Cubbit is the distributed cloud.
Its architecture is designed to fully leverage the interconnectivity potential of the internet, while at the same time offering state-of-the-art performances and user experience.
The Swarm: hybrid distributed architecture
The heavy load, namely storing and transferring files, is carried out by a peer-to-peer network of Cubbit Cells, single-board devices that together forms what we call “the Swarm”. Meanwhile, a coordinator server handles authorizations, metadata, and applies optimization algorithms and machine learning to optimize the network.. Thanks to the distribution of workload, this hybrid architecture has ~1/1000 the maintenance costs of traditional data centers. This way, Cubbit enables the first collaborative cloud storage that is at the same time hack-proof, sustainable, and free from monthly fees.
The Cubbit Cell: a node in the network
The Cubbit Cell is the node of the swarm. It contains an optimized single board computer and a hard disk of 1 TB. By connecting the Cubbit Cell to the network, the owner immediately receives 512 GB of storage space on the Cubbit Cloud, as it contributes to the network with its internal hard disk space.
The Cubbit Cell: expandable storage
Additional space can be obtained by plugging in any storage device (sticks, hard disks, etc.) via USB to the Cubbit Cell. Half of the attached storage is made available to the user, up to 4TB of cloud space by providing 8TB to the network. The remaining storage is used for redundancy shards, and network operations.
Security and privacy are implemented by design. Before the upload, every file is encrypted with an AES-256 key that is randomly generated client-side and added to the user’s keychain. This keychain is then encrypted by a key generated from the user’s password before being saved on the coordinator server. This way, nobody, not even the Cubbit team, can access the user’s files. This is the zero-knowledge architecture of Cubbit.
In order to counter momentary disconnections and downtimes of single Cells, Cubbit implements a redundancy procedure based on Reed Solomon error-correcting codes: each encrypted file is divided into 24 pieces, which are processed into 36 redundancy shards. Out of the 36 shards, only 24 of them are necessary to retrieve the original encrypted file. This procedure alone ensures a statistical uptime of ~ 99.9999% (probability of downtime 10-6). Moreover, the coordinator monitors the uptime status of each Cell and triggers a recovery procedure when the total number of online shards reaches a threshold of 30/36, virtually achieving 100% uptime.
Once the file has been encrypted and segmented, the client communicates with the coordinator to obtain authorization for uploading it to the distributed cloud. The coordinator, in turn, verifies the authorization and finds the optimal set of 36 Cells to store the file by minimizing a cost function that accounts for geographical proximity, mean uptime, free space and other metadata. It then acts as handshake server to initiate the peer-to-peer connection between the hosting Cells and the client, which distributes the shards on the network.
While the upload procedure looks complex, it is actually invisible for the final user. Thanks to the coordinator server, authorizations and initiation of connections to the hosting peers are handled in fractions of a second. Once p2p connections are initiated, Cubbit takes full advantage of parallel transfers to ensure upload and download speeds, which are at least equal or even superior to centralized data centers (especially as regards the parallel download).
The internet infrastructure is responsible, as of today, for the 10% of the total worldwide energy demand [1,2]. Data centers account for one third of it, making “the Cloud”, despite the ephemeral name, an ecological monster that consumes as much as the entire United Kingdom (66 millions inhabitants and 5 th world’s economy).
Cubbit is based on small, optimized single-board computers, which have an impact per GB that is 10 times smaller than data center racks. Moreover, it can leverage on geographical proximity to avoid long data transfers, which, in certain cases, can be as much consuming as storage itself . The result, detailed in our green paper, is that an average storage plan of ~5 TB will save, choosing Cubbit over traditional cloud storage, the equivalent consumption of an always-on fridge in a year.
-  Greenpeace International. How clean is your cloud? catalysing an energy revolution. Technical report, 2012
-  Mark P Mills. The cloud begins with coal. Digital Power Group, 2013.
-  Jayant Baliga, Robert WA Ayre, Kerry Hinton, and Rodney S Tucker. Green cloud computing: Balancing energy in processing, storage, and transport. Proceedings of the IEEE, 99(1):149–167, 2011.
Leave your email and get all the inside news from our Kickstarter campaign