The dropbox endgame

I'm not sure I agree at all with Jacques maths, the cost of local storage is tiny compared to the cost of Dropbox itself, and that's unlikely to ever change

A hard drive costs maybe $30 for 100GB, higher capacities have much better GB/$ ratios. Dropbox costs $240 per year for 100GB, which has to cover both the storage and bandwidth costs.

At the moment, the growth in data stored is increasing at much higher rate than the rate at which the cost of storage and bandwidth is decreasing.

Sure Dropbox is great and a local/remote sync process is likely to become normal over time, but it won't be for cost saving reasons.

I'm pretty sure this is also the current vision of the Dropbox team. They are making a lot of effort to become a platform not just a product.

Does the author mistake the past with the future? A cloud-based home dir is basically a network file system (samba,afs,nfs,...) We had them for decades and they suck. Dropbox is an improvement, because is not necessary anymore to be "always on".

Doesn't seem wise to entirely trust your data to a single entity. Can we be sure dropbox will still exist in 5, 10 years time?

What happens when your net connection fails?

Local storage will forever be around, in some form.

Dropbox's core competency isn't cloud storage, it's continuous client syncing. If anything, moving towards not using the local file system at all hurts Dropbox, as they lose the part of their service that makes them special - syncing all files locally.

On top of that, I can't imagine removing a local file system would work as a general purpose solution. By the time network bandwidth catches up to today's bus speeds (probably years away), the new bus speeds will be that much faster, and that's what people will want. Managing files locally may be a pain, but that's a UI issue. Users want native performance. Cloud files are a leaky abstraction. Dropbox gives you local files and abstracts away cloud storage and syncing. That's the way to do it.

This article is very interesting. The iTunes example in particular shows that these measures have to be transparent or they won't work very well.

I think this was very much Microsoft's vision for Livemesh; they were thinking about desktop apps running data across the cloud.

There's a real potential for dropbox if they could somehow hook into a variety of web apps and provide a common back end.

The problem at the moment is that if you assume much of the app work people do will move into the browser, you're losing the common file base. You have several app-specific silos across the web.

Very interesting idea. However, it made me think that one major flaw of Dropbox will limit its potential as the "ultimate storage."

Dropbox is very memory intensive, even for my 5 gigs and my 40,000 files in the Box. If you cover my entire hard drive, that's over 100 gigs and millions of files. Each file (and each sub part of a file, sort like like a Merkel Tree) must have its finger print stored in memory. Currently those 40,000 files use about 70 MB, and when syncing, consume an entire CPU. I can't imagine how resource-intensive it would become if I indexed my entire hard drive now.

I don't think Dropbox can push order-of-magnitude improvements in resource usage, even if they move away from Python. I'm sure they're already doing very well. I have multiple friends working there, and I have talked to Drew and Arash. All of them are very brilliant people and I doubt that they haven't consumed all the low hanging fruit in performance.

Sure, in the future using 20x the memory may be OK. That's just 4.3 doublings, so maybe memory increases may balance it out. But don't forget that the amount of data I have will also increase.

So I don't think it's feasible (with the current trajectory of technology) to have Dropbox serve as the ultimate storage medium, unless we develop a different way of syncing files.

I'm not trolling - but this is hardly a 'weird idea' as the article states ... I thought of this and did it last weekend :)

My flatmates harddisk broke down the other day in his laptop, so I put Ubuntu live CD in, a USB stick and hooked up dropbox. A script later, and he's using dropbox for permanent storage online, and considering not buying another hard disk.

Of course he will :P .. but my point is that this seems a pretty obvious progression :)

Apple already made the first step by releasing an OS that is geared around the idea of cloud-based computing (Lion). It's just a matter of time before this happens, and when it does; Dropbox might have a hard(er) time competing.

http://www.patentlyapple.com/patently-apple/2011/02/apple-in...

How will this work out in a world where the telcos seem to be bent on increasing costs and introducing data caps?

I live in South Africa, where the normal ADSL account gives you 3 Gb of traffic a month (10USD/gig for each gig after that) at a maximum of 4 Mbit/s (10 in a few areas). I know we are probably an outlier, but judging by what happened in Canada recently, and a few cases in the US, its seems possible a reverse case of Moore's law will come into play.

Also, most people (speaking globally), wont have the network capacity to support this kind of system for many years. There would likely have to be a truly massive increase in network capacity before we ever see this really take off.

For fun, try running your computer off an NFS - even with a 100Mbit connection, it gets annoying quickly.

Dropbox sits in that transition period from local to cloud storage. I said earlier in my comments that cloud will be an OS feature and Apple will lead that.

However it will not start with pure storage(their approach of experience rather than specs). It will start with your music,video, pictures. Than your ios+local apps (with preferences and save files). And then Computer setup (remember your OS setup). Thats why Apple is easy with offering SSD harddisks in their top tier products.

I believe this would mean giving up the main advantage DropBox has over all competitors: It's the first cloud storage provider that sucessfully mixes local storage and cloud storage by providing transparent synchonization. After all, the DropBox folder is just a plain folder, but it is sync'ed whenever something changes. Without the local storage part there would be no reason to use DropBox as one could simply use for example a WebDAV share or SFTP.

Sounds like the Chrome OS vision to me.

Dropbox is a great place to store media and created content but doesn't easily accommodate your registry hive or myriad support directories that you need to make programs actually run.

If you re-architect to achieve all of those goals and still have sync, you get ChromeOS. Barring that you have a nice replicated USB key (which Drew describes as the original original vision)

As long as there's no 3G or other (affordable) wireless internet access on the London Underground or anywhere else where I would need my files, I'm simply not interested in purchasing a product that doesn't have local storage. I mean, imagine an iPhone that would stream everything, I couldn't even listen to music during my daily commute.

This is how I use Google. My docs, calendar, and email are everywhere at once - no syncing necessary.

The thing I don't like about cloud services is this: They can be served a search warrant to let government go through what belongs to me without my ever being notified.

I in a way agree with the OP however I think we are decades away from the proposed future. The problem is multi-tiered

- If you are outside of a major city, chances are you don't have internet wireless coverage everywhere. In fact many places in the city face this issue. This is more of a physics issue than anything. Inside my office building I get fairly bad coverage. Don't forget tunnels and such. My apartment has ZERO coverage. Only my wifi. And not all apartments I visit have wifi or coverage.

- Price and transfer rate. Sending my hard drive to dropbox is about $50 a month (or more) and not to mention taking about a month to actually upload.

- I don't want to rely on a company like Dropbox to exist. If they fail (see delicious) I don't want to be stuck with nothing. In fact single points of failure are bad. And I don't mean Dropbox can have a hard drive failure, I mean what if DB decides to triple their price.

- America's internet infrastructure kinda sucks with current insanity pushed by isps regarding data caps/etc. Just wait till people REALLY put our infrastructure to the test.