ForgeCDN is a high-tech product, grounded in combination of breakthrough technologies for content delivery and blockchain and cryptocurrencies operations.
The ForgeCDN network utilizes decentralized capacities of the network participants working for the reward. It is called fog due to its topological peculiarity: the nodes of the system (computers, laptops, tablets, etc.) are literally everywhere; the size of these capacities is relatively small in comparison to general CDNs, and reminds micro-droplets that are suspended in the air causing fog. Our software and hardware control this fog, allowing us to use it as a homogeneous environment.
Blockchain, Ethereum and smart contracts
Ethereum is a platform for decentralized online services creation operating on the blockchain basis and working with smart contracts. ForgeCDN smart contract of this platform is programmed to transfer funds from content providers to system participants and developers only after fulfilling their obligations to transfer content to the end-users, which gives the guarantee to all the participants.
High degree of content protection
We use a symmetric algorithm of block-based content encryption AES 128, accepted as a standard for encryption by the US government. On the one hand content is stored in encrypted form at user side, which does not allow to dispose the content by participant at its own discretion, protecting the rights holders, on the other hand, it protects the network participants from accidental transfer of illegal content (in case content-provider managed to bypass our system of legality of content tracking).
The network uses smart managers of operational and disk memory for network members. This technology allows increasing the speed of delivery, placing the most popular content in RAM and less demanded to store on the hard drive. Additionally, managers can reduce the load on the hard drive, which significantly reduces system participant's equipment depreciation.
The speed of the system
Core of the system is written in C ++ programming language, which can combine the properties of both high-level and low-level language. It allows optimizing the network in such a way that even at 40 GB output data from one device (which is equivalent to 40,000 simultaneous connections with an average speed and corresponds to the average video quality broadcasted in the Internet), for instance processor Core i7 gets loaded only by 10%. Such a low consumption of the processor's resource will allow using the equipment for its intended purpose while simultaneously connecting to the ForgeCDN network. In addition, it allows one to connect devices with weak processors to the network.
One of the unique features of the system is the presence of artificial intelligence. The algorithm is in a constant process of self-learning, and its main function is to redirect users requesting content to the participant's CDN on the basis of 36 parameters, such as geographical distance, channel bandwidth, cached data in RAM and so on. Thus, the algorithm allows one to deliver content to the end-user in the shortest possible time.
Load balancing, fault tolerance
The traffic balancer constantly monitors the conditions of network participants and redistributes the priority of routing in conjunction with the algorithm of artificial intelligence. Core redirects end-user to network node in accordance with the weight assigned by AI. The system keeps track on cases of undefined participant and translates it by distributing traffic. Decentralization at the network architecture level contributes to fault tolerance increase. And the failure of one of the kernel servers only leads to automatic redirection of nodes and content consumers to the nearest working kernel server.
We developed a rating system for the network participants, which is one of the criterias for selecting them as the final node for the transfer of content on demand. Artificial intelligence takes into account this criterion, together with many others, and thus controlling the balancer and increasing the possibility of getting a reward to the most stable participant.