Friday, March 15, 2013

The File System Evolution and Dropbox Strategy

Dropbox acquired high-profile email startup Mailbox today. Why? I think it’s about the evolution of the file system.

In the beginning in the ’60s and ’70s, computing was done entirely on a shared server in a time-sharing-style system. The client was thin and dumb, nothing more than a terminal and a keyboard. All the data and processing lived on the centralized computer (think Cloud, but just one box, where the box is the size of a whole room).

With the rise of the PC in the ’80s, computing moved to a fat client model where all the data and processing resided locally. Everyone owned their own computer and bought boxed software to engage with new apps locally. This stage was the advent of the local file system paradigm that we know well and still use today.

With the rise of web services in the ’00s, more of the processing and application logic are moving back into the server again; echoing the timesharing model of the past. The client is thinner again, mainly just a browser window and javascript GUI rendering. This time around, the cloud is distributed; it’s a field of cargo containers filled with servers in multiple geographies instead of a single box in another room. For some applications the data has moved into the cloud too (gmail stores all your mail in the cloud). For other web services, the data is still local.

Dropbox has done a wonderful job building a service that gives users the convenience of local data files with the distributed availability of cloud storage. It’s an invaluable, amazing product that I depend upon everyday.

But, their potential risk is the swinging pendulum of power between the server and the client. Dropbox was conceived in an era where clients were thick and held the primary responsibility of data storage. Data storage locally is convenient but not accessible everywhere and Dropbox fixed that. However, as more and more of the file system moves into silos in the cloud, Dropbox’s utility decreases.

So, what will Dropbox do? Buy some silos (or access to silos). The Mailbox acquisition is a way for Dropbox to create a relationship between the end-user and the data in their inbox.

Email data is incredibly rich. When I first joined Union Square Ventures in 2006 I asked the partners there how do I access our shared network drive. They said they had one, but no one uses it. I was surprised because in my prior job the shared network drive was the cornerstone of my team’s collaboration environment. I asked how people access shared files at USV, and I was told that people simply used email as their file system. The idea of an email-based file system is not uncommon apparently. I’ve encountered a number of organizations that continue to do business this way. USV ended up moving to a more centralized file system as the team grew (I have no idea what they use today as a file system; Spark uses Dropbox for Teams), but it was a very telling experience watching their initial email-based workflow.

Email is not the only place where the file system is moving into the cloud. I had two people tell me recently that they deleted their music libraries and now depend solely on streaming services like Rdio and Spotify to deliver their music. This is GBs of data that Dropbox used to control that is now moving away from their reach.

Rather than build competing silos in all these categories (like email or music), I expect Dropbox will make more smart acquisitions like Mailbox where a company has a direct relationship between a user and their silo’d data.


  1. eyesoftx reblogged this from thegongshow
  2. refineunity reblogged this from thegongshow
  3. khuyi reblogged this from thegongshow
  4. mdiabi reblogged this from thegongshow
  5. prosperitytech reblogged this from thegongshow
  6. newspeedwayboogie said: we use gdrive/gdocs now
  7. thegongshow posted this