Decentralized Distributed Incentivized Computing

in #obyte5 years ago (edited)

If you know about BOINC or been following Obyte's standing in the THOR challenge then this should be familiar to you.

BOINC is a distributed computing system which scientists and data analysts use to break a large computing problem down into bite size tasks and then distributing these tasks to anyone who wants to contribute; the results are collected in a centralized server which are valuable to scientists. Usually the computing problems are related to simulating molecule interactions, signal detection, mathematical curiosities and etc...

OBOTIC is a decentralized distributed incentivized computing platform

🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖 🤖

I think that's really cool, but here are just some advantages over BOINC:

  1. Low maintence, set and run indefinitely
  2. Open transparent, anyone can join, no one can cheat
  3. Incentivized computing, passive income while you sleep

To contribute and earn income, a Provider / Cruncher runs a nodejs obyte headless wallet which listens for computing jobs using the Obot Autonomous Agent.

For Scientists / Data Analyst, this system is most suitable for computing problems which takes little input data or a link to a big data set, then do complex or private computation. Real world examples are science data analysis like SETI or WCG or Zooniverse.

Since there is fee involved, prices can be set by crunchers based on the difficulty of the computation, and payments are made automatically once result is submitted. This avoids a major drawback in the BOINC system preventing mass adoption, since no one gets paid for contributing to their network.

This also means that paid private data computation is possible, where people hold on to their own data and only send results of calculation. For example, you might keep a log of where you been traveling throughout the day; a data analyst might want to calculate your and many users' total travel distances. In this case, using Obotic, your travel patterns are not released, but useful information can be sold to the data analyst. That's powerful right?

For computation results which requires verification, Obotic has a built in system to handle it. Which is a new AA technique I have developed as well, which I call RANK AND RECORD. Results which are dynamic and not pre-determined act as votes of themselves, so the best result is the one with the highest vote. All this data is recorded on the Obyte DAG.

Overall, the system uses 4 prior AAs which I developed since the first AA contest, see below diagram. Being the only contestant who has entered into every single contest, I can say this entry is the one I am most proud of, with ramification that is quite world changing.

IMG_20191024_181622.jpg

To test the AA, using a testnet wallet:

  1. Send 100,000 to this Obotic AA address UHW5KN3OR2VB6BZGCX7SC7MASZWC77L2 and
    Also send data... Field = "args" Value = "[ 7, 1, 3, 2 ]"
    Like this:

    This will create a new processing round and choose 3 Obots
    It should look like this:

  2. Take note of the round number in the response as you will need that in the next step. Distributed computation has already commence, need to wait some time (30 mins) for things to settle.

  3. Check the result of the distributed computation by
    Also send data... Field = "check" Value="<above round number>"

The demo should bounce response which show a sorted array "[ 1, 2, 3, 7 ]".

For data analyst /scientist, I am interested to hear what difficult large scale computational problem you have in mind, and we can set something up.