Why The InterPlanetary File System & Its Ilk Are A Big Deal For Blockchains & Beyond

in etc •  2 years ago


The Internet is amazing but it is not as good as it can be. It is several decades old and was designed for a different environment. It does not provide adequate protections and support for the ways people actually use it:

  • Security - Security was not a major design goal of the Internet. Data is vulnerable to hardware failures, censorship and attackers. Even today tools like Secure Sockets Layer (SSL), Pretty Good Privacy (PGP) and Domain Name System Security Extensions (DNSSEC) are not widely used. Security is optional :


  • Mobile Devices - The Internet was designed to allow static computers to effectively communicate. Movement requires new wireless connections to be constantly set up. Furthermore, mobile device marketshare is growing :


  • Broadcasting - The Internet was designed to allow two computers to effectively communicate. That is great for tasks such as sending email. However, today the majority of Internet traffic is broadcasting. Examples include Netflix and YouTube videos:


Glimpses Of A Better Internet


Virtual private networks (VPNs) and Tor provide better security. Cell phone networks provide better support for mobile devices. BitTorrent provides better support for broadcasting. These give an idea of what a future improved Internet can look like.

The InterPlanetary File System


The InterPlanetary File System (IPFS) is one noteworthy attempt to replace the current Internet. Security is built in by default. BitTorrent like broadcasting gives it better performance. Duplication of data across several computers provides censorship resistance and protection against hardware failures. There is even ongoing work to add support for mobile devices.

The Internet Protocol (IP) addresses devices. A major change in IPFS, relative to the current Internet, is to focus on addressing data. Notice that as long as copies of data are accurate, sources are irrelevant. Furthermore, IPFS addresses are based on cryptographic hashes making information virtually impossible to tamper with.



Blockchains are not designed to effectively store vast amounts of data. This is why blockchain systems can greatly benefit from systems such as IPFS. Imagine a video smart contract, on Ethereum or Ethereum Classic, for which it would be prohibitively expensive to store the required videos on the blockchain. It would make much more sense to store the videos on IPFS and just store the corresponding IPFS addresses on the blockchain.



There are efforts underway to greatly improve the Internet. These efforts will make the Internet faster, more secure and more mobile friendly for all of us. Blockchains will be able to effectively utilize these new architectures to allow the creation of many innovative applications in the near future. Stay tuned as it should be a wild ride.


Feel free to leave any comments or questions below. You can also contact me by clicking any of these icons:

twitter facebook linkedin


I would like to thank Van Jacobson for his brilliant Named Data Networking related papers and talks that have greatly helped clarify many of these ideas. I would also like to thank IOHK (Input Output Hong Kong) for funding this effort.



This work is licensed under the Creative Commons Attribution ShareAlike 4.0 International License.

Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!
Sort Order:  

What type of hardware and protocol changes will be required? Will it run on top of existing hardware and protocols using existing layers? Will it be like adding a new Layer 3 protocol? Will it be like IPv6 and require new hardware? IPv6 does quite a bit more than IPv4, but people still have seemed to be rather loathe in giving up on IPv4. So how it is implemented and what requirements to make it work are there will be a big part in determining its attractiveness and adoption.


Happy to help. IPFS runs on top of existing Internet hardware and software layers without any changes required. So, I would think adoption would continue apace.

Think of how the Internet was originally a new layer on top of the old telephone network, yet, doesn't //require// the old telephone network. It is conceivable in the future companies will eventually sell "IPFS optimized" hardware and TCP/IP might fade in importance and maybe even get replaced.

Can you add links to your Acknowledgements? I'd love to know more.


Sure. I've watched this Van Jacobson talk many times. I can't recommend it enough...


See also https://en.wikipedia.org/wiki/Named_data_networking for more info and links. Regards.