodan in 'Container Wars'

Dan analyzes the new fight brewing between Docker and its challengers, including the just-announced Rocket.

• By Dan Kusnetzky
• 12/04/2014

Docker launched itself and started talking in March 2013 about a container technology to make application development and migration much easier. Since then, many suppliers of hardware, software and cloud services have jumped on the Docker bandwagon in the hopes of harnessing a groundswell of interest in technology that would simplify developer's ability to run isolated workloads on one or more systems easier, and use a technology much lighter than virtual machines (VMs).

Recently a competitor, CoreOS, launched a Docker competitor, Rocket and published a post disparaging Docker's move from offering a simple Operating System Virtualization and Partitioning (OSVP) product to offering a complete platform based upon that technology. Docker responded quickly to CoreOS' assertions.

What's New Is Old Again
What we're seeing is a typical shouting match that often emerges in a new market. We've seen this type of behavior in the early market for UNIX, Linux, databases and many other types of software. As I've mentioned before, this really isn't a new market. We've seen technology of this type available on mainframes since the late 1960s, and on midrange UNIX systems since the late 1980s. It is a relatively recent addition to the Industry Standard (x86) market, however.

Is this type battle unusual? No, not really. We can expect to see other entrants to this race, and for the rhetoric to get louder and more heated before it's all done. When Linux was fresh and new in the late 1990s, my group at IDC was tracking nearly 400 different Linux distributions. Each supplier issued announcements saying that it was uniquely qualified to be an enterprise's Linux operating system, and then went on subtly -- or not so subtly -- to describe the problems with other suppliers' offerings. We saw a similar dynamic in the late 1980s when IDC was tracking nearly 35 UNIX competitors.

Typical Market Trajectory
Let's examine how the software market works: one supplier introduces a technology. This technology may have been developed in-house or come from research done by some academic or governmental research institution. If the technology appears useful and is priced correctly, enterprises quickly see the value and start to conduct pilot projects to learn how well it will perform in their environment. If the technology pilot is successful, a measured, careful rollout might follow.

Since software suppliers, like all businesses, follow the money, if someone is making money, or at least appears to have a strong potential of making money, other suppliers will develop a competitive technology. In the case of open source technology, both suppliers may be starting from the exact same source code.

These competitors yell at each other like Godzilla and Rodan. Later, when another suppliers jumps into the market, the shouting match gets louder – imagine Mothra appearing on the scene while Godzilla and Rodan duke it out.

Dan's Take: Splitting Up Tokyo
Customers who adopt technology early on are likely to discover that life isn't all roses. While the technology is likely to help these customers address their computing issues, they are also likely to introduce problems. As new competitors enter the market, customers are likely to face interoperability, development and management problems; this is because all the suppliers are trying to lock them in to purchasing products only from them.

What happens next is very predictable. Customers will complain and threaten to take their business elsewhere, unless the suppliers get together and make life easier for the customers. The suppliers will get together to create a standards body. The standards process will be contentious. The scene I imagine is Godzilla, Rodan and Mothra sitting down at the same table to decide how to split up Tokyo. After standards appear, the suppliers will -- grudgingly -- make their products comply with the standards.

Will life be happy for customers forever after? No, not really. If the standard isn't complete, suppliers are likely to try and slip in ancillary lock-ins into their products. Typical areas include: communications protocols, file formats, APIs or management tools.

This continues until the standards catch up and suppliers are forced to compete on product quality, support, pricing or integration with customer environments. Eventually, of course, these suppliers will go on to find some other technology to fight over. This is what makes the IT market so interesting and dynamic.