How Our Tech Works.

Polycloud is an innovative first-in-class distributed storage platform. It’s more affordable than traditional cloud data management and provides increased protection from events that disrupt traditional data centers.

Cloud object storage comparable to Amazon’s S3 object storage, Google Storage, or Microsoft Azure Blob storage.

Offers secure and private data storage with immediate access and high reliability.

Stores data as objects within buckets that can be accessed and managed via our basic browser based user interface or full functionality through our S3 compatible API.

We’re Not Your Typical Cloud Storage Provider

First, let’s compare us to a single site data center.

In a typical cloud data storage solution your data ultimately sits inside a data center. Data centers are large buildings full of servers and hard drives that are connected to the internet.

Data centers are very expensive to build, operate, and maintain.
Require large amounts of real estate, numerous buildings, huge amounts of power to not only operate the building and equipment, but also to keep all the equipment cool.
Use redundant infrastructure such as multiple backup generators and onsite fuel repositories to ensure that the risk of failure is minimized.

Overall, this infrastructure equates to large operational costs and large environmental impact. Which is why we try to minimize the amount of traditional data center infrastructure we utilize with Polycloud.

Our Way. A Better Way.

Instead of using expensive hardware, Polycloud uses software solutions to create a better storage system with cloud data management.

One.

Polycloud’s distributed network combines un-utilized storage and bandwidth resources that are already deployed and connected to the internet. Each storage device on the distributed network becomes a distinct node. We are currently utilizing a network of over 250,000 NAS devices in consumer homes and businesses, with a combined capacity of over 400 petabytes.

Two.

A virtual data center is constructed through these nodes. They are distributed geographically and provide durability and reliability that is much more than with individual components.

Three.

Because of the geographic distribution, our solution is protected from any outage caused by localized geographic events such as fires, floods, power outages, earthquakes, hurricanes, tornadoes, Internet outages, or even more malicious activity such as sabotage or terrorist attacks.

Four.

The result is a more cost effective solution that largely mitigates the environmental impact of our storage system. Our network inherently operates with dozens of independent power utilities and internet service providers. By using our software to create a distributed virtual data center, we are able to create all the benefits and more of a traditional data center at a fraction of the cost.

The Distributed Data Storage Process

It’s a three-step process that ensures your cloud data management is secure and safely guarded.

In a typical cloud data storage solution your data ultimately sits inside a cloud data management center, a large building full of servers and hard drives connected to the internet. Polycloud sometimes uses regionally distributed data centers to temporarily store data, but ultimately the data is stored on our distributed network. The three steps below are the process used to store data on our distributed network.

Encryption.

When an object is sent to Polycloud to be stored on our distributed network, we first encrypt the object with xsalsa20poly1305 encryption (configurable to others if needed). If it is larger than 64-128MB it is also broken down into smaller blocks. Each object is erasure encoded to break the encrypted object into multiple different pieces (typically 40) of which only a subset (typically 20) is needed to completely reconstruct the object.

Distributed to be Stored.

Each of these encrypted pieces is then stored on a different node within the network and repeated for each block of data, which constitutes the entire object. Because each block uses 40 nodes, that means a 5GB object will be stored on thousands of nodes, greatly increasing its durability and availability.

Meta-Data Capture.

The information about the location of each piece for each object is tracked, stored, and encrypted in a database in a public cloud storage provider, with multiple redundant backups. Because this metadata about the pieces is orders of magnitude smaller than the data itself, traditional techniques for multiple datacenter redundancy are more feasible and affordable than it would be for the data itself.

cloud object storage. 1. files are encrypted and split into multiple pieces 2. encrypted pieces are stored on nodes in a geographica.ly distributed network and 3. only a subset of the pieces that are stored are needed to retrieve the fil

The Data Retrieval Process

When an object is retrieved, each of the 40 pieces of that object are called for and streamed back in parallel.

Polycloud then retrieves the fastest 20 pieces and reconstructs the file, decrypts it, and delivers it back to the end user. Because we are able to download the objects in parallel, and because we only require 20 pieces to reconstruct the whole file, our system is able to retrieve data at very fast speeds.

How Is It So Fast?

Larger objects are broken down into smaller blocks, these speeds scale even as the size of the object increases. Of the 40 pieces stored we only need to download the fastest 20 pieces in parallel from the network to restore the object.

In addition to the built in redundancy of the storage platform (storing 40 pieces of data but only needing any 20 to reconstruct each object), the durability of the data is ensured through our active repair mechanism. Each day every object in the network is examined and the number of available pieces of each object is assessed.

How Does Cloud Object Storage Redundancy Work?

If the number of available pieces falls below the threshold where we can be certain that the object will be safe, then we will actively insert additional pieces of the object onto other nodes in the network, thereby keeping the number of available pieces well above the minimum of 20 that is required to reconstruct the file.

Cloud Data Management Safety

Polycloud uses state of the art xsalsa20poly1305 encryption both in transit and at rest to ensure that your data is safe and secure.

Your data is spread out across our distributed network. With our 20/40 encoding scheme a malicious attacker would need to physically access 20 different nodes within our network of almost 300,000 devices.

They then would have to break the state of the art encryption of each individual piece of the object.

Then even if they were successful to this point they would have to repeat the process with an entirely different set of 20 devices in order to obtain a second file of your data. In this way the distributed network creates an added layer of security because unlike a data center there is no centralized repository of data for a bad actor to target.

The Durability of Polycloud

Because of the geographically distributed architecture of Polycloud we are able to prevent the loss of data more effectively than solutions which are housed within data centers within a single geographic region.

Designed to provide better than 11 9’s (99.9999999%) durability. This translates to on average losing less than 0.000000001 % of objects a year, or in other words if you store 10,000 objects with Polycloud you should lose less than one object every 10,000,000 years.

Get Started with Polycloud.

Getting set up with Polycloud is easy

Head over to our registration page to sign up for an account. Once your account is created, you can use our basic web interface to create buckets as well as upload, download, and manage your files. If you would rather use an approved 3rd party application to manage your data storage use cases, you can obtain the necessary credentials using the same web interface. This will allow you to use our S3 compatible API to connect your 3rd party application to Polycloud.