Phenomenological Model for Distributed Systems
Info and Philosophy

The Common Model for Distributed Systems

Distributed systems provide many advantages over monolithic systems. Some of the advantages of using distributed systems include concurrency, performance, communication, resources sharing, economies of scale, reliability, and incremental growth.

A distributed system contains physical entities called nodes. On each node, any number of objects can reside. An object is functional unit that has some purpose such as manage data or perform a calculation.

Currently, the common model for designing a distributed system of this sort, is to view the objects from a data or functional centric viewpoint. Basically, we design an object by deciding what it knows, what it wants to know, what it can do, and what it wants to have done. We purposefully omit the questions of how it will tell you what it knows and how it will ask you what you know. And we purposefully pay no attention to how you ask it to do something or how it asks you. That is not to say that anyone can ask it to do something, there is authentication, but an authenticated node can always get the object to do what it wants no matter how it asks.

But what about in a distributed system where the surrounding nodes are unknown and all you can tell about a node is over what channel they made their request? This might be a known intimate communication channel and therefore you would want to accept this new node's request.

The common model has used these omissions to gain platform independance, communication protocol transparency, portability, and many other advantages. However, in some distributed systems these are not desirable attributes.
This project seeks to design a model for distributed systems where not only does it not omit this information, but it exploits it to design new and powerful systems.

The Senses (Interface Channels) are at the Forefront

We (The Objects) Are Only What we Experience

Phenomenology

20th Century philosophical movement dedicated to describing structures of experience as they present themselves to consciousness.

Phenomenologists tend to oppose the acceptance of unobservable matters and grand systems erected in speculative thinking.

It is a rejection of Cartesian dualism where there exists a separation of the mind and the body, and believe that the mind exists only in response to the senses.

Phenomenological Model for Distributed Systems

This project seeks to delve into the use of Phenomenological thought to model distributed systems and provide tools to implement these systems.

This will lead us down the path of designing objects whose "mind" exists in response to its "body". It has immediate applicability to sensor networks.

By approaching the process of modeling systems in this way, the standard designed becomes a work of art and the systems developed using this standard parallel human social existence.
Home | Info | People | Documents | Software Tools | Hardware Platforms
Example Code | Applications | Media | Internal

Responsive Environments Group
MIT Media Lab