LEARNET Architecture (LARCH)
Pearl: Proposal for Experimental Academic Research using LEANET
Authors: Simon Crosby, Ian Leslie, Cambridge University Computer Laboratory
Jon Crowcroft, UCL Computer Science DepartmentCS
Lionel Sacks, Chris Todd, UCL EE
[Thanks to Bob Briscoe at BT for supplying references]
The realisation of the Communications Centre joining the strengths of Cambridge with those of UCL, and the ongoing London Communications Centre involving Imperial College, Kings College, QMW and UCL represents a unique concentration of research skills and professional training facilities in communications, certainly in Europe and probably World Wide. The support that BT has given these initiatives and in particular the offer of the London East Anglian Network (LEANET) as a potential active experimental base represents a vital step forward.
Given the scale of LEANET and the implicit complexity of proposed collaborative experimental work on that network, there are a number of critical factors to consider.
Although these critical factors are demanding, UCL/CAM believe their solution is eminently justified by the scale of the research opportunity that collaboration on LEANET could provide at this critical juncture in the evolution of the communications business. The programme, which forms the basis of this proposal, has been targeted at this opportunity and is far too large for the envisaged two RA’s to complete in 3 years, given their other duties outlined above. To complete this programme, and the one or two others that are ‘a gleam in the eye’ but will take shape soon, will require a critical mass of researchers and the portfolio approach also outlined above.
The provision of new services on current and future high speed networks will require network elements to be flexible and easily controlled, because service components will require specialised communications features to be implemented within the provider network. The Asynchronous Transfer Mode (ATM) has been widely hailed as a solution to the problem of supporting multiple services, because it appears to offer the best prospects for solving a particularly awkward problem in an integrated communications environment, namely the accommodation of multiple services with widely differing resource and performance demands.
Research in recent years has focused on the traffic controls required to ensure that data transmissions from multiple sources of different types each receive the appropriate level of service from the network. Typically these controls are located in the data path, and include scheduling techniques, policing and shaping functions. Until recently, however, few researchers had seriously considered the control plane aspects of the provision of multiple services in broadband networks. Key to the provision of new services and the support of existing and standards-based services is the requirement for a powerful, and flexible network control mechanism capable of meeting the needs of multiple network services, such as native ATM connections, IP, IPv6, and specialised services tailored to specific communications needs, including video distribution, remote sensing or conferencing. ATM as a multiplexing technique can provide resource management for new services, however ATM as standardised by the ITU-T and ATM Forum is inflexible and does not encourage the development of services which require modes of communication or control which cannot be expressed in the standardised UNI protocols. Unfortunately the standardised protocols are not sufficiently expressive for the needs of new services, and even support for IP, the service type from which operators reasonably expect to earn a significant proportion of their revenue in the near future, is poor and does not scale.
Fortunately, recent developments from research in the area of "open signalling" (or more aptly, open control) suggests that this problem can be overcome. The use of sophisticated distributed systems architectures in the network control and management infrastructure can drastically simplify the design and implementation of truly multi-service networks. In addition, devolving the control of switches to off-switch control processors allows the basic switching functionality of the network to be exploited in innovative ways, for example as used by Ipsilon’s IP switching, or the Columbia control architecture. The fundamental ability of ATM to perform high speed switching with fixed-sized units of bandwith can thus be exploited by new services.
The LEANET offers a unique opportunity to build a large multiservice network based around a common switched core network with fast and versatile provisioning of connection services, and to test this in a realistic context with sufficient user load processing and communications complexity. The convergence of differing uses of telecommunications networks gives rise to ‘layer’ convergence in management and service provisioning; with the consequence that the responsibility for set-up, routing, error reporting etc shift between the control plane and distributed management systems and the division between these ‘layers’ becomes ill defined. It is clear that to build a complete service, each layer of the system must work harmoniously with the others. This requires more than the traditional stack-service based approach and ‘Layer Violation’ must be managed and the QoS concepts can be seen as a way to do this. In general, the LEANET test bed, as a complete captive network with full access to all ‘layers’, should provide an ideal context to look well beyond conventional layering criteria and to allow new full service definitions to be created; thus providing a knowledge base for future application development in BT.
The proposed work will range from control to service up and down the stack, and will use as a basis for its research a technique developed at the University of Cambridge Computer Laboratory which enables multiple network control architectures (CAs), such as ITU standard Q.2931 signalling, TAG switching, or Ipsilon IP switching to be supported on top of a single network. New services may be built upon the architecture, and crucially, virtual networks of any type can be created. UCL Computer Science will bring their previous experience to bear to attack complexity and manageability issues associated with IP(v6)/ATM service provision with RSVP control in its own virtual network, and seek some understanding of the very different Virtual Circuit World View and the IP World View of the meaning of multicast; how switch/routers deal with these different viewpoints is itself a significant research issue for the programme. UCL CS are also interested in open service provision, QoS in routers, IP traffic QoS, measurement and estimation of traffic resource requirements, and end system requirements for multimedia. UCL EE will be concerned with the quality of the connectivity management with respect to eg the call / data connection requests (e.g. connecting to data), (advanced IN type) personal services and to the availability of status information and in doing so will move beyond the recent experience gained in a number of ACTS Projects to develop Quality of the Quality of Service ie (Qo)2S concepts, in the context of an ongoing risk management research project for BT; there are particular issues here arising from the Cambridge virtual networks approach. UCL EE would seek to define ‘periphery’ agent platforms to link for example advanced personal service generics with the ‘network service on demand’ concepts of Cambridge and thereby build on other recent research in EE dealing with the complexity of mixed technology environments.
As discussed in Section 1, this proposed programme of research which is detailed in Section 3, presumes a fully operational LEANET. However it is assumed the the two RA post holders will need to spend a significant portion of their time working jointly with BT staff to enable a configuration that will support the research programme, and subsequently to aid with the management of LEANET as a multipurpose experimental platform. The proposed programme builds on current research strengths in the Cambridge and UCL groups concerned, and the summary demonstrates how individual research interests in the three groups have been assembled to compliment and be mutually supportive in the programme of work. The UCL and Cambridge groups have prior experience of effective interworking and a methodolgy for technical exchange.
The original area of work (
http://www.cs.ucl.ac.uk/staff/jon/pearl/pearl.htm) is around the complexity and manageability of IP(v6)/ATM service provision - we will help evaluate the results of the network that BT build, with appropriate RSVP control in its own virtual network, and the QoS support within the virtual networks.Since the writing of the proposal, the IETF and ATMF ideas have moved along. It would appear that the evolution of the Internet had been heading towards a classical telecommunications model, with RSVP providing a form (albeit "soft-state") of connection oriented networking. Apart from some minor details of how to define service level agreements, and how to carry out efficient fast IP/ATM forwarding, the "Integrated Services Internet" seemed to offer a technical and business case solution to the problem of provisioning, accounting and billing for the Internet with the same levels of assurance that GSTN offers.
However, the influence of ideas from pricing has led to the notion of Differentiated Services. This has led to the re-assertion of the "IP Classic" open model of networks-sans-frontiers created by UNI signalling. Instead, user profiles are installed at subscription time and distributed to ingress routers at the access points.
The assumption is that measurement and policing are still in place, but that the user does not need to use per session set-up, so that RSVP is not needed, nor is there necessarily an admission test, although there may be some equivalent function carried out to calculate a tariff if for congestion based pricing for some types of profile.
However, the notion of networks on demand (VPNs and similar) has grown more relevant, whether implemented at the IP level or below.
Currently, the Internet functions since sources are adaptive. This is not policed, but ubiquitous use of TCP means that all WWW and Email currently typically finds its level by adapting to available capacity.
Multimedia applications, whether streaming, one-to-one, multicast, or interactive, one-to-one or multicast, require some particular quality to be satisfactory for some user activity. The user can adapt to some extent (and who will not, if and when higher quality is free).
There are various time-scales for adaption, depending on whether we are looking at the available capacity, the user’s reaction to perceived quality, or adaption to transient losses incurred due to noise in transmission, or adaption to the burst losses caused by other adaptive and variable sources peak or near-peak level bursts.
T1: Adapting to users quality and affordable, or available capacity - >> RTT
T2: Adapting to bursts incurred due to other users ~~ RTT
T3: Adapting to arrival of more or less other users ~~#users and O(RTT)
An RTT is a round trip time – whether we have feedback from within the network (FECN or BECN in Frame Relay terms), or we have a full round trip time only affects this by a small constant multiplier.
With RTCP feedback, RTP/UDP sources can adapt (and with representatives, can do so scalable too!). So how do we choose what the target to adapt to is?
Well, control theory says that we want a relatively stable system. T2 is the shortest of our time frames, and T1 is dependant to some extent on T3 (and on users’ preparedness to pay!).
So we must use an aggressive backoff, and tentative increase algorithm for our adaption for T2, just as TCP does. For T1, after the system has settled, we can let the user see what quality they get, and whether a new price/quality is required, and then have them request a new coding/quality/transmission level.
Of course, there are then 3 different possible mechanisms within the network for fulfilling the required QoS (a hybrid can be imagined):
Q1. Dimension the network for adequate quality – e.g. minimum of 1 packet of quality per RTT
Q2. Ask the user to pay enough so that the current dimensioned network is already fast enough – increase the price (aggressively)as load goes up close to overload, decrease it (tentatively) as load decreases.
Q3. Ask the user to specify their flow requirements (possibly via measurement, done on their behalf), and run a admission algorithm; if this fails (possibly measurement-based admission), and arbitrarily reject a flow (rejection is not arbitrary if it results in a price increase); otherwise, like call blocking in the phone net, it is "last user loses" policy. Other role-based policies could be envisaged!
In the diagram below, we have switches and routers: this is for the purposes of illustration ; in fact, in the network at the time of writing, there are also SDH multiplexors, and site switches and site routers as well as core switches and routers – currently, ASX1000s inside the net, ASX200s outside, with FreeBSD PC based routers running CIARN s/w for IP forwarding, and Cambridge software for IP to ATM and other control planes.
Figure 111. Use of Switchlets to virtualise ATM switch control.
.
The idea is to provide middleware for the operator that allows them to partition the network at whatever level – for example, the ATM or SDH levels, into a set of quarantined service offerings, each of which can offer different (evolving) control planes. The rest of the work in the programme is made on top of this architecture, and acts to inform the evolution of control planes.
Figure 222. Virtual networks, each controlled using a different control.
Most of the application level work will use an IP control plane, but there may be more than one of these – e.g .providing some particular service discipline (e.g. Best effort, or Integrated Services with RSVP<->Q2931 interworking, or Differentiated Services with RSVP RAP <-> Measurement Based Admission interworking).
Mbone GUIs
Web GUIs
Evaluation GUIs
Plugins – for Browser, Server and SDR/Mbone toolset
User Perception and Understanding of tariff structures……
Multicast, Mobile and other services stress the model of who pays (and who signals!)
The work in the IETF on IP switching has moved on a pace – there are now at least four proposals for IP on ATM (and on other switched VC nets) although none address multicast well, or the problem of short lived flows actually observed in the Internet (95% of flows are HTTP, of which 90% are 4k byte downloads – typically a total of 14 packets for a particular source-destination pair – the temporal correlation in subsequent destinations for web access is very low.
ARIS seems a good fit for this type of traffic ,whereas tag and flow labelling will thrash.
CSR (the Tiny Tera is the best example of a CSR platform) is a nice way to implement these models, although the implications of multicast on an input queued switch router are not well understood (the fan-out in a multicast tree loads a crossbar order n, for a tree of degree n).
Never destined to take off in a big way, end to end ATM lacks a flexible open architecture for higher level protocols – the DAN and other work at Cambridge suggested lots of ways forward, but despite their valiant efforts, I think we can rule this out as being of low interest.
Users don’t know how to characterise their own traffic – whether the user is an individual running an application, or the author of a new tool, or a site subscribing to a backbone, it is hard for users to come up with a good measure to ask for a particular QoS – the problem is exacerbated by multi-metric signaling protocols, where there are QoS surfaces or regions of equivalent bandwidth. The distinction between different choices can be easier made within the network, by measureing the traffic. Lots of results show this is feasible in short order (e.g. fractions of an RTT), requiring no signaling at all – what remains, is perhaps an architecure which requires signaling feedback to the user of the costs of their current offered load given the current circumstances.
An important area to understand is that of timescales of this feedback – is it per session, per variation within a session, or is it something that, at least for most typical flows, is actually just a bill (an adddition to an account) and that varies slowly based on new services and applications emerging and being measured/characterised?
A problem with Class-of-Service, as Differentiated Service defines is that it is destinationless.
This ,means that there is a problem provisioning the network:
an interesting excercise is to look at
http://www.statslab.cam.ac.uk/~richard/PRICE/pricing-internet/
meanwhile
pricing work (for the internet and telecoms) covers a number of areas
0/ economics - networks (as transmissio nresource, not for content)
are closer to utilities (gas, electricity, water) than to commoditiy
markets - pricing work in economic theory has stemmed largely from
here - often including models of social welfare as well as free market
models of selling capacity as a scarce resource - hence a lot of work
on fairness
you don't have to be a capitalist to want a pricing mechanism -
pricing can just be a way of redistributing wealth, and creating
incentives to not misuse the network...
metrics (what is accouted and billed for) are essential, as well as
accuracy and timespans e.g. do we charge per bit-per-sec per time of
day, or just by volume, or by the web page-
metrics might include
and 2nd or higher order (variance etc(=jitter)averaged over some interval
and by some time of day, for source-destiantion possibly including
distance in some way...
Of course yo uneed arationale for this - are you recovering capital
costs (over what time frame) and maintenance, or are you trtying to
invest in the future? - depends on the common goals of providers and
subscribers, so it vari
es - the internet is no longer 1 constituency
pricing (current main internet model) you need accounting - what level
of detail depends on
TBA by angela et al
Kelly work on bursty and on elastic, and on proportional fairness
An interesting excercise is to look at
http://www.statslab.cam.ac.uk/~richard/PRICE/pricing-internet/
pricing work (for the internet and telecoms) covers a number of areas
of economics - networks (as transmission resource, not for content)
are closer to utilities (gas, electricity, water) than to commoditiy
markets - pricing work in economic theory has stemmed largely from
here - often including models of social welfare as well as free market
models of selling capacity as a scarce resource - hence a lot of work
on fairness
you don't have to be a capitalist to want a pricing mechanism -
pricing can just be a way of redistributing wealth, and creating
incentives to not misuse the network...
metrics (what is accouted and billed for) are essential, as well as
accuracy and timespans e.g. do we charge per bit-per-sec per time of
day, or just by volume, or by the web page-
metrics might include
of course you need a rationale for this - are you recovering capital
costs (over what time frame) and maintenance, or are you trtying to
invest in the future? - depends on the common goals of providers and
subscribers, so it varies - the internet is no longer 1 constituency
pricing (current main internet model) you need accounting - what level
of detail depends on
i) audit rules
ii) model of traffic
a system/model that doesnt factor in the price of charging and of
supprting more than 1 service is doomed...
the cost of monitoring may be high - many routers cannot monitor
accurately the traffic flows thru them - it costs to depoy a bunch
more routees, or seperate measurement systems.....on the other hand,
measurements are useful for other things (e.g. predicting when you need to
upgrade links...)
if you offer a range of services than you need mechanisms that support
them - these have costs...
service models (e.g. best effort, the range of integrated services
and the range of differentiated services, or in BISDN terms, UBR, ABR,
VBR, CBR, or in plain terms, elastic to inelastic) as well as other
factors (e.g. tiem of day, or congestion related pricing) affect the
complexisty of the system
if we had pervasive ecash (especially millicent systems) we might
consider using per call charges...
the network isn't standing still - capacity is deployed as demand
grows - the model of a stable system is a dead model
sets of people working on perception and understanding of price v.
performance
expectations are interesting here...(viz PSTN versus mobile phone
quality and reliability:-)
mechanism that support a range of the models above, and allow free
choice of subscription, flat fee, ecash, metric, service model, sender
v. recipient billing v. share, etc etc would be nce, so long as it
doesn't cost more than the net
capacity than to deploy a billing system - right now, ALL ISPs are
working by this rule.........the first to introduce a usage bill will
have to show a marked, and perceived edge, to the users....note
metcalfe's law: the power of the network is the square of the number
of users - this applies to service offerings within the internet -
they have to capture a significant piece of the net before they show
up in any users perception (otherwise a lot of users are accessing
servers on the new net, or from the new net that are on the old, and
see no different) - exceptions then (market entry points) are
a) intranets
b) new deployments (e.g. by cable tv or direct digital broadcast+internet,
by GSM + internet, or by large telcos in new areas (e.g. replace
analoge access line with adsl...etc)
[{ACL_perf?}] {Firewall 10% performance drop per ACL, via Dave Douglas}=20
this is about typical behaviour of boxes that identify individual
flows - but it will be out of date i nthe next generation of routers
(e.g. cisco GSR, tiny tera, juniper, lucent and other routers) which
can do per flow WFQ
[ActiveX] "Microsoft Internet Explorer 4.0 White Paper; Complete
Administration", Microsoft, Jul 1997,
<URL:http://www.microsoft.com/ie/press/whitepaper/iwhite/white016.htm>=20
company name:
nuff said....:--)
[Bailey95] J. Bailey, S. Gillett, D. Gingold, B. Leida, D. Melcher, J.
Reagle, J. Roh, R. Rothstein, and G. Seale, "Internet Economics Workshop
Notes," Research Program on Communications Policy, MIT, Mar 1995.
<URL:http://www.press.umich.edu/jep/works/BailWNotes.html>=20
this workshop was a key step i nthe work on pricing - essential
reading
[BaileyMcK95] Joseph P Bailey and Les W McKnight, (MIT), "Internet
Economics: What Happens When Constituencies Collide", Apr 1995,
<URL:http://aleph.pangea.org/HMP/PAPER/123/html/paper.html>=20
handwavy telco v. internet wats
[Barns89] Barns, W, "Defense Data Network Usage Accounting Enhancement
Approaches," The MITRE Corporation, 1989.=20
historical
[BlueSquirrel] WebWhacker and other Blue Squirrel Web robot products,
<URL:http://www.bluesquirrel.com/products.html>=20
an example of turbo applications
that beat up the net with heavy load, but have actual value -
why should they pay more, but why not? and wh ycan't they?
>
[Bohn93] Roger Bohn(Uni of CA-San Diego), Hans-Werner Braun & Kimberly C
Claffy (San Diego Supercomputer Center) and Stephen Wolff
(NSF DNCRI), "Mitigating the coming Internet crunch: multiple service
levels via Precedence", Technical Report, University of California-San
Diego, San Diego Supercomputer Center and NSF, Nov 1993,
<URL:http://www.nlanr.net/Papers/mcic.html>=20
this was the earliest paper talking seriously about using existing IP
priority to use different queues to get different throughput/delay
classes.
[Bolot94] Jean-Chrysostome Bolot (INRIA), "Cost-quality tradeoffs in the
Internet", Computer Networks and ISDN Systems, 1994,
<URL:ftp://www.inria.fr/rodeo/bolot/94.Tradeoffs.ps.gz>=20
this is good background, but jean's done a lot more on admission
control since.
[borfp96] OMG request for proposals (RFP) for Common Business Object
Specifications <URL:http://www.omg.org/pr96/bomrfp.htm>=20
OMG do CORBA _ a distributed platform for management systems - the
business model is neat - tal kto dave lewis at UCL CS for more
info....
[bow97] "OOPSLA'97 Business Object Workshop" Forthcoming, Proceedings, pub.
Springer-Verlag
<URL:http://www.tiac.net/users/jsuth/oopsla97/index.html>=20
see dave again if interested - particuallry for subscription and
workflow management
[Brownlee94] Nevil J. Brownlee (Uni. of Auckland), "New Zealand experiences
with network traffic charging", usage-based access charging,
ConneXions Volume 8, No. 12, Dec 1994,
<URL:http://www.press.umich.edu/jep/works/BrownNewZe.html>=20
this is interesting, although NZ and Oz are very special cases where
its REALLY easy to indeitfiy all users and all itnernational link
users - but worth a read....its the basis for the UK charging that
just started (amybe people didnt knwo this, but each university gets a
volume
usage bill for inbound data from US link since january - see
http://www.cs.ucl.ac.uk/staff/jon/jisc.html
for a proposal on how to do this better
[Cairncross97] Frances Cairncross, "The Death of Distance : How the
Communications Revolution Will Change Our Lives", pub. Harvard
Business School Press, ISBN: 0875848060, Oct 1997=20
standard argument about the end of distance related bills (e.g .a lot
of PSTN providers esp. in US are removing more and more distance and
time related bills as the system becomes saturated with capacity so
that demand (except in flash calls e.g. television phone in vote,
world cup or rock concert ticket sales) cannot exceed supply...
[Calhoun98] Pat R. Calhoun (Sun), Allan C. Rubens (Ascend), "DIAMETER; Base
Protocol", work in progress: Internet Draft, Mar 1998,
<URL:draft-calhoun-diameter-02.txt> (also see Internet Drafts index for
many other DIAMETER drafts)=20
metric work
[Carpenter96] Brian Carpenter,"Metrics for Internet Settlements", EXPIRED
Internet Draft, May 1996, but the main information from this draft is
recorded by Claffy at <URL:http://www.nlanr.net/Papers/telegeog96.html>=20
worth a look - in a multiprovider environemnt, there may be traffic
imbalances - we looked at this along the lines of the phone company
nets.....its tricky (but internet access providers are different from
backbone providers and already have settlements!)
[Cisco] Cisco Systems, "Queuing", informal white paper,
<URL:http://mmgp.cisco.co.jp/technologydoc/queue.html>=20
what cisco can do in their routers....
[Clark95] David D Clark (MIT), A Model for Cost Allocation and Pricing in
the Internet, presented at MIT Workshop on Internet Economics,
Mar 1995, <URL:http://www.press.umich.edu/jep/works/ClarkModel.html>=20
has a very nice discussion of sender v. receiver based pricing
[Clark97] David D Clark and John Wroclawski (MIT), "An Approach to Service
Allocation in the Internet", EXPIRED Internet Draft, Jul 1997,
<URL:http://diffserv.lcs.mit.edu/Drafts/draft-clark-diff-svc-alloc-00.txt>=
discusses the diff serv model
[Cocchi93] Ron Cocchi (Hughes/USC), Scott Shenker(Xerox PARC), Deborah
Estrin (USC) and Lixia Zhang (Xerox PARC), "Pricing in Computer
Networks: Motivation, Formulation, and Example", in: IEEE/ACM Transactions
on Networking, vol. 1 (6), pp. 614-627, Dec 1993
<URL:http://www.wiwi.hu-berlin.de/~brupp/pricing/PricingInComputerNetworks.p
s.gz>=20
nash equilibrium
[{Crowcroft1n?}] {Advantage of one network to manage - via Crowcroft}=20
wafle about why we'd like the internet to take out the PSTN and TV
nets:-)
[Crowcroft96Kn] Jon Crowcroft, "QoS - Plenty of $$$ but just 1 Bit",
Keynote speech presented at Protocols for High Speed Networks '96, Oct
1996, <URL:http://www.cs.ucl.ac.uk/staff/jon/hipparch/dollarbit/> based on
HIPPARCH discussions between Christophe Diot,
Jean-Chrysostome Bolot, Jon Crowcroft & David Clark=20
idea (1st) that we could get away with a price based QoS mechanism
without any other parameters - just a delay preference for the
price...
[Crowcroft96] Crowcroft, Jon, "Pricing the Internet", in IEE Colloquium on
Charging for ATM (ref. no. 96/222) 12 Nov 1996, pp1/1-4.=20
similar to above...
[Crowcroft98] Jon Crowcroft & Philippe Occhslin (UCL), "Differentiated End
to End Internet Services using a Weighted Proportional Fair
Sharing TCP", Apr 1998, <URL:ftp://cs.ucl.ac.uk/darpa/multcp.ps> or <URL:
http://www.cs.ucl.ac.uk/staff/J.Crowcroft/hipparch/pricing.html>=20
a mechanism to get faster TCP throughput for people who pay more to
the sender system, with NO network supprt needed at all (nearly)
[Danielsen95] K. Danielsen (Molde College, No) and M. Weiss (Uni. of
Pittsburgh), "User Control Modes and IP Allocation," Journal of
Electronic Publication, University of Michigan Press,
<URL:http://www.press.umich.edu/jep/works/DanieContr.html>, 1995.=20
similar to mackie mason/varian smart market idea
[diffserv] Differentiated Services (diffserv) IETF working group charter
<URL:http://www.ietf.org/html.charters/diffserv-charter.html>=20
more about diff serve
[Edell95] Richard Edell (Uni of CA-Berkeley), Nick McKeown (Stanford), and
Pravin Varaiya (Uni of CA-Berkeley), "Billing Users and Pricing for
TCP." IEEE JSAC Special Issue on Advances in the Fundamentals of
Networking, Sep 1995,
<URL:http://tiny-tera.stanford.edu/~nickm/papers.html>=20
another approach to charging for different TCP thruput
[Faupel97] Matthew Faupel (APM), "Java Distribution and Deployment", ANSA
report APM.2038, Oct 1997
<URL:http://www.ansa.co.uk/ansasponsors/phase3-doc-root/sponsors/APM.2038.02
.ps.gz> (ANSA sponsor access only)=20
worth a look - from the ANSA people (who bought you the prototype of
CORBA) - a discussion of Java's usefulness and pervasiveness
[Finlayson98] Ross Finlayson, Discussion at Third Reliable multicast
research group meeting, Orlando, FL, and following on the mailing list,
Feb 1998 <URL:http://www.east.isi.edu/rm/>=20
RM - see meeting at UCL in 6 weeks...
[{Floyd?}] {Absorbing bursts with one network - via Crowcroft or Floyd}=20
why stat muxing and bursts are a good thing
[Floyd93] Sally Floyd and Van Jacobson, (LBL) "Random Early Detection
Gateways for Congestion Avoidance", IEEE/ACM Transactions on
Networking, Vol.1, No. 4, pp. 397-413. Aug 1993,
<URL:ftp://ftp.ee.lbl.gov/papers/early.ps.gz>=20
random early detection is a mechanims for making TCP fairer, and
making IP router qwueues shorter on average without chaning end
systems
[FollowMe] "FollowMe", ANSA/ESPRIT research project,
<URL:http://www.ansa.co.uk/Research/FollowMe.htm>=20
[Freestone97] D Freestone and M Owen, (BT), "Operational support systems
for the future", BT Technology Journal, Vol.15, no.1, p172, Jan
1997. <URL:http://www.labs.bt.com/library/bttj/v15n1/16.htm> (abstract &
order details only) or
<URL:http://sss.info.bt.co.uk/BOOKS_AND_JOURNALS/BTTJ/v15n1/16.htm> (BT
access only)=20
pass (how about BT Tech journal being online for free to the customers
eh? like ACM SIGCOMM ? :-)
[Herzog95] Shai Herzog (IBM), Scott Shenker (Xerox PARC), Deborah Estrin
(USC/ISI), "Sharing the cost of Multicast Trees: An Axiomatic
Analysis", in Proceedings of ACM/SIGCOMM '95, Cambridge, MA, Aug. 1995,
<URL:http://www.research.ibm.com/people/h/herzog/sigton.html>=20
first wortk on how to apportion a bill to a st of receivers o na
multicast tree (actually i think its broken - why not just charge tyhe
sener by the NUMBER of receivers...a la TV net or UUNETs UUCAST
service which does this * data rate)
[Huitema95] Christian Huitema, "Routing in the Internet", Prentice Hall,
ISBN: 0131321927, Apr 1995=20
very good intro to IP routing
[IEE802.1p] IEEE, "Draft Standard for Traffic Class and Dynamic Multicast
Filtering Services in Bridged Local Area Networks (Draft Supplement
to 802.1D)", IEEE802.1p, 1997=20
how bridges do ip multicast IGMP snooping etc)
[IGMPv3] Brad Cain (Bay), Steve Deering (cisco), Ajit Thyagarajan
(Torrent), "Internet Group Management Protocol, Version 3", Work in
progress: IETF Internet Draft, Nov 1997 (expires May 1998),
<URL:draft-ietf-idmr-igmp-v3-00.txt>=20
IGMP with fast leave
[intserv] Integrated Services (intserv) IETF working group charter
<URL:http://www.ietf.org/html.charters/intserv-charter.html>=20
the int serve model
[ippm] IP Performance metrics (IPPM) IETF working group charter
<URL:http://www.ietf.org/html.charters/ippm-charter.html>=20
good provider metric framework
[IPv6] S. Deering, (cisco), R. Hinden (Ipsilon), "Internet Protocol,
Version 6 (IPv6) Specification", work in progress: IETF Internet draft, Nov
1997, <URL:draft-ietf-ipngwg-ipv6-spec-v2-01.txt> or latest draft through
IPng working group charter,
<URL:http://www.ietf.org/html.charters/ipngwg-charter.html>=20
V6 (not very relevant really)
[Jacobsen88] Van Jacobsen (LBL), "Congestion avoidance and control", In
Proc. ACM SIGCOMM'88, 314-329,
<URL:ftp://ftp.ee.lbl.gov/papers/congavoid.ps.Z>=20
how van saved the internet from complete failure in 1988....TCP
control loop
[Jacobsen95] Referred to in [Shenker96] as "Van Jacobsen [LBL], private
communication, 1995."=20
refinements of above (i think...)
[{Kelly?}] {Kelly? Statistical analysis of elastic/inelastic utility curves}=
pass
[Kelly97] Frank P Kelly (Cambridge Uni.), "Charging and Accounting for
Bursty Connections", in "Internet Economics" (Editors Lee W.
McKnight and Joseph P. Bailey) MIT Press, 1997. 253-278.
<URL:http://www.statslab.cam.ac.uk/~frank/charge.html>=20
this is neat - basically, VBR (or CL or GS in internet) - an
optimisatio napproach to figuring out the bill
>
[Kelly98] Frank P Kelly, Aman K. Maulloo and David KH Tan, "Rate control
for communication networks: shadow prices, proportional fairness
and stability", Journal of the Operational Research Society, 49,1998,
<URL:http://www.statslab.cam.ac.uk/~frank/rate.html>=20
this is the theory base for our TCP work (multcp - see workshop at UCL
on june 15/16 where panos is talking about measurements of it)
[Kramer96] Douglas Kramer, "The JavaTM Platform", Sun White Paper, May 1996,
<URL:http://www.javasoft.com/docs/white/platform/CreditsPage.doc.html>=20
more background on java
[Lehr96] William H Lehr (Columbia Uni.) & Martin BH Weiss (Uni. of
Pittsburgh), "The political economy of congestion charges and settlements
in packet networks", Telecommunications Policy, 20 (1996), 219-231,=20
more smart markets
[Lin97] Dong Lin and Robert Morris (Harvard Uni) "Dynamics of Random Early
Detection", in Proc. SIGCOMM'97, Computer Communication
Review, ACM SIGCOMM, volume 27, number 4, ISSN # 0146-4833, Oct 1997
<URL:http://www.inria.fr/rodeo/sigcomm97/program.html#ab078>=20
Fair RED - a refinement of RED to treat slow connections better
[MacKieVar92] Jeffrey K MacKie-Mason and Hal Varian, (UMich), "Some
Economics of the Internet=92, Tenth Michigan Public Utility Conference
at Western Michigan University, March 25-27, 1993:
<URL:http://www.sims.berkeley.edu/~hal/people/hal/papers.html>=20
the original congestion billing paper
[MacKieVar93] Jeffrey K MacKie-Mason and Hal Varian, (UMich), "Pricing the
Internet", Public Access to the Internet, JFK School of
Government, Apr 1993
<URL:http://www.sims.berkeley.edu/~hal/people/hal/papers.html>=20
a refinement of the above
[MacKieVar95] Jeffrey K MacKie-Mason and Hal Varian, (UMich), "Pricing
Congestible Network Resources", IEEE Journal on Selected Areas in
Communications, "Advances in the Fundamentals of Networking'', 1995,
<URL:http://www.sims.berkeley.edu/~hal/Papers/pricing-congestible.pdf>=20
ditto
[Newman96] Peter Newman, Tom Lyon & Greg Minshall, (Ipsilon) "Flow Labelled
IP: A Connectionless Approach to ATM", Proc. IEEE
Infocom'96, vol. 3, pp. 1251-60, Mar 1996
<URL:http://www.ipsilon.com/~pn/papers.html>=20
interestying since it shows that you dont need connection setup to
detect a longish lived flow...
[Nichols98] Kathleen Nichols (Bay), Steven Blake (IBM), (eds),
"Differentiated Services Operational Model and Definitions", work in=
progress,
IETF Internet Draft, Feb 1998
<URL:http://www.ietf.org/internet-drafts/draft-nichols-dsopdef-00.txt> (see
[diffserv] for latest version)=20
how diff serve is really going to work in detail
[Odlyzko97] Andrew Odlyzko (AT&T Research), "A modest proposal for
preventing Internet congestion", Sep 1997,
<URL:http://www.research.att.com/~amo/doc/recent.html>=20
a possibly flawed idea based on Paris Metro Pricing (only diff between
1st and 2nd class is the price)
[Peyton85] H. Peyton-Young, editor, "Cost Allocation Methods, Principles,
Applications. pub. Elseviers, 1985=20
baseline economics
[Putland97] Paul Putland, Jake Hill, and Dimitrios Tsapakidis, "Electronic
payment systems", BT Technology Journal, Vol.15, no.2, p32, Apr
1997. <URL:http://www.labs.bt.com/library/bttj/v15n2/03.htm> (abstract &
order details only)
<URL:http://sss.info.bt.co.uk/BOOKS_AND_JOURNALS/BTTJ/v15n2/03.htm> (BT
access only)=20
pass
[radius] Remote Authentication Dial-In User Service (radius) IETF working
group charter
<URL:http://www.ietf.org/html.charters/radius-charter.html>=20
important architecturally becvause of legacy access networks such as
modem+phone
[rap] RSVP Admission Policy (RAP) IETF working group charter
<URL:http://www.ietf.org/html.charters/rap-charter.html>=20
signaling (if you want it)
[Reeder98] Tony Reeder & James Cochrane (BT), "A Review of Internet
Charging Models", Mar 1998,
<URL:http://www.jungle.bt.co.uk/projects/lsma/charging/qos-based/busmodel/re
view/review~1.html> (BT access only)=20
pass (this is gettign annoying:-)
[RFC768] Jon Postel (USC/ISI), "User Datagram Protocol", STD 6, IETF RFC
768, Aug 1980, <URL:rfc768.txt>=20
how connectionless (e.g. multicast) apps send packets
[RFC791] Jon Postel (ed), "Internet Protocol", IETF RFC 791, Sep 1981,
<URL:rfc791.txt>=20
how all packets look
[RFC793] Jon Postel (ed), "Transmission Control Protocol", STD 7, IETF RFC
793, Sep 1981, <URL:rfc793.txt>=20
how reliable ordered delivery is done i nthe internet
[RFC1272] C. Mills, D. Hirsh, G. Ruth, "Internet Accounting: Background"
IETF RFC 1272, Nov 1991 <URL:ftp://ftp.isi.edu/in-notes/rfc1272.txt>=20
some old ideas on what we might monitor
[RFC1633] R. Braden, D.Clark, S.Shenker, "Integrated Services in the
Internet architecture: an overview", IETF RFC 1633, Jun 1994.
<URL:http://www.isi.edu/div7/rsvp/pub.html>=20
good intro to the (excessivley complex) int serv model of getting
guaranteed or controlled load services
[RFC1905] J. Case, (SNMP Research), K. McCloghrie (cisco), M. Rose (Dover
Beach Consulting) & S. Waldbusser (International Network
Services), "Protocol Operations for Version 2 of the Simple Network
Management Protocol (SNMPv2)", IETF RFC 1905, Jan 1996,
<URL:rfc1905.txt>=20
network management protocol
[RFC1906] J. Case, (SNMP Research), K. McCloghrie (cisco), M. Rose (Dover
Beach Consulting) & S. Waldbusser (International Network
Services), "Transport Mappings for Version 2 of the Simple Network
Management Protocol (SNMPv2)", IETF RFC1906, Jan 1996,
<URL:rfc1906.txt>=20
how network management protocol packets get around
[RFC2001] W. Stevens (NOAO), "TCP Slow Start, Congestion Avoidance, Fast
Retransmit, and Fast Recovery Algorithms" IETF RFC2001, Jan
1997 <URL:rfc2001.txt>=20
the details on Van's 88 paper - what TCP does now
(i.e. how relaible unicast applications react to network problems and
adapt)
[RFC2005] J. Solomon (Motorola), "Applicability Statement for IP Mobility
Support", IETF RFC 2005, Oct 1996, <URL:rfc2005.txt>=20
mobile IP (really only nomadic clients at the moment)
[RFC2063] N. Brownlee, C. Mills, G. Ruth, "Traffic Flow Measurement:
Architecture" IETF RFC 2063, Jan 1997 <URL:rfc2063.txt>=20
yet more metrics
[RFC2208] F. Baker, B. Braden, S. Bradner, A. Mankin, M. O''Dell, A.
Romanow, A. Weinrib, L. Zhang, "Resource ReSerVation Protocol (RSVP)
Version 1 Applicability Statement Some Guidelines on Deployment", IETF RFC
2208, Jan 1997, <URL:rfc2208.txt>=20
where RSVP meets its limits
[RFC2236] B. Fenner (Xerox PARC), "Internet Group Management Protocol,
Version 2", IETF RFC2236, Nov 1997, <URL:rfc2236.txt>=20
[rsvp] Resource Reservation Setup Protocol (rsvp) IETF working group
charter <URL:http://www.ietf.org/html.charters/rsvp-charter.html>=20
[rtfm] Realtime Traffic Flow Measurement (rtfm) IETF working group charter
<URL:http://www.ietf.org/html.charters/rtfm-charter.html>=20
useful for RTP on UDP flows ratehr than TCP models
[Ruth97] Gregory R. Ruth (GTE), "Usage Accounting for the Internet", In
Proc. INET'97 Kuala Lumpur,
<URL:http://www.isoc.org/isoc/whatis/conferences/inet/97/proceedings/F1/F1_1
.HTM>=20
more on ippm and other stuff
[Sarkar95] Mitrabarun Sarkar, "An Assessment of Pricing Mechanisms for the
Internet - A Regulatory Imperative," Journal of Electronic=20
Publication, University of Michigan Press, Mar 1995,
<URL:http://www.press.umich.edu/jep/works/SarkAssess.html>=20
my counter proposal is for Oftel to set a minimum price for Internet
access and remove the price setting for telephony:-)
[Sharkey95] W W Sharkey, "Network Models in Economics", in MO Ball et al
(eds), "Handbook of Operations Research and Management
Science: Networks", Vol 8, pp713-765, 1995=20
background
[Shelton97] Robert Shelton ed., "Common Business Objects", bom/97-12-04
Object Management Group white paper,
ftp://ftp.omg.org/pub/docs/bom/97-12-04.pdf=20
more background
[Shenker95] Scott Shenker (Xerox PARC), "Fundamental Design Issues for the
Future Internet", IEEE Journal on Selected Areas in
Communications, 1995. URL:
<http://www.spp.umich.edu/spp/courses/744/docs/FundamentalDesign-shenker.pdf=
led to some of the diff serve work
>
[Shenker96] Scott Shenker (Xerox PARC), David Clark (MIT), Deborah Estrin
(USC/ISI) and Shai Herzog (USC/ISI), =91Pricing in Computer
Networks: Reshaping the research agenda=92, SIGCOMM Computer Communication
Review Volume 26, Number 2, Apr 1996, <URL:
http://www.statslab.cam.ac.uk/~frank/PRICE/scott.ps>=20
very very important paper - discusses the idea that we should not put
in mechanism that stop any particular policy for pricing (e.g. flat
fee v. usage v. smart market v. anything else)
[Skevington97] Peter J Skevington and Tim P Hart (BT), "Trusted third
parties in electronic commerce", BT Technology Journal, Vol.15, no.2,
p39, Apr 1997, <URL:http://www.labs.bt.com/library/bttj/v15n2/04.htm>
(abstract & order details only) or
<URL:http://sss.info.bt.co.uk/BOOKS_AND_JOURNALS/BTTJ/v15n2/04.htm> (BT
access only)=20
shame - but we know about TTP - ask farez or others in CS...
[Somayaji97] A. Somayaji, S. Hofmeyr, and S. Forrest, "Principles of a
Computer Immune System" New Security Paradigms Workshop, Sep 1997
<URL:http://www.cs.unm.edu/~steveah/papers.html>=20
more neat ideas on securing things
[Speakman98] Tony Speakman, Dino Farinacci, Steven Lin, Alex Tweedly,
(cisco) "PGM Reliable Transport Protocol Specification", Work in
progress: IETF Internet Draft, Jan 1998 (expires Jul 1998),
<URL:draft-speakman-pgm-spec-01.txt>=20
usaeful for some new applications (games, share trafding) but we havnt
thought much about billing for it yet...
[Stoica97] Ion Stoica and Hui Zhang, (CMU), "A Hierarchical Fair Service
Curve Algorithm For Link-Sharing, Real-Time and Priority Services", in
Proc. SIGCOMM'97, Computer Communication Review, ACM SIGCOMM, volume 27,
number 4, ISSN # 0146-4833, Oct 1997 <URL:
http://www.inria.fr/rodeo/sigcomm97/program.html#ab011>=20
fancier scheduling (basically CBQ + WF**2Q(
[Tassel97] J=E9r=F4me Tassel (Uni of Kent at Canterbury), "Quality of Servic=
e
(QoS) adaptation using Reflective Java", MSc dissertation, Sep 1997,
<URL:http://www.labs.bt.com/people/briscorj/projects/lsma/jt_thesis/thesis_f
inal.htm>=20
interesting work.....
[TasselBri97] J=E9r=F4me Tassel, Bob Briscoe, Alan Smith, (BT), "An End to E=
nd
Price-Based QoS Control Component Using Reflective Java", in
Lecture Notes in Computer Science from the 4th COST237 workshop, pub.
Springer-Verlag, Dec 1997,
<URL:http://www.labs.bt.com/people/briscorj/papers.html#QoteS>=20
hmmmm - yes - good ideas.......
[Thomas96] Stephen A. Thomas, "IPng and the TCP/IP Protocols : Implementing
the Next Generation Internet", pub Wiley, ISBN: 0471130885,
Jan 1996=20
details on v6 - useful but apart from increased address space, not
sure of relevance much....
[Tsirtsis98] George Tsirtsis (BT), "RADIUS - DIAMETER Comparison", March=
1998
<URL:http://www.gideon.bt.co.uk/george/radvsdiam/RADvsDIAM.htm> (BT access
only)=20
worj a look - i think there's an i-d on this
[Varian96] Hal R Varian, "Differential Pricing and Efficiency", f =A1 =AE s=
T -
m o =F1 d @ =A5 , Vol.1 No.2, Aug 1996,
<URL:http://www.firstmonday.dk/issues/issue2/different/>=20
update on smart markets and index work
[Xiaoming96] Hao Xiaoming (Nanyang Technologica University, Singapore),
Huang Yu (Baptist University, Hong Kong) & Kewen Zhang
(University of Missouri), "The Internet and Information Control: The Case
of China", in Proc. INET'96,=20
<URL:http://cios.llc.rpi.edu/getfile/Xiaoming_V6N296> (subscription=
required)=20
total control freak approach to access and $$$
[Zhang93] Lixia Zhang (Xerox PARC), Stephen Deering (Xerox PARC), Deborah
Estrin(USC/ISI), Scott Shenker (Xerox PARC) and Daniel
Zappala (USC/CSD), "RSVP: A New Resource ReSerVation Protocol", IEEE
Network. Sep 1993. <URL:http://www.isi.edu/div7/rsvp/pub.html>=20
more about signaling.......