Ity pejy ity dia navaozina tamin'ny 2020-07 ary mifanaraka amin'ny router modely 0.9.46.

Overview

I2P ships with a generic naming library and a base implementation designed to work off a local name to destination mapping, as well as an add-on application called the address book. I2P also supports Base32 hostnames similar to Tor's .onion addresses.

The address book is a web-of-trust driven secure, distributed, and human readable naming system, sacrificing only the call for all human readable names to be globally unique by mandating only local uniqueness. While all messages in I2P are cryptographically addressed by their destination, different people can have local address book entries for "Alice" which refer to different destinations. People can still discover new names by importing published address books of peers specified in their web of trust, by adding in the entries provided through a third party, or (if some people organize a series of published address books using a first come first serve registration system) people can choose to treat these address books as name servers, emulating traditional DNS.

NOTE: For the reasoning behind the I2P naming system, common arguments against it and possible alternatives see the naming discussion page.

Naming System Components

There is no central naming authority in I2P. All hostnames are local.

The naming system is quite simple and most of it is implemented in applications external to the router, but bundled with the I2P distribution. The components are:

  1. The local naming service which does lookups and also handles Base32 hostnames.
  2. The HTTP proxy which asks the router for lookups and points the user to remote jump services to assist with failed lookups.
  3. HTTP host-add forms which allow users to add hosts to their local hosts.txt
  4. HTTP jump services which provide their own lookups and redirection.
  5. The address book application which merges external host lists, retrieved via HTTP, with the local list.
  6. The SusiDNS application which is a simple web front-end for address book configuration and viewing of the local host lists.

Naming Services

All destinations in I2P are 516-byte (or longer) keys. (To be more precise, it is a 256-byte public key plus a 128-byte signing key plus a 3-or-more byte certificate, which in Base64 representation is 516 or more bytes. Non-null Certificates are in use now for signature type indication. Therefore, certificates in recently-generated destinations are more than 3 bytes.

If an application (i2ptunnel or the HTTP proxy) wishes to access a destination by name, the router does a very simple local lookup to resolve that name.

Hosts.txt Naming Service

The hosts.txt Naming Service does a simple linear search through text files. This naming service was the default until release 0.8.8 when it was replaced by the Blockfile Naming Service. The hosts.txt format had become too slow after the file grew to thousands of entries.

It does a linear search through three local files, in order, to look up host names and convert them to a 516-byte destination key. Each file is in a simple configuration file format, with hostname=base64, one per line. The files are:

  1. privatehosts.txt
  2. userhosts.txt
  3. hosts.txt

Blockfile Naming Service

The Blockfile Naming Service stores multiple "address books" in a single database file named hostsdb.blockfile. This Naming Service is the default since release 0.8.8.

A blockfile is simply on-disk storage of multiple sorted maps (key-value pairs), implemented as skiplists. The blockfile format is specified on the Blockfile page. It provides fast Destination lookup in a compact format. While the blockfile overhead is substantial, the destinations are stored in binary rather than in Base 64 as in the hosts.txt format. In addition, the blockfile provides the capability of arbitrary metadata storage (such as added date, source, and comments) for each entry to implement advanced address book features. The blockfile storage requirement is a modest increase over the hosts.txt format, and the blockfile provides approximately 10x reduction in lookup times.

On creation, the naming service imports entries from the three files used by the hosts.txt Naming Service. The blockfile mimics the previous implementation by maintaining three maps that are searched in-order, named privatehosts.txt, userhosts.txt, and hosts.txt. It also maintains a reverse-lookup map to implement rapid reverse lookups.

Other Naming Service Facilities

The lookup is case-insensitive. The first match is used, and conflicts are not detected. There is no enforcement of naming rules in lookups. Lookups are cached for a few minutes. Base 32 resolution is described below. For a full description of the Naming Service API see the Naming Service Javadocs. This API was significantly expanded in release 0.8.7 to provide adds and removes, storage of arbitrary properties with the hostname, and other features.

Alternatives and Experimental Naming Services

The naming service is specified with the configuration property i2p.naming.impl=class. Other implementations are possible. For example, there is an experimental facility for real-time lookups (a la DNS) over the network within the router. For more information see the alternatives on the discussion page.

The HTTP proxy does a lookup via the router for all hostnames ending in '.i2p'. Otherwise, it forwards the request to a configured HTTP outproxy. Thus, in practice, all HTTP (I2P Site) hostnames must end in the pseudo-Top Level Domain '.i2p'.

We have applied to reserve the .i2p TLD following the procedures specified in RFC 6761.

If the router fails to resolve the hostname, the HTTP proxy returns an error page to the user with links to several "jump" services. See below for details.

Address Book

Incoming Subscriptions and Merging

The address book application periodically retrieves other users' hosts.txt files and merges them with the local hosts.txt, after several checks. Naming conflicts are resolved on a first-come first-served basis.

Subscribing to another user's hosts.txt file involves giving them a certain amount of trust. You do not want them, for example, 'hijacking' a new site by quickly entering in their own key for a new site before passing the new host/key entry to you.

For this reason, the only subscription configured by default is http://i2p-projekt.i2p/hosts.txt (http://udhdrtrcetjm5sxzskjyr5ztpeszydbh4dpl3pl4utgqqw2v4jna.b32.i2p/hosts.txt), which contains a copy of the hosts.txt included in the I2P release. Users must configure additional subscriptions in their local address book application (via subscriptions.txt or SusiDNS).

Some other public address book subscription links:

The operators of these services may have various policies for listing hosts. Presence on this list does not imply endorsement.

Naming Rules

While there are hopefully not any technical limitations within I2P on host names, the address book enforces several restrictions on host names imported from subscriptions. It does this for basic typographical sanity and compatibility with browsers, and for security. The rules are essentially the same as those in RFC2396 Section 3.2.2. Any hostnames violating these rules may not be propagated to other routers.

Naming Rules:

  • Names are converted to lower case on import.
  • Names are checked for conflict with existing names in the existing userhosts.txt and hosts.txt (but not privatehosts.txt) after conversion to lower case.
  • Must contain only [a-z] [0-9] '.' and '-' after conversion to lower case.
  • Must not start with '.' or '-'.
  • Must end with '.i2p'.
  • 67 characters maximum, including the '.i2p'.
  • Must not contain '..'.
  • Must not contain '.-' or '-.' (as of 0.6.1.33).
  • Must not contain '--' except in 'xn--' for IDN.
  • Base32 hostnames (*.b32.i2p) are reserved for base 32 use and so are not allowed to be imported.
  • Certain hostnames reserved for project use are not allowed (proxy.i2p, router.i2p, console.i2p, *.proxy.i2p, *.router.i2p, *.console.i2p, and others)
  • Keys are checked for base64 validity.
  • Keys are checked for conflict with existing keys in hosts.txt (but not privatehosts.txt).
  • Minimum key length 516 bytes.
  • Maximum key length 616 bytes (to account for certs up to 100 bytes).

Any name received via subscription that passes all the checks is added via the local naming service.

Note that the '.' symbols in a host name are of no significance, and do not denote any actual naming or trust hierarchy. If the name 'host.i2p' already exists, there is nothing to prevent anybody from adding a name 'a.host.i2p' to their hosts.txt, and this name can be imported by others' address book. Methods to deny subdomains to non-domain 'owners' (certificates?), and the desirability and feasibility of these methods, are topics for future discussion.

International Domain Names (IDN) also work in i2p (using punycode 'xn--' form). To see IDN .i2p domain names rendered correctly in Firefox's location bar, add 'network.IDN.whitelist.i2p (boolean) = true' in about:config.

As the address book application does not use privatehosts.txt at all, in practice this file is the only place where it is appropriate to place private aliases or "pet names" for sites already in hosts.txt.

Advanced Subscription Feed Format

As of release 0.9.26, subscription sites and clients may support an advanced hosts.txt feed protocol that includes metadata including signatures. This format is backwards-compatible with the standard hosts.txt hostname=base64destination format. See the specification for details.

Outgoing Subscriptions

Address Book will publish the merged hosts.txt to a location (traditionally hosts.txt in the local I2P Site's home directory) to be accessed by others for their subscriptions. This step is optional and is disabled by default.

Hosting and HTTP Transport Issues

The address book application, together with eepget, saves the Etag and/or Last-Modified information returned by the web server of the subscription. This greatly reduces the bandwidth required, as the web server will return a '304 Not Modified' on the next fetch if nothing has changed.

However the entire hosts.txt is downloaded if it has changed. See below for discussion on this issue.

Hosts serving a static hosts.txt or an equivalent CGI application are strongly encouraged to deliver a Content-Length header, and either an Etag or Last-Modified header. Also ensure that the server delivers a '304 Not Modified' when appropriate. This will dramatically reduce the network bandwidth, and reduce chances of corruption.

Host Add Services

A host add service is a simple CGI application that takes a hostname and a Base64 key as parameters and adds that to its local hosts.txt. If other routers subscribe to that hosts.txt, the new hostname/key will be propagated through the network.

It is recommended that host add services impose, at a minimum, the restrictions imposed by the address book application listed above. Host add services may impose additional restrictions on hostnames and keys, for example:

  • A limit on number of 'subdomains'.
  • Authorization for 'subdomains' through various methods.
  • Hashcash or signed certificates.
  • Editorial review of host names and/or content.
  • Categorization of hosts by content.
  • Reservation or rejection of certain host names.
  • Restrictions on the number of names registered in a given time period.
  • Delays between registration and publication.
  • Requirement that the host be up for verification.
  • Expiration and/or revocation.
  • IDN spoof rejection.

Jump Services

A jump service is a simple CGI application that takes a hostname as a parameter and returns a 301 redirect to the proper URL with a ?i2paddresshelper=key string appended. The HTTP proxy will interpret the appended string and use that key as the actual destination. In addition, the proxy will cache that key so the address helper is not necessary until restart.

Note that, like with subscriptions, using a jump service implies a certain amount of trust, as a jump service could maliciously redirect a user to an incorrect destination.

To provide the best service, a jump service should be subscribed to several hosts.txt providers so that its local host list is current.

SusiDNS

SusiDNS is simply a web interface front-end to configuring address book subscriptions and accessing the four address book files. All the real work is done by the 'address book' application.

Currently, there is little enforcement of address book naming rules within SusiDNS, so a user may enter hostnames locally that would be rejected by the address book subscription rules.

Base32 Names

I2P supports Base32 hostnames similar to Tor's .onion addresses. Base32 addresses are much shorter and easier to handle than the full 516-character Base64 Destinations or addresshelpers. Example: ukeu3k5oycgaauneqgtnvselmt4yemvoilkln7jpvamvfx7dnkdq.b32.i2p

In Tor, the address is 16 characters (80 bits), or half of the SHA-1 hash. I2P uses 52 characters (256 bits) to represent the full SHA-256 hash. The form is {52 chars}.b32.i2p. Tor has a proposal to convert to an identical format of {52 chars}.onion for their hidden services. Base32 is implemented in the naming service, which queries the router over I2CP to lookup the LeaseSet to get the full Destination. Base32 lookups will only be successful when the Destination is up and publishing a LeaseSet. Because resolution may require a network database lookup, it may take significantly longer than a local address book lookup.

Base32 addresses can be used in most places where hostnames or full destinations are used, however there are some exceptions where they may fail if the name does not immediately resolve. I2PTunnel will fail, for example, if the name does not resolve to a destination.