Allmydata 3.0: The Ultimate Online Storage Solution
I have to admit Allmydata has had a rough time. The original team had difficulty executing on the vision and most had to be replaced. Using a peer-to-peer client to have lower bandwidth and storage costs and hence lower pricing have not been a comparative advantage in an over-funded world where many money losing free competitors have emerged. Moreover, despite the security and encryption users hated the idea that their files were stored on other people’s computers. The product we had originally devised was much more difficult to use than we had anticipated. Frankly the software was slow, bloated and buggy!
Over the past year we rebooted the company. We changed the team, rewrote the specs, redesigned the product and changed the business model:
- We redefined the problem Allmydata is trying to solve: Allmydata is the solution for online storage, backup and sharing.
- The way we decided to solve this problem is by building a virtual drive on your computer that acts exactly like a regular drive if you use the Allmydata client:
- Allmydata shows up as a drive
- You can store as many files as you want of any type
- You can run files directly from the virtual drive or copy them to your local drive
- We decided not to build our own backup application, but instead to integrate to the backup applications built into Vista, XP and Mac OS or allow users to use any backup client they want
- We created a web interface that looks exactly like the client to give remote access to files without needing to download the software client
- We simplified the business model dramatically removing tiered pricing and offering an all you can eat plan for $4.99 per month for unlimited storage with a 1 Gb free plan.
- We removed peer to peer from the downloadable client as it was too resource hungry and slowed uploads down dramatically. We did keep it on our back-end to split the files on our own servers to allow us to use cheap Linux boxes and hence have much lower costs than all our competitors, Amazon’s S3 included.
- We rewrote the software client such that memory usage went from as much as 200Mb to less than 30Mb.
- Simultaneously with the Allmydata 3.0 release we are releasing a beta version of the Mac client both for Tiger and Leopard.
All the changes above happened incrementally over the past year or so. None of them on its own was a game changer, but we would like to believe that the sum total of the improvements create a “Skype-like” (meaning easy to use and understand solution) for users’ storage, backup and sharing needs.
We’ll see what the future holds but for the first time in a long time, I am becoming optimistic about Allmydata!
On the subject of file backup, sharing and storage …
Online backup is becoming common these days. It is estimated that 70-75% of all PC’s will be connected to online backup services with in the next decade.
Thousands of online backup companies exist, from one guy operating in his apartment to fortune 500 companies.
Choosing the best online backup company will be very confusing and difficult. One website I find very helpful in making a decision to pick an online backup company is:
http://www.BackupReview.info
This site lists more than 400 online backup companies in its directory and ranks the top 25 on a monthly basis.
Fabrice,
Quick question, how do you do it?? Being CEO of AllmyData and having problems like these, finding time for OLX and raising funding, then I believe you are on the board of several other investments, and you happen to find time to ski at Whistler and play tennis !!! Don’t tell me you sleep just 2 hours a day 🙂
I am not CEO of Allmydata. I am just an investor and on the board. Peter Secor runs Allmydata. It’s good to focus in life 🙂
In terms of time management, I am fairly effective. I focus on whatever I do at the present and I don’t do anything I don’t want to do (cleaning, cooking, meeting people I don’t appreciate, etc.).
I find it interesting that there is no comparative advantage from p2p / reduced costs in the backup world, while Skype clearly found one in the VoIP world. I imagine that backups are an even larger user of bandwidth than VoIP. This suggests that this market is severely overfunded as you mention.
But what happens when the music stops? What happens when competitors can’t make money? An economic downturn could separate the wheat from the chaff, and a p2p model could win the day (as the early VoIP providers disappeared, and Skype won). I guess your hope is that you truly will remain the low cost provider with p2p on the backend, but that rings a little hollow to me. Amazon’s services are built on cheap linux PC’s, as is Google’s new offering.
I find it a shame your original model couldn’t succeed. I would argue it had less to do with p2p than with the product itself – if a product is dead simple to use, the nuts and bolts on the back end shouldn’t account for much. Resource usage is a problem, but this could be solved with batching I imagine.
Did you consider open sourcing your code? That could also be a competitive advantage – and you could charge for additional back end server side storage.
Another approach would be to allow users to select their peers. This could work for SOHO type businesses, where they have plenty of spare space on work PC’s, but no good backup solution. Or families, that would be comfortable if they could *easily* help each other backup files.
I was pulling for Allmydata’s model, I find it a bit sad to see them become just like the rest. Still, best of luck, there are no clear winners yet.
Wow – I didn’t realize you had open sourced your code, I do look forward to digging in and learning more. Like I said in my earlier post, I loved Allmydata’s p2p approach when you originally launched, and I look forward to seeing how this all plays out with time.
Thanks for the response – great news all around!
Hi Murray,
Thanks for the thoughts on the P2P model within the context of this space.
There are definitely competitive advantages (cost of equipment and bandwidth) to using P2P in the backup world which is why we’ve continued to use it internally for our managed service. We chose not to expose the clients (yet) to the P2P-ness of our architecture for two main reasons: 1) added development complexity meant longer delivery and debug cycles 2) user misunderstanding of what P2P meant in this context meant longer sales cycles. By deferring the solution of these issues to a future date, we have been able to more quickly deliver a lightweight and useful service. In future releases we can deploy the P2P architecture to our user base as needed/desired.
Note that we have open sourced our code (http://allmydata.org) – feel free to play with it. We run a test grid and people have already started setting up private “friend-nets”. It has proved very valuable with great feedback and analysis of our codebase, particularly on the data security front.
Hi Fabrice!
I just visited allmydata.com to see how it’s going and was pleased to see the “iphone app” sticker slapped on the homepage.
So I clicked, and saw the app but I also saw the amount of one-star reviews. The reason for those basically negative reviews isn’t the people that are actually using it, it’s the dumb kids that download it thinking they can use it without having to sign up for anything.
I bet if you refreshed the design a bit, worked out any kinks, and called it version 2.0 you could start over. By charging 99 cents for it you keep out the punk kids.
On a related note, I designed the official “X-Men Origins Wolverine” app in the store, as you can see, I also have issues with one-star brats!
Best,
Tim Pedersen