Shared Object Networking and IPFS

A few years ago, I worked on the idea of Shared Object Networking.  At the time, it was too different from the then predominant uses of the internet to be useful.  With the wider adoption of block chains and post-CDN systems, like IPFS, the time may have come to revisit this idea as many of the underlying prerequisites have emerged as adoptable standards.

The basic idea of SON was to get away from existing internet search models, like Google, which treat the internet as a massive salt mine.  You go in and search around, get millions of hits, and try to find the best one for your needs.  Google has made an industry of improving and gaming these results.

SON uses a different approach.  It thinks of everything as an object.  These objects have inherent relationships which can be observed and navigated.  Inherent in any SON type system is the ability to store and consume metadata about objects, and therefore an object model is a precursor to codifying data into objects in a useful way.

In several proprietary formats, we see this today.  A simple example is the ID3 tag of an MP3.  It contains the performer, title, publish date, and other aspects of the file that are relevant to anyone wanting to find a specific file, artists, genre, etc.  In a more mature implementation of this, these object containers have a more standardized protocol for declaring metadata and a published schema capability to allow for applications to more readily consume that data.  JSON and XML seem like likely targets for this.

Akin to some prior articles on crypto-economics, these objects are likely to occasionally become well known.  That is, one can use object anchors to tie to well-known objects so that future inferences are easier to make.  As an example of this, you could create a ‘person’ object instance of ‘Miles Davis’ which could be connected to an MP3 object instance of ‘Bitches Brew’.  These connections between well-known objects allow for reusability and inheritance relationships.

The IPFS and crypto-economics uses for this is to move object identification into a hash function and store on a system like IPFS.

The idea behind SON that makes these underpinning technologies interesting is the ability to consume objects through a browser.  In essence, to browse objects and their inherent relationships.  While objects and their direct relationships can be considered on a two-dimensional axis, the relationship model works best on the z-axis.  For example, a search layer like Google might be a good ‘layer’ in which to browse object relationships.  A better and more relevant example for my prior post on cyber reputational scoring would be the blockchain.  Individual chains could be consumed as transactions on the chain that relate to specific well-known objects.

In this way, blockchain exploration and computational analysis on account references can be more easily made.  In the cyber reputational risk model, cryptocurrency inferences are made by who accounts transfer monies between.  In a wide SON implementation, these macro identities are attached to specific objects within the overall model at an identity level.  Creating a single navigable model creates an analysis layer that would allow for wider inference and algorithmic scoring.

Outside of the business intelligence usability of this system, there is a ‘new internet’ function here that cannot be understated.  The ability to consume information as objects allows for intelligent navigation between content types and providers in a logical, intuitive, and user-friendly model.  This browser approach can be the underpinning for all application models.

The underpinning technologies that are missing is macro/micro identities to create object relationship and privacy models, a well-known object schema including content objectification and publication to move legacy content to the IPFS, and policy modules for content consumption which both facilitate how the content is consumed and how the policies that are attached to that consumption are enforced.

The last, but a probably most transformational aspect of SON, is the object browser.  Certainly, application developers will more immediately see and embrace the programmatic value of implementing a SON-based internet; but once users can consume the internet as objects in a friendly manner, this will drive critical adoption that will result in readdressing the fundamental flaws in the current internet architecture.

Leave a Reply