IRTF PeertoPeer Research Group

1
IRTF Peer-to-Peer Research Group

• Paris 2005 Meeting
• Co-Chairs Bill Yeager and Bobby Bhattacharhee

2
Agenda

• A short history
• Overview of research groups charter
• What weve done so far
• Mailing list organization
• New proposals
• What research might be standardized in the IETF?

3
A Short History

• Spring of 2002
• JXTA Protocols were submitted as a draft-rfc to
  bootstrap the process leading to an IETF P2P
  Working Group
• Summer of 2002 at Yokohama IETF 54
• JXTA IPR (Some 40 Sun patents) were disclosed to
  the IETF as per rfc2026,2028 (updated by
  rfc3979).
• Bill Yeager and Jeff Altman hosted a BAR BOF of
  the IETF area directors and other interested
  parties to discuss a P2P WG with the JXTA
  protocols as a starting point
• This was summarily rejected
• It was felt that more study was needed before
  starting a WG to understand what is Peer-to-Peer,
  and given a reasonable definition, then just what
  are the interfaces with the current IETF
  standards, etc.

4
A Short History

• Fall of 2002 at Atlanta IETF 55
• After a months worth of email exchanges with
  area directors and IAB members, Jeff Altman and
  Bill Yeager were required to create a less than
  ten line description of P2P and its IETF goals.
  We managed to our own surprise and delight of the
  area directors and IAB. The 10 lines
• The proposed IETF P2P working group will
  define the architecture and suite of protocols
  necessary to implement an Internet, P2P overlay
  network (IP2PONET) to re-enable application
  layer, global end-to-end communication. Such an
  IP2PONET is a collection of Internet nodes that
  can communicate with one another and can run
  app-services wherein they play client/server
  roles. Nodes may be ad-hoc and autonomous, or
  use traditional, centralized,client/server
  technologies. The IP2PONET may or may not allow
  the definition of arbitrary namespaces. The WG is
  expected to use pre-existing RFCs relating to
  IPv4/v6, URIs, http, and current work on TLS,
  S/MIME, PKI, SACRED, XMLDSIG, DNSEXT, and MSEC.
  Depending on the WG directions, DHCP, BEEP and
  XMPP might be utilized.

5
A Short History Continued

• In Atlanta it was then decided that an IRTF p2p
  RG is appropriate for this area and to this end
  the ball was passed to Vern Paxson who was the
  IRTF chair at that time.
• Early winter 2002
• As a first step, Vern, Bill and Jeff wrote the
  research groups charter. This took about 3
  months to finish.
• Neutrality is important and is difficult to
  achieve.
• Late winter 2003
• Finally, the IRTF P2P RG was launched in early
  2003 with Bill Yeager and Bobby Battacharjee as
  the co-chairs.

6
P2P research groups charter

• A charter is a starting point, a way to set the
  direction and focus of our research it is not a
  mandate, and is mutable.
• 1. Begin with a simple definition of P2P
• Peer-to-Peer (P2P) is a way of structuring
  distributed applications such that the individual
  nodes have symmetric roles. Rather than being
  divided into clients and servers each with quite
  distinct roles (such as Web clients vs. Web
  servers), in P2P applications a node may act as
  both a client and a server. P2P systems are in
  general deployable in an ad-hoc fashion, without
  requiring centralized management or control. They
  can be highly autonomous, and can lend themselves
  to anonymity.
• 2. Give IETF relevant historical examples (P2P is
  an old concept)
• Some historical examples of P2P systems are
  USENET servers, built on top of NNTP, and
  inter-domain routing, built on top of BGP.
• Note Beginning in the late 70s with the ARPANET
  and continuing in the 80s with the Internet all
  connectivity was end-to-end, and almost all
  hosted systems were peers of one another. The
  local- and wide-area-networks as well as the
  Internet were used for massive exchange of data.
  Mounting one anothers file systems with either
  NFS or AFS to copy data was generic P2P MTAs
  are also peers, etc.
• 3. Point out the return to end-to-end
  connectivity
• A key concept for P2P systems is to permit any
  two peers to communicate with one another in such
  a way that either ought to be able to initiate
  the contact.

7
P2P research groups charter
4. P2P can enable social interaction on the
Internet As such, P2P is a powerful tool for
organizing cooperative communities - both in the
research and commercial domains - with common
goals. 5. Issues of research versus commercial
interests However, in practice, we find that the
research and commercial worlds are driven by
different needs. The former often focus on
developing generalized building blocks that can
then be composed to realize P2P systems with
quantifiable properties. These building blocks
sometimes arise out of analysis of the
deficiencies of existing P2P systems, attempting
to overcome discovered shortcomings in areas such
as peer-node organization, content caching and
distribution, lookup, search, discovery, routing,
security and trust. The commercial P2P world, on
the other hand, is driven by the concerns of
time-to-market and viable business models. Many
commercial systems have little concern for the
research issues mentioned above, while the
short-term concerns of commercial entities are
often not within the purview of academic
research. Such discontinuities in perspective
have led to a rift between the two communities,
bridging which will be of significant short- and
long-term benefit. 6. Try to motivate these
communities towards alignment, agreement The P2P
Research Group attempts to serve as such a
bridge. First, the group offers a forum for
researchers to explore a broad range of
fundamental P2P issues such as peer-node
identity, naming, configuration and capabilities
P2P network organization and scope resource
discovery, content lookup, search and
distribution request routing and operation in
the presence of mobility adaptation to expected
peer-node instability monitoring of P2P
operations security of P2P systems involving
reputation-based trust for ad-hoc systems or more
centralized, CA-like approaches etc.
8
P2P research groups charter
4. Historically commercial deployment has always
raced ahead of research, thus pointing out areas
where research and standards are required In
addition, as commercial P2P deployment on the
Internet has raced ahead of research and
standards, issues as basic as interoperable,
scalable, P2P communication protocols have been
set aside. There is no foundation upon which one
can build a unified, P2P network on the Internet
today's P2P protocols create disjoint islands of
isolated Internet nodes. 5. Our near-term
goals To this end, the research group also
emphasizes the following near-term goals
classifying the P2P problem space (both
currently, and as it evolves) into those problems
for which there are existing solutions and those
for which solutions require longer-term
development developing descriptive model(s) of
peer-node organization whose interpretation can
be applied to these solutions articulating the
scope as to what sort of P2P applications the
models encompass and what sort they do not
understanding the unique security-related
problems and opportunities P2P systems pose
exploring interfaces to IETF protocols to realize
the models and offering input to the IETF as a
starting place for possible groups standardizing
new protocols that are useful in building P2P
applications.
9
What weve done so far

• A growing bibliography of P2P references by Bobby
  with RG input
• http//www.cs.umd.edu/projects/p2prg/bib/
• A problem statement draft from the P2P RG
  mobility subgroup by Frank-Uwe Andersen, Luca
  Caviglione and Oliver Waldhorst
• (Currently email archive) It will be online in
  the next few weeks.
• A summary is found at
• http//www-info3.informatik.uni-wuerzburg.de/st
  aff/mopi/p2prgIRTFwashington2004.pdf
• A Survey of Research Towards Robust P2P
  Networks by John Risson and Tim Moors of the
  University of South Wales
• http//www.cs.umd.edu/projects/p2prg/bib/
• Mailing list activities
• Lists maintained at both the IRTF and University
  of Maryland. More on this later.

10
Mailing List Organization

• The primary list has a link at
• http//www.irtf.org/charters/p2prg.html
• Weve six research subgroups each with a mailing
  list at
• http//mailman.cs.umd.edu/mailman/listinfo/p2prg-
  
• Applications apps
• Discovery, and Resource Location Protocols
  drlp
• Metadata metadata
• Mobility mobility
• Overlay overlay
• Security security

11
Mailing List Reorganization

• Our consensus is that the mailing list
  organization is too complicated
• The proposed solution is to have a single p2prg
  mailing list at the University of Maryland and
  use
• Subject subgroup text
• to track research threads.
• Unless there are overwhelming objections, this
  will be adopted after this conference.

12
Accepted New Research Proposals

• Peer-to-Peer over IPV6 (Overlay subgroup)
• Proposed by Luca Caviglione
• luca.caviglione_at_cnit.it
• Peer-to-Peer Content, Resource and Service
  Discovery (Discovery and resource location
  protocol subgroup)
• Proposed by John Buford
• buford_at_research.panasonic.com
• Both proposals will have a link on the IRTF P2P
  RG web-site in the next week or two.
• All interested parties are both encouraged and
  welcome to participate in the research.

13
What Might Be Standardized?

• It is important to read RFC2014 on this matter. A
  relevant excerpt
• The IRTF does not set standards, and thus has
  somewhat different and
• complementary philosophy and procedures. In
  particular, an IRTF
• Research Group is expected to be long-lived,
  producing a sequence of
• "products" over time. The products of a
  Research Group are research
• results that may be disseminated by
  publication in scholarly journals
• and conferences, as white papers for the
  community, as Informational
• RFCs, and so on.
• In addition, it is expected that technologies
  developed in a Research Group will be brought to
  the IETF as input to IETF Working Group(s) for
  possible standardization. However Research
  Group input carries no more weight than other
  community input, and goes through the same
  standards setting process as any other proposal.

14
Questions? Comments, Discussion