Many VMware watchers were interested when Jerry Chen, a well-respected executive with the virtualization giant, left his role to become a VC with Greylock Partners. What, we all wondered, would Chen come up with for the firm?
Well, we didn't have long to wait: Chen's first investment was in Docker, the company that has taken an existing technology (Linux Containers) and built a huge movement around it. While many of us scratch our heads to work out exactly how Docker will live up to its astronomical valuation, at least on paper it looks like Chen hit the mother lode with that deal.
So all eyes were watching to see Chen's next move. And today we have it with news that Greylock (via Chen) is investing in little-known database startup Noms. Noms, which has been in stealth for about nine months, is a new decentralized database with ideas taken from a plethora of different sources, Git, Camlistore, bup and IPFS among them.
Aaron Boodman, the founder of Noms, explained Noms as being like source code repository Git in some ways, and unlike it in others. per his blog post:
The easiest way to understand Noms is by comparison to Git
Like Git, Noms is a fully-decentralized database with no single source of truth. A Noms database can be replicated and edited concurrently on many disconnected computers, any set of which can later come together and synchronize their edits.
Like Git, Noms is a content-addressed, immutable, and versioned. When you write data to Noms, you are appending, not overwriting. Edits are not destructive. You can rewind to any previous version of the database, fork from any previous point, and quickly see the differences between any two versions.
Unlike Git, Noms stores structured data rather than text files, and is intended to be used primarily by software, rather than directly by humans.
Unlike Git, Noms is designed to scale to very large data, while retaining the ability to diff, fork, and merge quickly.
He suggests that Noms will be particularly useful for a few key use cases: import and archival, aggregation and transformation, synchronization and merge.
In explaining his rationale for investing, Chen penned a blog post that puts the case. As he points out, the world is made of data and, increasingly, it is data that will define the products and services we use and trust in the future. The key problems with data, however, center around sharing and enabling individuals to build applications with the data pool which is, obviously, highly distributed.
Data stuck in silos, in databases, in spreadsheets and in web pages all requires different techniques to use it -- versioning simply compounds this problem. By extension, building applications on top of all of this data is increasingly difficult since the days of a single database and monolithic stack upon which to build solutions are long gone.
Which is where Noms comes in. Like Docker, Noms is an open source solution, in its case a fully decentralized database that, as Boodman pointed out, allows users to publish and consume data. Noms allows users to diff, branch and sync any data, and ingests any data from photos to stock prices.
The Noms team (in addition to Boodman, a small team is involved) have tackled these problems before -- they helped build Google Chrome and the Chrome extensions platform, Chrome OS, ECMAScript and Greasemonkey. As Chen sees it, they have thought more about moving data around the internet than almost any other people in the world and have written multiple generations of sync code, from AvantGo to Google Gears to offline Gmail to apps for wearables. Clearly, this team has been thinking about this problem for a long time.
Chen is a smart investor and a skilled technologist. The problem that Noms is trying to solve is very real. It is going to be interesting to watch this company progress.
This article is published as part of the IDG Contributor Network. Want to Join?