Goal: ability to make pretty graphs locally with both credit mining and multichain (two different branches).
Task: make an Ubuntu channel or other large validated legal collection, with actual "from the wild" downloads.
Problem Description (7 pages) 2.1 Challenges in creating credit mechanisms 2.1.1 Gain return and performance by spending credit 2.2 Prior Credit Mining Research 2.3 Prospecting good investment (Research Question 1) 2.3.1 Credit mining as investment tool 2.4 Substituting investment cache (Research Question 2) 3.
Credit Mining System Design (11 pages) 3.1 Libtorrent 3.1.1 Priority Management 3.1.2 Share Mode 3.1.3 Peer Discovery 3.2 Credit Mining Architecture 3.2.1 Mining Sources 3.2.2 Resource Optimization 4.
"5.4 Finding best parameters", more text for experimental section Please add details: 13000 torrent are downloaded for 1 hour each in batches of 500. 59%(8176) are OK when only downloading first 4 pieces, a mere 5%(668) for rarest first.
Investigate what goes wrong with those 7508 swarms.
The goal of this research is to devise an automatic caching layer which ensures content availability for both long-lived years-old and freshly created content mere seconds ago.
Current 2.1 and 2.2 have too much overlap with Section 1. Move all related work and citations to a dedicated section. uses Bit Torrent protocol to increase user download speed and at the same time reduce datacenters load." Section 2.4.1 needs to move to either 1 or 3, too much engineering details for 2.
Strategy: Till end of Feb focus on credit mining, then include relaying bandwidth to avoid too much complexity at the start. Next sprint: accurate investment information, torrent_integration, discovered swarms, downloaded torrents, DHT lookup failure, failed to download .torrent, used bandwidth for checking, swarms connected to, dead swarms, seeder/leecher info determined. Also problem to TCP check trackers, as tunnels only support UDP.
Explain everything at a deeper level, for instance: Comments on this thesis version Please focus on readability and understandability of results section.
Part-time: API in the core swarm stats credit earning stats.
Our first validation experiment tests the basic credit mining swarm selection algorithm.
We create a dozen swarms, all a different number of seeders.