[smart (OP)]

I just stumbled upon this: https://ipfs.io/

"A peer-to-peer hypermedia protocol to make the web faster, safer, and more open."

[chris]

Can you summarise what IPFS is for us?

[smart (OP)]

Here's the abstract from the latest draft specification of IPFS:

ABSTRACT
The InterPlanetary File System (IPFS) is a peer-to-peer distributed le system that seeks to connect all computing devices with the same system of files. In some ways, IPFS is similar to the Web, but IPFS could be seen as a single BitTorrent swarm, exchanging objects within one Git repository. In other words, IPFS provides a high throughput content-addressed block storage model, with content-addressed hyper links. This forms a generalized MerkleDAG, a data structure upon which one can build versioned file systems, blockchains, and even a Permanent Web. IPFS combines a distributed hashtable, an incentivized block exchange, and a self-certifying namespace. IPFS has no single point of failure, and nodes do not need to trust each other.

https://github.com/ipfs/papers/raw/master/ipfs-cap2pfs/ipfs-p2p-file-system.pdf

[evan_au]

New protocols are proposed all the time. Some of them even get widely used (you are probably viewing this over IPv4 or IPv6 protocol).

Internet protocols generated by the IETF are listed here: https://tools.ietf.org/rfc/  - today I count 8128 of them. And that is just one standards organization!

I subscribe to a journal where perhaps 25% of the articles are proposing new protocols, or extensions to existing ones. And another 25% are explaining the behavior of existing protocols. 

There have been numerous proposals for networks that retrieve data by content. They tend to have a problem that you need a fairly long "address" to uniquely specify any piece of data in the world  - or any piece of data that could potentially be generated in the future, on or off the Earth! (This includes the conversation you are having on your mobile phone). IPv4 used 32-bit addresses, and IPv6 uses 128 bit addresses, while some of the content retrieval proposals have content descriptors that are 256 or 512 bits long.

That is a somewhat heavy tax on all communications, especially the very short packets used to carry speech.

[jeffreyH]

Many years ago I came across an indian university website that had an article called roll your own internet. It told you how to send raw connection packets to your isp to sign in and then how to send udp and tcp packets. I downloaded the example source code and was communicating with the internet within a day or so. That was very simple to achieve once explained. As things progress it becomes progressively more difficult to achieve the same results.

[smart (OP)]

IPFS, the future of healthcare IT using content-addressed storage (CAS) ?

http://www.cio.com/article/3171132/innovation/an-ipfs-addressable-storage-model-for-healthcare-with-blockchain.html

[evan_au]

the future of healthcare IT

Health care records are a challenging subject.

• Every week there are new medical discoveries, new medications and devices approved, and new recommendations on how best to use them.
• While half of your doctors entries would have been recognized by the same doctor 20 years ago (blood pressure, weight, smoking status, pain in the right knee, etc), many of the medications, medical tests and government-dictated codes would be new.
• It only takes one MRI image to add a gigabyte or more to your medical file, which dwarfs the hundred or so bytes manually added by your doctor at a typical visit.
• Once DNA analysis becomes affordable, a large fraction of everyone's medical records from birth (or before) will consist of DNA sequences, which would be incomprehensible to most doctors today.

While Content-Addressable storage may address retrieval aspects, it doesn't address backward-compatible extension and data interpretation in a very dynamic database schema, let alone importing and accessing previous unstructured data sources from 20 years ago.

[smart (OP)]

While Content-Addressable storage may address retrieval aspects, it doesn't address backward-compatible extension and data interpretation in a very dynamic database schema, let alone importing and accessing previous unstructured data sources from 20 years ago.

IPFS is being built on top of IPV4/IPV6. Consequentially, it should be backward-compatible with the "old" HTTP protocol.

[evan_au]

IPFS is being built on top of IPV4/IPV6. Consequentially, it should be backward-compatible with the "old" HTTP protocol.

One of the problems of IPv4/IPv6 is that the address implies a geographical location, which is not ideal if people are always on the move, and the best data storage location is always changing. So IPv4/IPv6 transport is considered merely a transition phase by supporters of content-addressable data.

But by backward compatibility, I was referring to the application data, not the transport of the data.

Despite great strides in standardization of medical imaging formats, some medical data was created before file standards were developed, and standards tend to lag behind the newest technologies. So some old data and some new data may not be readable in an electronic health records system.

[smart (OP)]

Democratizing the web through a decentralized (peer-to-peer) architecture is how the Internet should delegate content-addressing and authority to the users.