Bitcasa: Unlimited storage, version history & sync

bitcasa-sBitcasa has just emerged from “skunkworks” mode. The cloud storage startup made waves in 2011 as finalist at TechCrunch Disrupt and runner up at Startup Battlefield. After burning through an initial $2 million, they landed an additional $7 million in June 2012. While there were few updates during 2012, some analysts noted that they filed for 20 patents—a few are really slick! Now, during Feb 2013, they have unveiled a cloud service with an edge over all others (SkyDrive, iDrive, Dropbox, Sugarsync, etc). In my opinion, only Symform and SpaceMonkey come close to the model that I described 3 years ago (search for ‘Ellery’ and ‘RDDC’).

Bitcasa gives every user folder sync, a timeline for version recovery, and cloud storage without limits. And, I really mean limitless! By the end of next month, I may be using petabytes, as in millions of gigabytes! The space available to me shows exabytes are still available.  That’s more than all the grains of sand on the world’s beaches and all the stars in the heavens. How much does this cost? Just $99 a year, or $49 if you sign up early this month. (Promo Code: BETATHANKS). WildDucks can help this Blog by using our referral link. It tacks a free month onto your editor’s subscription.

I can’t guarantee that Bitcasa will be around next year. After all, most startups fail. But in this case, I crafted a substantially identical network architecture years ago. I understand the business model. Even with a high fraction of data hogs, the venture can profitably service users for the long haul. If an understanding of the secret sauce isn’t sufficient to assuage hesitation, this interview with CEO Tony Gauda will floor you. He combines the technical and marketing genius of Steve Jobs with the showmanship of Siegfried and Roy, and the smile of Barak Obama

https://www.youtube.com/watch?v=3ImZWVxAQ_Q

Damon Michaels, a WildDuck contributor wrote:

Seems like a virtual drive. I need automatic backup of
my important data. I use Carbonite for this right now.

The folder-sync defaults to all drives in their entirety—even external drives and network attached storage! If you accept the default, it always backs up everything. But more importantly, Bitcasa reverses the model. As connectivity becomes more ubiquitous and speedy, they want you to use the cloud as your primary active storage. Eventually, it will even host your live EXE files (your apps) and your “bootable” OS. The synchronized copy on your PC will be the backup – as well as the one that is used when you cannot connect.

I proposed the fundamental principles used in Bitcasa architecture in this Blog, and 3 years ago in other articles. I called it a “Reverse Distributed Data Cloud” (RDDC). My spec adds distributed, P2P storage to the model. This reduces cost, creates redundancy, and makes a far more robust system. Not only does it get rid of the data center completely. With my model, it is unnecessary for the service provider to perform any backups. In effect, the live cloud is a RAID 10,000 constellation.

One architectural trade-off is the desire for massive de-duplication –vs– the compelling need for end-to-end encryption, in which only the individual users have the keys. These two features are incompatible. DropBox and Bitcasa claim that files are encrypted at the sender and that private keys are never given to the service. While technically true, that claim covers up a nasty little detail. They use a method called Convergent Encryption in which encryption keys are derived from a character string within the encrypted file. Although the service cannot decrypt a unique file (for example, your income taxes), they could compare a hash of your file to one provided by a government or alleged rights owner, thus proving that you have stored a copy of contested media. They could block access to movies and music that you have stored or even block your original upload. The good news is that with a full RDDC implementation, the need for de-duplication is greatly reduced or even eliminated. Therefore, a properly implemented RDDC can truly empower its uses with strong, end-to-end encryption.

I’ll report more about Bitcasa after a few months of use. For now, I feel ratified to see my dream taking shape at several American ventures. If you find this field as fascinating as me, check out Symform, SpaceMonkey and Digital Lifeboat. That last venture is floundering, and may be bankrupt by the time you read this. But they have some very compelling technology for p2p, distributed storage.

12 thoughts on “Bitcasa: Unlimited storage, version history & sync

  1. Seems like a virtual drive. I need automatic backup of my important data. I use Carbonite for this right now.

  2. Yes. Bitcasa claims that there is truly no cap to your storage space. You may store thousands of terabyte drives, if that’s what you wish to do. Of course, at some level, the claim must be hyperbole. If you attempt to match the storage of archive.org, you would certainly run into a problem (they catalog the entire Internet, including historical snapshots of every web site on earth). At an upload rate of 100 Mbps (a very fast internet ISP service), you would require several million millennium.

    Clearly, Bitcasa hope to “de-dupe” the majority of your stored content–-perhaps even in your PC (before you upload), based on a digital hash. What appears on your cloud drive as your OS files or your collection of 5,000 songs, would be stored in your region of their servers as pointers on their servers to common files stored on behalf of many users.

    But they cannot de-dup your personal photos and documents. Other users do not have identical files. Even so, they can do some remarkable “cross-file compression”. For example, they can build efficient Huffman-encoding tables to improve compression on similar files, even if files are not identical.

    But what about privacy? You could layer your own encryption onto the connection (such as TrueCrypt), which prevents Bitcasa from saving any storage space. They do not prohibit you from doing this, but they discourage it by claiming that your data is encrypted anyway. They additionally claim that they do not have the private key to your encrypted files. But Caveat Emptor! While the statement is technically true, this form of encryption has a serious drawback for any files that are not unique to you, such as commercial movies, music, or software applications.

    For many users, this niggling, little detail will be of no concern. But for those that don’t want to upload evidence of their file origin to a cloud service, there is still a simple way to circumvent the problem—and without layering another encryption system on top of every upload. In fact, I am adding the technology to my own RDDC specification. I have discussed the approach with a few readers in the past.

  3. Hi Ellery,

    DARPA and other agencies are looking into (what you call) “reverse” storage architecture and (especially) p2p architecture for some really interesting projects. {Redacted} Some in-house military R&D may give way to using commercial technology, if it is convincingly secure & robust. (a very big “IF”).

    I suspect that RDDC architecture is robust. In my mind, this makes sense, because it is effectively a “crowd sourced”, self-healing, n-way RAID matrix (a robust, storage “swarm”). But this still raises some questions about individual players and how we can best identify the partners that can manage a nexus…

    === Is Bitcasa truly on the path to a full RDDC implementation? It seems to be your latest “darling”
    === What about your past darlings: Symform & SpaceMonkey? Are they credible players with great tech?
    === Are any of these companies doing torrent re-acquisition? It was a pillar of your original model; a most intriguing feature.
    === Can RDDC be configured to avoid the admin/nexus altogether? We don’t even want to be dependent on DNS.
    === Is Digital Lifeboat really floundering? Is our old Microsoft buddy still involved?

    That last question scares the bejeebers out of me. It is one of the companies on which we might have staked our national security (Figure of speech. Any project would be phased in series of carefully staged trials with fail-over mechanisms. But still…). I have read some of your early papers and I know that you consult in this field. My goal is to let my people pull back the sheets just a few inches and see if you can help us to quantify the risks. To do this, you will need to know a little about our goals, and so it could lead to a consulting gig. With this in mind, your reputation as a privacy advocate should not be a problem, because in this case, it coincides nicely with our our own requirements.

    P.S. Bitcasa CEO, Tony Gauda, sure presents a better face-forward than DropBox founder, Drew Houston. Ha! Talk about charming and ebullient! Gauda’s got it all. That goes far in carrying investors and partners.

    Did you know that Houston was approached by Steve Jobs a buyout discussion?

  4. You cover a lot of ground. Too much for this Blog. A few quick responses might interest WildDucks, but let’s take the remainder of our discussion off line.

    o Yes, Bitcasa is my current ‘darling’ (Your term, but accurate). Yet, I have not had contact with Bitcasa except to purchase their Infinite Storage service.
    o I don’t know if Bitcasa is moving toward additional features of a full-blown RDDC. I don’t even know if they have looked into it. (That’s a lot of not-knowing). Eventually, I may reach out.
    o Yes, Tony Gauda is certainly magnetic. I don’t see him as showman. Rather, he believes in a new paradigm with a passion.
    o I have no basis on a comparison to Drew Houston. In fact, with the exception of a few engineers who picked up on this Blog, Dropbox never responded to my ticker. I have the impression that they like my architectural suggestions, but were directed to avoid external discussion.
    o I like your term “swarm” for a communication p2p hand-off architecture. I first became interested in the profound potential of IP packet swarms while reading Michael Crichton’s Prey. In that novel, fleets of microscopic robots not only acted with a collective conscious, they could actually assemble into human form to better communicate with our species. In the late 90s, I filed a patent on “Fungible Spectrum” that became the basis of antenna-hopping schemes used by Cisco and Nokia.
    o I was aware that DropBox CEO, Drew Houston, turned down a buyout offer from Steve Jobs. I saw it in a video interview with Houston and mentioned it in my Blog here.

    If WildDucks wish to discuss cloud architecture, please don’t hesitate to use the Contact form at the top of this page. The topic is my very favorite subject (along with Blind Signaling and Response). I encourage a lively and active exchange. It may seem that I am discouraging Chuck’s inquiries. Not so. Rather, the issues he raises are too specialized and speculative for this forum.

  5. Thank you for catching this error, Alyssa. I was researching several cloud storage start-ups as I prepared the article and I goofed. Sources of the $2 and $7 million financing are described in this press release.

    • Hi WP,

      Give your network folder a drive letter. If you use Windows, try mapping first. If that doesn’t work, use the Subst command. On my PC, Bicasa picks it up as a real drive.

  6. Pingback: Some cloud storage / online backup providers compared (DropBox, Sugarsync, Bitcasa, …) « The Wiert Corner – irregular stream of stuff

Leave a Reply

Your email address will not be published. Required fields are marked *