From: Brian Warner Date: Tue, 3 Jun 2008 01:38:32 +0000 (-0700) Subject: move historical docs from wiki pages into the source tree, clearly marked as historical X-Git-Url: https://git.rkrishnan.org/pf/content/-?a=commitdiff_plain;h=aa2c6937641f857fe29efa3236bada7df60fcdba;p=tahoe-lafs%2Ftahoe-lafs.git move historical docs from wiki pages into the source tree, clearly marked as historical --- diff --git a/docs/historical/peer-selection-tahoe3.txt b/docs/historical/peer-selection-tahoe3.txt new file mode 100644 index 00000000..0cd88ceb --- /dev/null +++ b/docs/historical/peer-selection-tahoe3.txt @@ -0,0 +1,66 @@ += THIS PAGE DESCRIBES HISTORICAL ARCHITECTURE CHOICES: THE CURRENT CODE DOES NOT WORK AS DESCRIBED HERE. = + +When a file is uploaded, the encoded shares are sent to other peers. But to +which ones? The PeerSelection algorithm is used to make this choice. + +In the old (May 2007) version, the verifierid is used to consistently-permute +the set of all peers (by sorting the peers by HASH(verifierid+peerid)). Each +file gets a different permutation, which (on average) will evenly distribute +shares among the grid and avoid hotspots. + +This permutation places the peers around a 2^256^-sized ring, like the rim of +a big clock. The 100-or-so shares are then placed around the same ring (at 0, +1/100*2^256^, 2/100*2^256^, ... 99/100*2^256^). Imagine that we start at 0 with +an empty basket in hand and proceed clockwise. When we come to a share, we +pick it up and put it in the basket. When we come to a peer, we ask that peer +if they will give us a lease for every share in our basket. + +The peer will grant us leases for some of those shares and reject others (if +they are full or almost full). If they reject all our requests, we remove +them from the ring, because they are full and thus unhelpful. Each share they +accept is removed from the basket. The remainder stay in the basket as we +continue walking clockwise. + +We keep walking, accumulating shares and distributing them to peers, until +either we find a home for all shares, or there are no peers left in the ring +(because they are all full). If we run out of peers before we run out of +shares, the upload may be considered a failure, depending upon how many +shares we were able to place. The current parameters try to place 100 shares, +of which 25 must be retrievable to recover the file, and the peer selection +algorithm is happy if it was able to place at least 75 shares. These numbers +are adjustable: 25-out-of-100 means an expansion factor of 4x (every file in +the grid consumes four times as much space when totalled across all +StorageServers), but is highly reliable (the actual reliability is a binomial +distribution function of the expected availability of the individual peers, +but in general it goes up very quickly with the expansion factor). + +If the file has been uploaded before (or if two uploads are happening at the +same time), a peer might already have shares for the same file we are +proposing to send to them. In this case, those shares are removed from the +list and assumed to be available (or will be soon). This reduces the number +of uploads that must be performed. + +When downloading a file, the current release just asks all known peers for +any shares they might have, chooses the minimal necessary subset, then starts +downloading and processing those shares. A later release will use the full +algorithm to reduce the number of queries that must be sent out. This +algorithm uses the same consistent-hashing permutation as on upload, but +instead of one walker with one basket, we have 100 walkers (one per share). +They each proceed clockwise in parallel until they find a peer, and put that +one on the "A" list: out of all peers, this one is the most likely to be the +same one to which the share was originally uploaded. The next peer that each +walker encounters is put on the "B" list, etc. + +All the "A" list peers are asked for any shares they might have. If enough of +them can provide a share, the download phase begins and those shares are +retrieved and decoded. If not, the "B" list peers are contacted, etc. This +routine will eventually find all the peers that have shares, and will find +them quickly if there is significant overlap between the set of peers that +were present when the file was uploaded and the set of peers that are present +as it is downloaded (i.e. if the "peerlist stability" is high). Some limits +may be imposed in large grids to avoid querying a million peers; this +provides a tradeoff between the work spent to discover that a file is +unrecoverable and the probability that a retrieval will fail when it could +have succeeded if we had just tried a little bit harder. The appropriate +value of this tradeoff will depend upon the size of the grid, and will change +over time. diff --git a/docs/historical/peer-selection.txt b/docs/historical/peer-selection.txt new file mode 100644 index 00000000..0d16c3bb --- /dev/null +++ b/docs/historical/peer-selection.txt @@ -0,0 +1,28 @@ +When a file is uploaded, the encoded shares are sent to other peers. But to +which ones? Likewise, when we want to download a file, which peers should we +ask for shares? The "peer selection" algorithm is used to make these choices. + +During the first tahoe meeting, (actualy on the drive back home), we designed +the now-abandoned "tahoe1" algorithm, which involved a "cabal" for each file, +where the peers involved would check up on each other to make sure the data +was still available. The big limitation was the expense of tracking which +nodes were parts of which cabals. + +Formerly (until v0.6, ticket #132), we used the "tahoe3" algorithm (see +peer-selection-tahoe3.txt), but now we use the "tahoe2" algorithm (see the +PEER SELECTION section of docs/architecture.txt), which uses a permuted +peerid list and packs the shares into the first 10 or so members of this +list. (It is named "tahoe2" because it was designed before "tahoe3" was.) + +In the future, we might move to an algorithm known as "denver airport", which +uses Chord-like routing to minimize the number of active connections. + +Different peer selection algorithms result in different properties: + * how well do we handle nodes leaving or joining the mesh (differences in the + peer list)? + * how many connections do we need to keep open? + * how many nodes must we speak to when uploading a file? + * if a file is unrecoverable, how long will it take for us to discover this + fact? + * how expensive is a file-checking operation? + * how well can we accomodate changes to encoding parameters?