We announced today that I am joining Gluster as its new CEO.
People have been asking me why a guy from a Web 2.0 consumer company like Plaxo (or a consumer entertainment company like Plaxo’s parent, Comcast) would take on the CEO role at a fairly technical start up in the storage space.
The answer? I believe that first step in building a great company is to take on a big challenge. From my experiences at Comcast and Plaxo, I have seen one of these challenges first-hand.
A Big Challenge
The consumer web has seen an explosion of unstructured data over the past few years, as hundreds of millions of users have been creating and accessing unprecedented volumes of videos, MP3s, photos, and more. Managing, storing, and delivering these huge files presents a set of economic and technical challenges that can’t be addressed by conventional storage solutions, which were designed for structured data (e.g. a financial database) used by much smaller numbers of users. Similar issues face consumer entertainment companies, who are trying to deliver vast libraries of on-demand digital content– including high def and 3-D video— to millions of geographically dispersed users.
However, this problem isn’t limited to the consumer space. Enterprises across the globe are experiencing the proliferation of huge, unstructured data files in the form of medical images, seismic data, surveillance video, atmospheric & climate models, virtual machine images, and seismic, nuclear, and DNA data. According to a recent report from IDC, humanity’s total digital output is expected to surpass 1.2 zettabytes (1.2 x 10^23 bytes) this year—the storage equivalent of a stack of DVDs that reaches to the moon and back. In the past year, the total amount of unstructured data created has grown by some 80%, and is expected to increase over 40-fold by the end of the decade. Much of this data needs to be stored permanently—and must be available in near real time.
Meanwhile, the cost of storing a gigabyte of HDD storage is dropping by only about 25-30% per year. And, to deliver equivalent performance, most conventional solutions reach a point where storage capacity needs to be added exponentially. See a problem?
A Big Idea
Fortunately, Gluster has been purpose-built to address today’s unstructured data storage and management environment. Starting from a clean sheet, Gluster’s founders made a number of intelligent decisions that have enabled them to deliver a storage solution that is easy to deploy, scales linearly, and delivers excellent performance at a fraction of the cost of conventional solutions. For proof, you can look to the hundreds of customers, across Digital Media, Healthcare, Energy, Biotech, using Gluster in production deployments from a few terabytes to petabyte-scale.
What were those decisions?
First—the team at Gluster recognized that storage is increasingly becoming a software problem. As such, they created a software solution. Gluster works with a broad range of hardware—from high-end SANs to commodity drives– thus providing flexibility, reducing vendor lock-in, and dramatically reducing hardware cost.
Second-they embraced the open source model. I’m a huge believer that open source increases the speed and quality of development, improves the feedback loop from customers, and fundamentally increases flexibility and reduces risk for enterprises. Gluster is being continuously improved by a large and passionate user base, who have downloaded GlusterFS over 100,000 times.
Third-By starting from a clean sheet, Gluster was able to challenge the architectural ideas that guided the development of structured storage and file systems. For example, by designing a file system without a centralized metadata store, Gluster provides true parallelism and eliminates the performance bottlenecks, reliability issues, and data corruption vulnerabilities associated with metadata servers. By fully embracing existing filesystems and networking protocols, they let users easily deploy in existing environments without rewriting applications, implementing a new APIs, or replacing hardware. And, by pooling disk and memory resources in a single global namespace that scales to multiple petabytes Gluster operates well in cloud storage environments.
A WORLD CLASS TEAM
Perhaps most importantly, Gluster’s founders assembled an incredibly talented team, which operates both in California and in Bangalore. Obviously, I think very highly of the founders, AB Periasamy and Hitesh Chellani). And, I can’t say enough about the value of the worldwide Gluster user community. Combine that sort of global talent with rapidly growing revenues, great financial backers, and a low cost sales & development model—and you have the makings of a world class company. Addressing global challenges with big ideas and a world class company? It doesn’t get much better for a guy like me.