While we love talking about what the NDN Protocol has achieved so far and it’s practical applications in the real world, we also are eager to know what further developments are going to be made and what use cases we can expect in the nearest future.

This brings us to one of the recent researches that were done by eight computer scientists from University of London, the University of California, Los Angeles, and Osaka University in Japan. The research paper that was published by those scientists on September 25, 2019 talks about a proposal of a brand-new IPFS data retrieval algorithm.

Their whitepaper provides examples of how the NDN Protocol can increase IPFS’s data retrieval capabilities by more than 20 times.

The main focus of the study was on video content retrieval as a result of the increasingly growing demand on the market for such content. A comparison study of the current TCP/IP Infrastructure was done with the ICN/NDN Networks to get the performance data of each technology and conduct a thorough analysis as to the effectiveness of NDN in P2P data retrieval.

Consequently, the analysis numbers have shown that the effectiveness of IPFS data retrieval without the NDN Protocol was significantly low. So the scientists proposed a solution in the form of a Satisfied Interest Table for adaptive routing algorithm, which eventually increased the retrieval speed of existing IPFS networks by more than 20 times.

Why was there a research regarding NDN and IPFS in the first place? Couldn’t we just stick to the way things are now with the current Internet (TCP/IP)?

Over the last 5 years the global IP traffic has tripled. This increase was mainly due to the demand for video content. According to statistics the increase of video content is not going to be decreasing anytime soon, moreover, it is forecasted that it will only double by 2020.

With that demand, data consumption becomes more expensive because it drives the cost for maintenance and operation for the current infrastructure up, requiring more storage space and more speed for content delivery.

Content Delivery Networks (CDN) play a huge role in content delivery to end-users. While the producers of content are willing to pay higher fees to distribute their content, CDNs on the other hand, are looking for ways to drive the cost of delivery down, moreover they are constantly looking for ways to optimize the delivery to increase speed.

As a result Peer-to-Peer (P2P) networks have an advantage over CDNs, because they drive the cost of content distribution down.

The benefit of a P2P network is the way the data is being stored and distributed. Each piece of content is stored on a node nearby for fast retrieval. As opposed to CDNs where content is stored on centralized servers. But the limitation of a P2P network is in the fact that it’s mainly used for downloading large static files and it doesn’t not have a commercial use case.

That’s why projects like IPFS/Filecoin were born to give a P2P network a commercial use case through the distributed storage concept. Distributed storage provides decentralized data storage and data retrieval capabilities through distributed ledger and cryptography.

The IPFS interplanetary file system (IPFS) is a content-centric design similar to ICN and NDN networks. It maps the CID of the content to the host through a self-verification content hash, and human-readable names. In addition, IPFS provides data block transmission services through the BitSwap protocol, and rewards resource providers through FileCoin.

But the interest of scientists in this research was if P2P has an advantage over CDN and IPFS has an advantage over P2P, can NDN make IPFS even better?

Yes, it can!

In the paper, scientists conducted an experiment to verify the performance difference for video content retrieval between a raw IPFS system and an IPFS system with the help of an NDN protocol.

For the experiment, a video file with the length of 10 minutes and a resolution of 480p was used. IPFS uses NS-3 to run the IPFS product model. NDN uses the NDN SIM end client. The routing cache function is temporarily disabled. There are three forwarding strategies of NDN: ASF, NCC and best route strategy respectively.

The results shown that the retrieval performance of a pure IPFS system is relatively low, mainly due to fetching address speed issue of the DHT technology used by IPFS. The transmission efficiency of the BitSwap protocol is also low, and it will pass duplicate file fragments at the same time causing loses.

The comparison data shows that NDN-based IPFS has a 20 times improvement in addressing performance compared to pure IPFS; Moreover, NDN-based IPFS has a 20-times increase in throughput compared to pure IPFS.

Judging from this experiment we see that having the NDN protocol for the IPFS system is rather beneficial. This combination would give an adaptive state forwarding, and by using SNAMP it would ensure the security of the namespaces. With the help of SIT routing the system would be able find content in the same area of ​​the network.

This shows that IPFS running on an NDN protocol is far more efficient than IPFS running on the IP Protocol.

This paper shows just a fraction of how NDN could be beneficial for the global network. If you wish to explore other use cases of NDN, you can find them in the NDN Link project.