Re-thinking IO

Have you noticed that “IO” means different things to different people? “IO” to an administrator or developer of Cisco gear means network, whereas “IO” to an administrator or developer of EMC gear means storage.

To confuse things even more, “storage” means different things to different people. EMC, Netapp, Brocade, Q-Logic, and Emulex are all storage companies, but they are very different from one another. EMC and Netapp are storage target companies. Brocade is a storage fabric company. Q-Logic and Emulex are storage initiator companies. They each represent disparate functions in the storage stack. (now Diamanti) believes that “IO” encompasses ALL of the above – networks, initiators, and targets – anything that is outside the scope of the host CPU and its memory.

The challenges in managing, scaling, and achieving predictable behavior in the data tier is well documented in blogs, white papers, videos, et al. which many are trying to address by fixing a piece of the IO stack. There are projects and whole companies being funded for “storage for containers” or “network for containers”.

But to properly address the data tier, Step 1 is to solve IO end-to-end holistically. That means a re-think of initiator-network-target architecture. Solving only one of those silos or cobbling together parts of the stack won’t cut it. At the center of’s (now Diamanti) technology is the marriage of network AND storage, because it’s about both data in motion and data at rest.

Step 2 is to make the technology consumable by those who need it: application developers and infrastructure administrators. That means plugging into frameworks like Kubernetes, Docker, and Mesos and contributing back to ecosystems that drive the innovation and adoption of new architectures such as cloud native applications and containers.’s (now Diamanti) solutions are about further enabling the fundamental values of these new architectures – speed of innovation and time to market.

At the end of the day, it has been the open source community that has led people to re-think how they develop and deploy applications and infrastructure. We believe it is now time to re-think IO.