Gnutella2

In November 2002, Michael Stokes announced the Gnutella2 protocol to the Gnutella Developers Forum. While some thought the goals stated for Gnutella2 are primarily to make a clean break with the gnutella 0.6 protocol and start over, so that some of gnutella's less clean parts would be done more elegantly and, in general, be impressive and desirable; other developers, primarily those of LimeWire and BearShare, thought it to be a "cheap publicity stunt" and discounted technical merits. Some still refuse to refer to the network as "Gnutella2", and instead, refer to it as "Mike's Protocol" ("MP").[2]

The Gnutella2 protocol still uses the old "GNUTELLA CONNECT/0.6" handshake string for its connections[1] as defined in the gnutella 0.6 specifications. This backwardly compatible handshake method was criticized by the Gnutella Developers Forum as an attempt to use the gnutella network for bootstrapping the new, unrelated network, while proponents of the network claimed that its intent was to remain backwards-compatible with gnutella and to allow current gnutella clients to add Gnutella2 at their leisure.

With the developers entrenched in their positions, a flame war soon erupted, further cementing both sides' resolve.[3][4][5][6]

The draft specifications were released on March 26, 2003, and more detailed specifications soon followed. G2 is not supported by many of the "old" gnutella network clients, however, many Gnutella2 clients still also connect to gnutella. Many Gnutella2 proponents claim that this is because of political reasons, while gnutella supporters claim that the drastic changes don't have enough merit to outweigh the cost of deep rewrites.[7]

Gnutella2 divides nodes into two groups: Leaves and Hubs. Most Leaves maintain two connections to Hubs,[8] while Hubs accept hundreds of Leaf connections, and an average of 7 connections to other Hubs. When a search is initiated, the node obtains a list of Hubs, if needed, and contacts the Hubs in the list, noting which have been searched, until the list is exhausted or a predefined search limit has been reached. This allows a user to find a popular file easily without loading the network, while, theoretically, maintaining the ability for a user to find a single file located anywhere on the network.

Hubs index what files a Leaf has by means of a Query Routing Table, which is filled with single bit entries of hashes of keywords, which the Leaf uploads to the Hub, and which the Hub then combines with all the hash tables its Leaves have sent, it in order to create a version to send to their neighboring Hubs. This allows for Hubs to reduce bandwidth greatly by simply not forwarding queries to Leaves and neighboring Hubs, if the entries which match the search are not found in the routing tables.

Gnutella2 relies extensively on UDP, rather than TCP, for searches. The overhead of setting up a TCP connection would make a random walk search system, requiring the contacting of large numbers of nodes with small volumes of data, unworkable. However, UDP is not without its own drawbacks. Because UDP is connectionless, there is no standard method to inform the sending client that a message was received, and so if the packet is lost, there is no way to know. Because of this, UDP packets in Gnutella2 have a flag to enable a reliability setting. When an UDP packet with enabled reliability flag is received, the client will respond with an acknowledge packet to inform the sending client that their packet arrived at its destination. If the acknowledge packet is not sent, the reliable packet will be retransmitted in an attempt to ensure delivery. Low importance packets, which do not have the flag enabled, do not require an acknowledge packet, reducing reliability, but also reducing overhead, as no acknowledge packet needs to be sent and waited upon.

Gnutella2 has an extensible binary packet format, comparable to an XML document tree, which was conceived as an answer for some of gnutella's less elegant parts. The packet format was designed so that future network improvements and individual vendor features could be added without worry of causing bugs in other clients on the network.[9]

For the purpose file identification and secure integrity check of files it employs SHA-1 hashes. To allow for a file to be reliably downloaded in parallel from multiple sources, as well as to allow for the reliable uploading of parts while the file is still being downloaded (swarming), Tiger tree hashes are used.[10]

To create a more robust and complete system for searching, Gnutella2 also has a metadata system for more complete labeling, rating, and quality information to be given in the search results than would simply be gathered by the file names.[11] Nodes can even share this information after they have deleted the file, allowing users to mark viruses and worms on the network, without requiring them to keep a copy.

Gnutella2 also utilizes compression in its network connections to reduce the bandwidth used by the network.[10]

Shareaza has the additional feature to request previews of images and videos though only FilesScope takes limited advantage of this.

gtk-gnutella extended the protocol to further reduce the gap between Gnutella and G2. In particular, the semi-reliable UDP layer was enhanced to add cumulative and extended acknowledgments in a way that is backward compatible with legacy G2 clients.[12] Further extensions include the "A" string in /Q2/I [13] and the introduction of /QH2/H/ALT, /QH2/H/PART/MT, /QH2/HN, /QH2/BH and /QH2/G1 in the query hits.[14]

Overall, the two networks are fairly similar, with the primary differences being in the packet format and the search methodology.

• GnucDNA
• Shareaza

• Gnutella2 wiki
• g2paranha - G2 Network Crawler