Local Production, Local Consumption Peer-to-Peer Architecture for a
Dependable and Sustainable Social Infrastructure
Kenji Saito
∗
Keio University
Eiichi Morino
Gesell Research Society Japan
Yoshihiko Suko
Keio University
Takaaki Suzuki
Articulate, Inc.
Jun Murai
Keio University
Abstract
Peer-to-peer (P2P) is a system of overlay networks such
that participants can potentially take symmetrical roles.
This translates itself into a design based on the philoso-
phy of Local Production, Local Consumption (LPLC), orig-
inally an agricultural concept to promote sustainable local
economy. This philosophy helps enhancing survivability of
a society by providing a dependable economic infrastruc-
ture and promoting the power of individuals.
This paper attempts to put existing works of P2P designs
into the perspective of the five-layer architecture model to
realize LPLC, and proposes future research directions to-
ward integration of P2P studies for actualization of a de-
pendable and sustainable social infrastructure.
1 Introduction
Peer-to-peer (P2P) has potential to be a basis of a de-
pendable and sustainable social infrastructure for enhanced
survivability, because of its robustness and plasticity.
Designs of P2P systems are characterized by their usage
of overlay networks such that participants can potentially
take symmetrical roles. This implies distribution of author-
ities, not only p reventing introduction of single points of
failure, but also possibly assuring the level of autonomy for
self-organization, where any subsystem can spontaneously
start, maintain itself, or recover from its failures.
This paper illustrates this potential of P2P becoming a
basis of a dependable and sustainable social infrastructure,
and sets forth research agendas for its actualization.
1.1 Example: p ower outage in Tokyo
In August 14, 2006, there was a huge power outage in
Tokyo area, affecting about 1.39 million households and
∗
E-mail: ks91@sfc.wide.ad.jp
major public transportation systems in City of Tokyo and
Kanagawa prefecture (west of Tokyo) for about three hours.
What caused this trouble was a crane ship, which acciden-
tally damaged three high voltage power lines running 16
meters above water, across the river that separates Tokyo
and Chiba prefecture (east of Tokyo).
The power lines convey 275kV of electricity generated
by large power plants in Chiba to Tokyo and Kanagawa.
There are two sets of such lines which can work without the
other for a short period of time, but both the primary and
secondary sets are supported by the same pylons. Unfortu-
nately, the crane ship damaged one or two lines from each
set, disabling them to convey enough voltage.
This incident has shown that City of Tokyo and the
surrounding area had a single point of failure. Although
Tokyo Electric Power Company could successfully switch
the power supply to alternative routes after three hours, as-
suming that those alternative r outes are all configured in a
similar manner, it should be possible to cut off the power
supply to the entire Tokyo area by just blowing up several
pylons, affecting lives of more than 10 million people.
This is a kind of problem where the concept of P2P be-
comes particularly useful.
1.2 Philosophy: local production, local
consumption
Let us apply the concept of symmetrical roles among
participants to the situation of the power supply. Then ev-
ery household should be generating power, or at least every
small region should be equipped with some minor power
plants (perhaps using natural energy sources such as wind).
In order to stabilize power supply to each household or
small region, those power generators should form some net-
work (a power grid) so that excessive power can be shared,
but primarily among neighbors to prevent loss of power
from long-distance conveyance and to localize the effects
of accidents or sabotages.
Proceedings of the 2007 International Symposium
on Applications and the Internet Workshops (SAINTW'07)
0-7695-2757-4/07 $20.00 © 2007
- Lower transmission cost
Better
to purchase
from B
- Lower latency
- Higher fault-resistance
A
B
Client
Atoms (Orange)
Bits (Music File)
Time Slots (Storage)
1
2
2
3
4
Figure 1. Purchasing goods or services on a
network
This solution is generalized as follows: what consumed
locally should be produced locally, and when such resources
are unavailable, they should be conveyed from the nearest
producers. This is the philosophy of Local Production, Lo-
cal Consumption (LPLC).
LPLC is originally an agricultural term to promote sus-
tainable local economy, which can enhance mutual under-
standing between producers and consumers (because they
live near to each other), and has positive effects on local in-
dustry because the money spent by the local consumers do
not exit from the region. But it has a wide range of applica-
tions in the designs of economic systems.
In a technological context of P2P, LPLC takes the form
of emphasizin g locality of communication as some systems
such as Plaxton Mesh[10] or Tapestry[15] address.
As a total architecture not only for efficiency of com-
munication, LPLC necessarily shows characteristics of P2P,
because it promotes autonomy of subsystems in every scale.
On the other hand, it is possible to design a P2P system not
to show characteristics of LPLC, as it can be designed with-
out co nsidering the lo cality of comm unication. Therefore,
LPLC is a strengthened concept of P2P.
2 Rationales for LPLC P2P
The concept of LPLC has quite an engineering value, and
it makes more sense to design LPLC systems than just P2P.
The principle is to use the nearest resource among multiple
candidates if they are considered the same. Resources in-
clude atoms (physical goods), bits (information) and time
slots (labors or CPU time) as illustrated in Figure 1.
2.1 Purchasing atoms
Let u s consider the case of purchasing oranges online.
We cannot evaluate the values of the oranges easily be-
cause we cannot take them on our hands or taste them. Trust
with the venders or producers (for the sake of arguments,
they ar e called producers altogether henceforth) is also a
problem; it is possible that the producers d o not send the
goods to us after receiving money for them.
Perhaps if there are predecessors who have made trans-
actions with the producers before, they could provide useful
information. A reputation system will be useful.
Suppose we could successfully evaluate the qualities of
the oranges. If there are several producers with the same
grade of oranges, it makes more sense if we purchase them
from the one nearest to us in the logistic network. Then
the oranges will be fresher, and possibly cheaper because
of smaller transportation cost. Moreover, if a road is closed
somewhere because of some disaster, it is more probable
that we can still purchase oranges from the nearest producer.
LPLC is a rational strategy in an economic sense.
Note that this is always a question of a distance within
a (logistic) netwo rk, which does not necessarily equal the
physical distance.
2.2 Purchasing time slots
Let us then consider purchasing some storage service
from another computer on a P2P overlay network. As it is
the case for purchasing oranges, the quality of the services
cannot be predetermined (availability of peers, performance
of disks, etc.). Trust is also a problem because the peers
might not give access to the storage after receiving the fee.
Perhaps if there are predecessors who have made trans-
actions with the peers before, they could provide useful in-
formation. Again, a reputation system will be useful.
Suppose we could determine the quality of the storage
services. Then it makes more sense if we purchase the ser-
vice from the computer nearest in the communication path
among the ones who provide the same level of services, as it
will have less delay, less affected by faults because it passes
through smaller number of failure points, and if the network
is divided, more certain to be able to use the service.
Again, LPLC is more economical, and provides more
survivability.
2.3 Purchasing bits
Purchasing bits is another story because bits can be
copied exactly at low cost, making the concept of producers
ambiguous.
However, if there are the same music files on different
paths from a consumer, for example, it will make more
Proceedings of the 2007 International Symposium
on Applications and the Internet Workshops (SAINTW'07)
0-7695-2757-4/07 $20.00 © 2007
Identities of
Entities
Location &
Routing
Reputation
Rendezvous
Large-Scale
LPLC Network
Dependable P2P
Dependable
Social Infrastructure
Exchange
Media
- Sustaining entities
- Identifying by loose IDs
or names
- Guaranteeing
its correctness
- Grading entities
- Drafts in real terms
as media of
exchange
- LPLC network to
facilitate bartering
- Searching near-by
entities with
required specs
- Browsing near-by
entities
- Acquisition of
locations of
(copies of)
entities
- Tracking entities
- Accesses and
messages to
entities
Figure 2. Five-layer architecture model
Table 1. Five-layers and existing works
Layers Existing research works
Large-scale LPLC network BitTorrent[3], Samsara[4], MojoNation[1],
Karma[13], Ripple[7], PPay[14], i-WAT[11]
Rendezvous Peer Group Rendezvous[5]
Reputation EigenTrust[8], Stamp trading[9]
Location and routing Chord[12], Tapestry[15], etc.
Identities of entities Freenet[2], E cosystem of Naming Systems[6]
sense to access the nearest copy for the reasons we have
just investigated. Some lookup systems, such as Tapestry,
are designed to exploit this type of economy.
2.4 LPLC as a fundamental
LPLC is also a fundamental concept underlying many
of the hot topics of networking: Ubiquitous computing is
about amenities being always provided by near-by entities.
Location-based information system is about selecting en-
tities by vicinity to provide information. Sensor network
makes sensor information particularly useful for near-by en-
tities. Mobility means that a sustained image of services is
provided by a group of near-by entities even when the con-
sumers move. Ad-hoc networking creates a field of infor-
mation with near-by entities.
3 Five-layer architecture model for LPLC
The question now is how we design LPLC P2P systems.
Another question is how close the P2P research community
is to actualization of LPLC.
The authors have made the five-layer architecture model
(Figure 2), into which we have attempted to put existing
works of P2P designs as shown in Table 1.
3.1 The perspective
The model is designed in a top-down way from the up-
permost layer of large-scale LPLC network.
Some economic system is necessary to make a system
usable in reality if we share time slots such as storage
or bandwidth over a P2P overlay network. For example,
BitTorrent[3] has a mechanism of choking the download
bandwidth to peers who do not provide upload bandwidth.
In order for those facilities for purchasing or exchang ing
goods or services to work, one needs to determine what to
purchase or exchange, b y searching near-by entities satisfy-
ing required specifications (rendezvous). For such searches
to work, grading entities in a distributed manner and guar-
anteeing its correctness are r equired (reputation).
For the reputation systems to work, as well as other basic
communicatio ns and conveyance, we need to locate entities
and forward messages to them (location and routing). To do
so, we need the most fundamental mechanisms of identify-
ing entities, assuring anonymity of them if necessary (iden-
tities of en tities).
3.2 Future research directions
Currently, the focuses of the P2P research community
are mostly on location and routing and identities of entities,
into which many existing works fall.
The authors believe that we need to shift our focuses to
the upper layers, and at the same time, redesign the lower
layers based on the requirements from the upper layers.
To proceed, we need to determine h ow the uppermost
layer is designed.
4 Designs of the exchange media
An LPLC network requires exchange media, otherwise it
always has to r equire an unrealistic condition of double co-
incidence of wants
1
. The question we need to ask ourselves
is how we design digital exchange media while bits can be
easily copied (even if they are protected, people will find a
way). An answer may be that some bits will not be copied
easily (people will not want to), such as bits as a proof of
debt of the holder.
4.1 Commo dity money in LPLC network
Autonomous economy, where participants can generate
exchange media themselves, is not a new concept. In old
Japan, rice was a common value being used as a medium
1
A situation in which A wants what B can provide, and B wants what
A can provide, as it is the case for the mechanism of BitTorrent. This is
still useful in solving specific economic problems in P2P.
Proceedings of the 2007 International Symposium
on Applications and the Internet Workshops (SAINTW'07)
0-7695-2757-4/07 $20.00 © 2007
of exchange (commodity money). In P2P overlay networks,
such common values are rather found easily.
Samsara[4] is a fair P2P storage infrastructure in which
each peer that requests storage of another must agree to
hold a claim, or incompressible space, in proportion to their
consumption. Claims can be forwarded along the chain of
nodes (that requests storage of another), eliminating them-
selves when cycles are found.
The authors would argue that claims can be forwarded
in exchange with services other than storage, making them
commodity money in the context of P2P (services and
claims go the same direction). This is a possibility we
would like to pursue, although some notable problems, par-
ticularly the efficiency of storage and bandwidth, may arise.
4.2 Drafts in real terms
If rice is a common value, promise of a specific amount
of rice of a specific grade can become a guarantee believable
enough to work as money. This is the concept of drafts in
real terms, which was a exchange medium used in Japan in
around 12th-13th century. I f we are to receive rice of the
same grade in exchange of such a medium, the nearer the
producer is, the cheaper the transactional cost is. Therefore
this medium connotes the concept of LPLC.
i-WAT[11] is a protocol which can implement such
drafts in real terms in an electronic way. It is a digital ver-
sion of the WAT System, a barter currency, whose medium
is a ticket resembling a bill of exchang e.
By issuing a digital ticket promising a certain grade of
storage service, a peer in a P2P overlay network can pur-
chase a service from another. Those tickets can be for-
warded along the chain of nodes in exchange with services
(services and tickets go opposite directions), eliminating
themselves when they return to their issuers. This is an-
other possibility we would like to pursue; actually, we have
been experimenting with usages of i-WAT for several years.
5 Conclusions
LPLC, a strengthened concept of P2P, is a good philos-
ophy for designing sustainable, cooperative (with nature),
efficient, dependable social infrastructure that can support
our lives.
We have made the five-layer architecture m odel to real-
ize LPLC, and attempted to put existing research works in
the areas of P2P so that we can set forth research agendas
towards actualization of LPLC.
We introduced two candidates for the designs of ex-
change media in LPLC networks that can coexist: claims
for storage, originally an idea from Samsara, and i-WAT
tickets as drafts in real terms. We will investigate those pos-
sibilities further in the future.
References
[1] AZI. Mojo Nation technology overview.
Online archive. Available electronically at
http://web.archive.org/web/20020127125928/
www.mojonation.net/docs/technical
overview.shtml.
[2] I. Clarke, O. Sandberg, B. Wiley, and T. W. Hong. Freenet:
A distributed anonymous information storage and retrieval
system. In Proceedings of the ICSI Workshop on Design
Issues in Anonymity and Unobservability, 2000.
[3] B. Cohen. Incentives build robustness in bittorrent. In Pro-
ceedings of the First Workshop on Economics of Peer-to-
Peer Systems, May 2003.
[4] L. Cox and B. Noble. Samsara: Honor among thieves in
peer-to-peer storage. In Proceedings of the ACM Symposium
on Operating Systems Principles, October 2003.
[5] Y. Doi. Peer group rendezvous using intersection among
peer groups on DHT. In Proceedings of 2004 Symposium
on Applications and the Internet (SAINT 2004 Workshops).
IEEE Computer Society Press, January 2004.
[6] Y. Doi, S. Wakayama, M. Ishiyama, S. Ozaki, T. Ishihara,
and Y. Uo. Ecosystem of naming systems: Discussions on
a framewo rk to induce smart space naming systems devel-
opment. In Proceedings of the First International Workshop
on Dependable and Sustainable Peer-to-Peer Systems (DAS-
P2P 2006), pages 772–778, April 2006.
[7] R. Fugger. Money as IOUs in social trust networks
& a proposal for a decentralized currency network pro-
tocol. Hypertext document. Available electronically at
http://ripple.sourceforge.net/.
[8] S. D. Kamvar, M. T. Schlosser, and H. Garcia-Molina. The
EigenTrust algorithm for reputation management in P2P net-
works. In Proceedings of the Twelfth International World
Wide Web Conference, 2003.
[9] T. Moreton and A. Twigg. Trading in trust, tokens, and
stamps. In Proceedings of the Workshop on the Economics
of Peer-to-Peer Systems, June 2003.
[10] C. G. Plaxton, R. Rajaraman, and A. W. Richa. Accessing
nearby copies of replicated objects in a distributed environ-
ment. In Proceedings of ACM SPAA, June 1997.
[11] K. Saito. i-WAT: The Internet W AT System – An Architecture
for Maintaining Trust and Facilitating Peer-to-Peer Barter
Relationships –. PhD thesis, Graduate School of Media and
Governance, Keio University, February 2006.
[12] I. Stoica, R. Morris, M. F. K. David Karger, and H. Balakr-
ishnan. Chord: A scalable peer-to-peer lookup service for
internet applications. In Proceedings of ACM SIGCOMM,
August 2001.
[13] V. Vishnumurthy, S. Chandrakumar, and E. G. Sirer.
KARMA: A secure economic framework for p2p resource
sharing. In Proceedings of the Workshop on the Economics
of Peer-to-Peer Systems, June 2003.
[14] B. Yang and H. Garcia-Molina. PPay: micropayments for
peer-to-peer systems. In Proceedings of the 10th A CM con-
fer ence on Computer and communications security (CCS
’03), October 2003.
[15] B. Y. Zhao, J. D. Kubiatowicz, and A. D. Joseph.
Tapestry: An infrastructure for fault-tolerant wide-area lo-
cation and routing. Technical Report UCB//CSD-01-1141,
U.C.Berkeley, April 2001.
Proceedings of the 2007 International Symposium
on Applications and the Internet Workshops (SAINTW'07)
0-7695-2757-4/07 $20.00 © 2007
