Centralized cloud storage stores data in massive data centers run by a single organization that consume large amounts of energy. Decentralized cloud storage encrypts data, breaks it up, and distributes for storage on drives that are run by different organizations in lots of different places — each with a different power supply and network connection—creating something much less wasteful. There are no data centers, no storage oligopolies and no vendor lock-in.
At first blush, it sounds a bit crazy. How can that work? How can it be feasible? How can it even be real?
The good news is it is very real and very feasible today. Here is our take on the common misconceptions we hear every day about decentralized cloud storage and the facts to bust them.
Myth #1: Decentralized Cloud Storage Can’t be Real
Decentralized cloud storage sounds a bit like magic. The idea of bringing together lots of independent individuals who are not coordinated and finding a way to deliver a great outcome does sound a bit far-fetched. However, that it’s not only possible, but exists today and is an actively-used strategy. Thanks to advancements such as bandwidth quality and availability as well as peer-to-peer networking technology, the path toward viable decentralized storage was created.
Decentralization has been around since the beginning of the internet. The internet actually got its start as a decentralized system with Arpanet, a loose collection of universities and other institutions. Even today, when we get on our dreaded Zoom meetings, our data is behaving in a decentralized way. Our audio and video is broken up into lots of little bits, encrypted and distributed over routers that are run by thousands of companies all around the world. As a participant in that Zoom meeting, you aren’t thinking about where the routers are located or who runs them. It just works.
Decentralized cloud storage is quite similar in concept. We have the technology to provide encryption, resiliency and incredible performance, all while utilizing disparate nodes throughout the world to provide the storage capacity.
Myth #2: If a Storage Node Goes Offline I Could Lose Data
Lots of people worry about the challenges of keeping a Node online. We hear it all the time, “What if someone turns the computer off? What happens to my data?”
Within a decentralized network, there are multiple measures taken to ensure no data will be lost if a Node goes offline.
- Auditing: Storage nodes are constantly audited to make sure they are storing the data they are supposed to be storing.
- Incentives: There is an incentive structure for node operators that is designed to elicit trusted behavior from storage nodes.
- Detection: If malicious behavior occurs, it is easily detected by the system.
- Redundancy: A layer of redundancy is built in using erasure coding so data can’t be lost. When an upload occurs a file is broken up into a number of pieces. To rebuild the file, you only need 29 of 80 pieces. So at any given time, up to 51 nodes could be offline and the file would still be available.
- Math - The probability that 51 Nodes out of tens of thousands would go offline at one time is astronomically low. So even if one node goes offline your data's still available.
- Restoration: The system keeps track of nodes that do go offline and even without the encryption keys, data can be healed. Those pieces can be recreated and redistributed to new storage nodes.
In summary, decentralized storage is designed to ensure that files will always be available.
Myth #3: Storing Data in People’s Homes Can’t be Safe
We often think about physical location data storage, but computers are connected to the internet and are readily reachable from everywhere. When there are vulnerabilities in firewall software or edge perimeter software, there is always a risk that anyone can reach your data.
People think about where their data is being stored (like it being stored by their shady neighbor) when really they should be thinking how their data is stored (is it encrypted, encoded, redundant?). This matters so much more than the physical location.
Ironically, centralized cloud storage is quite prone to attacks because of its lack of encryption and because you have to trust the provider. Decentralized networks create zero knowledge, zero trust environments where the host couldn’t access the data even if they wanted to.
Myth #4: Distributing the Data Must Make it Slow
Another common myth is that decentralized cloud storage is slow. This misconception comes from the idea that if you have to distribute all these pieces all over the network, retrieve them and encrypt them, it must be slow. The reality is that decentralized cloud storage is amazingly performant.
Instead of downloading a single file from a single location and serializing that transfer, what you get is this parallelism where an endpoint can download the fastest 29 pieces from 80 locations, in parallel. This means there are no bottlenecks. Your data is only limited by the bandwidth where the actual download is occurring. If you're in a high bandwidth area, you can really get massive performance out of that parallelism.
Myth #5: The Decentralized Model is too Complex to Work With
When people first learn about decentralized cloud storage, they tend to see it as incredibly complex — almost like a Rube Goldberg version of cloud storage. The fact is, a lot of the early decentralized projects were quite complicated to utilize. With advancement, comes simplification.
At Storj, we’ve been highly focused on building a decentralized cloud storage system that delivers an incredible user experience for developers. We've made it super easy to get all of these advantages of decentralized cloud storage, the enhanced privacy, the enhanced security, the enhanced performance, and the great economics — without creating a huge layer of complexity. We're S3 compatible. We take credit cards. Fundamentally, we’ve eliminated the barriers and made it comfortable and risk-free to get started with a free trial.
While it takes a lot to create a decentralized cloud storage network like ours and keep it running, from the developer standpoint, it's very easy to use and very easy to get started.
The Future of Cloud Storage is Decentralized
Decentralized storage is viable and available for developers to use for object storage. It excels at backing up large data sets from academic research to autonomous vehicles, point-to-port file transfer for large files, media serving and software distribution to name a few use cases. Actually using decentralized cloud storage is the best way to bust any lingering misconceptions you may have. So, start your free trial and see for yourself why decentralized cloud storage is a fantastic reality.
This article originally published at The New Stack.