RFC relevance of congestion

As per Relevance of the word congestion, we have this rfc below:

Network Working Group D.
Request for Comments: 3272 Movaz
Category: Informational A.
Celion
A.
I.
Lucent
X.
Redback
May 2002

Overview and Principles of Internet Traffic

Status of this

This memo provides information for the Internet community. It
not specify an Internet standard of any kind. Distribution of
memo is unlimited

Copyright

Copyright (C) The Internet Society (2002). All Rights Reserved

This memo describes the principles of Traffic Engineering (TE) in
Internet. The document is intended to promote better
of the issues surrounding traffic engineering in IP networks, and
provide a common basis for the development of traffic
capabilities for the Internet. The principles, architectures,
methodologies for performance evaluation and performance
of operational IP networks are discussed throughout this document

Table of

1.0 Introduction...................................................3
1.1 What is Internet Traffic Engineering?.......................4
1.2 Scope.......................................................7
1.3 Terminology.................................................8
2.0 Background....................................................11
2.1 Context of Internet Traffic Engineering....................12
2.2 Network Context............................................13
2.3 Problem Context............................................14
2.3.1 Congestion and its Ramifications......................16
2.4 Solution Context...........................................16
2.4.1 Combating the Congestion Problem......................18
2.5 Implementation and Operational Context.....................21

Awduche, et. al. Informational [Page 1]

RFC 3272 Overview and Principles of Internet TE May 2002

3.0 Traffic Engineering Process Model.............................21
3.1 Components of the Traffic Engineering Process Model........23
3.2 Measurement................................................23
3.3 Modeling, Analysis, and Simulation.........................24
3.4 Optimization...............................................25
4.0 Historical Review and Recent Developments.....................26
4.1 Traffic Engineering in Classical Telephone Networks........26
4.2 Evolution of Traffic Engineering in the Internet...........28
4.2.1 Adaptive Routing in ARPANET...........................28
4.2.2 Dynamic Routing in the Internet.......................29
4.2.3 ToS Routing...........................................30
4.2.4 Equal Cost Multi-Path.................................30
4.2.5 Nimrod................................................31
4.3 Overlay Model..............................................31
4.4 Constraint-Based Routing...................................32
4.5 Overview of Other IETF Projects Related to
Engineering................................................32
4.5.1 Integrated Services...................................32
4.5.2 RSVP..................................................33
4.5.3 Differentiated Services...............................34
4.5.4 MPLS..................................................35
4.5.5 IP Performance Metrics................................36
4.5.6 Flow Measurement......................................37
4.5.7 Endpoint Congestion Management........................37
4.6 Overview of ITU Activities Related to
Engineering................................................38
4.7 Content Distribution.......................................39
5.0 Taxonomy of Traffic Engineering Systems.......................40
5.1 Time-Dependent Versus State-Dependent......................40
5.2 Offline Versus Online......................................41
5.3 Centralized Versus Distributed.............................42
5.4 Local Versus Global........................................42
5.5 Prescriptive Versus Descriptive............................42
5.6 Open-Loop Versus Closed-Loop...............................43
5.7 Tactical vs Strategic......................................43
6.0 Recommendations for Internet Traffic Engineering..............43
6.1 Generic Non-functional Recommendations.....................44
6.2 Routing Recommendations....................................46
6.3 Traffic Mapping Recommendations............................48
6.4 Measurement Recommendations................................49
6.5 Network Survivability......................................50
6.5.1 Survivability in MPLS Based Networks..................52
6.5.2 Protection Option.....................................53
6.6 Traffic Engineering in Diffserv Environments...............54
6.7 Network Controllability....................................56
7.0 Inter-Domain Considerations...................................57
8.0 Overview of Contemporary TE Practices in
IP Networks...................................................59

Awduche, et. al. Informational [Page 2]

RFC 3272 Overview and Principles of Internet TE May 2002

9.0 Conclusion....................................................63
10.0 Security Considerations......................................63
11.0 Acknowledgments..............................................63
12.0 References...................................................64
13.0 Authors' Addresses...........................................70
14.0 Full Copyright Statement.....................................71

1.0

This memo describes the principles of Internet traffic engineering
The objective of the document is to articulate the general issues
principles for Internet traffic engineering; and where appropriate
provide recommendations, guidelines, and options for the
of online and offline Internet traffic engineering capabilities
support systems

This document can aid service providers in devising and
traffic engineering solutions for their networks.
hardware and software vendors will also find this document helpful
the development of mechanisms and support systems for the
environment that support the traffic engineering function

This document provides a terminology for describing and
common Internet traffic engineering concepts. This document
provides a taxonomy of known traffic engineering styles. In
context, a traffic engineering style abstracts important aspects
a traffic engineering methodology. Traffic engineering styles can
viewed in different ways depending upon the specific context in
they are used and the specific purpose which they serve.
combination of styles and views results in a natural taxonomy
traffic engineering systems

Even though Internet traffic engineering is most effective
applied end-to-end, the initial focus of this document document
intra-domain traffic engineering (that is, traffic engineering
a given autonomous system). However, because a preponderance
Internet traffic tends to be inter-domain (originating in
autonomous system and terminating in another), this document
an overview of aspects pertaining to inter-domain
engineering

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in
document are to be interpreted as described in RFC 2119.

Awduche, et. al. Informational [Page 3]

RFC 3272 Overview and Principles of Internet TE May 2002

1.1. What is Internet Traffic Engineering

Internet traffic engineering is defined as that aspect of
network engineering dealing with the issue of performance
and performance optimization of operational IP networks.
Engineering encompasses the application of technology and
principles to the measurement, characterization, modeling,
control of Internet traffic [RFC-2702, AWD2].

Enhancing the performance of an operational network, at both
traffic and resource levels, are major objectives of Internet
engineering. This is accomplished by addressing traffic
performance requirements, while utilizing network
economically and reliably. Traffic oriented performance
include delay, delay variation, packet loss, and throughput

An important objective of Internet traffic engineering is
facilitate reliable network operations [RFC-2702]. Reliable
operations can be facilitated by providing mechanisms that
network integrity and by embracing policies emphasizing
survivability. This results in a minimization of the
of the network to service outages arising from errors, faults,
failures occurring within the infrastructure

The Internet exists in order to transfer information from
nodes to destination nodes. Accordingly, one of the most
functions performed by the Internet is the routing of traffic
ingress nodes to egress nodes. Therefore, one of the
distinctive functions performed by Internet traffic engineering
the control and optimization of the routing function, to
traffic through the network in the most effective way

Ultimately, it is the performance of the network as seen by end
of network services that is truly paramount. This crucial
should be considered throughout the development of
engineering mechanisms and policies. The characteristics visible
end users are the emergent properties of the network, which are
characteristics of the network when viewed as a whole. A
goal of the service provider, therefore, is to enhance the
properties of the network while taking economic considerations
account

The importance of the above observation regarding the
properties of networks is that special care must be taken
choosing network performance measures to optimize. Optimizing
wrong measures may achieve certain local objectives, but may

Awduche, et. al. Informational [Page 4]

RFC 3272 Overview and Principles of Internet TE May 2002

disastrous consequences on the emergent properties of the network
thereby on the quality of service perceived by end-users of
services

A subtle, but practical advantage of the systematic application
traffic engineering concepts to operational networks is that it
to identify and structure goals and priorities in terms of
the quality of service delivered to end-users of network services
The application of traffic engineering concepts also aids in
measurement and analysis of the achievement of these goals

The optimization aspects of traffic engineering can be
through capacity management and traffic management. As used in
document, capacity management includes capacity planning,
control, and resource management. Network resources of
interest include link bandwidth, buffer space, and
resources. Likewise, as used in this document, traffic
includes (1) nodal traffic control functions such as
conditioning, queue management, scheduling, and (2) other
that regulate traffic flow through the network or that
access to network resources between different packets or
different traffic streams

The optimization objectives of Internet traffic engineering should
viewed as a continual and iterative process of network
improvement and not simply as a one time goal. Traffic
also demands continual development of new technologies and
methodologies for network performance enhancement

The optimization objectives of Internet traffic engineering
change over time as new requirements are imposed, as new
emerge, or as new insights are brought to bear on the
problems. Moreover, different networks may have
optimization objectives, depending upon their business models
capabilities, and operating constraints. The optimization aspects
traffic engineering are ultimately concerned with network
regardless of the specific optimization goals in any
environment

Thus, the optimization aspects of traffic engineering can be
from a control perspective. The aspect of control within
Internet traffic engineering arena can be pro-active and/or reactive
In the pro-active case, the traffic engineering control system
preventive action to obviate predicted unfavorable future
states. It may also take perfective action to induce a
desirable state in the future. In the reactive case, the
system responds correctively and perhaps adaptively to events
have already transpired in the network

Awduche, et. al. Informational [Page 5]

RFC 3272 Overview and Principles of Internet TE May 2002

The control dimension of Internet traffic engineering responds
multiple levels of temporal resolution to network events.
aspects of capacity management, such as capacity planning, respond
very coarse temporal levels, ranging from days to possibly years
The introduction of automatically switched optical transport
(e.g., based on the Multi-protocol Lambda Switching concepts)
significantly reduce the lifecycle for capacity planning
expediting provisioning of optical bandwidth. Routing
functions operate at intermediate levels of temporal resolution
ranging from milliseconds to days. Finally, the packet
processing functions (e.g., rate shaping, queue management,
scheduling) operate at very fine levels of temporal resolution
ranging from picoseconds to milliseconds while responding to
real-time statistical behavior of traffic. The subsystems
Internet traffic engineering control include: capacity augmentation
routing control, traffic control, and resource control (
control of service policies at network elements). When capacity
to be augmented for tactical purposes, it may be desirable to
a deployment plan that expedites bandwidth provisioning
minimizing installation costs

Inputs into the traffic engineering control system include
state variables, policy variables, and decision variables

One major challenge of Internet traffic engineering is
realization of automated control capabilities that adapt quickly
cost effectively to significant changes in a network's state,
still maintaining stability

Another critical dimension of Internet traffic engineering is
performance evaluation, which is important for assessing
effectiveness of traffic engineering methods, and for monitoring
verifying compliance with network performance goals. Results
performance evaluation can be used to identify existing problems
guide network re-optimization, and aid in the prediction of
future problems

Performance evaluation can be achieved in many different ways.
most notable techniques include analytical methods, simulation,
empirical methods based on measurements. When analytical methods
simulation are used, network nodes and links can be modeled
capture relevant operational features such as topology, bandwidth
buffer space, and nodal service policies (link scheduling,
prioritization, buffer management, etc.). Analytical traffic
can be used to depict dynamic and behavioral traffic characteristics
such as burstiness, statistical distributions, and dependence

Awduche, et. al. Informational [Page 6]

RFC 3272 Overview and Principles of Internet TE May 2002

Performance evaluation can be quite complicated in practical
contexts. A number of techniques can be used to simplify
analysis, such as abstraction, decomposition, and approximation.
example, simplifying concepts such as effective bandwidth
effective buffer [Elwalid] may be used to approximate nodal
at the packet level and simplify the analysis at the
level. Network analysis techniques using, for example,
models and approximation schemes based on asymptotic
decomposition techniques can render the analysis even more tractable
In particular, an emerging set of concepts known as network
[CRUZ] based on deterministic bounds may simplify network
relative to classical stochastic techniques. When using
techniques, care should be taken to ensure that the models
reflect the relevant operational characteristics of the
network entities

Simulation can be used to evaluate network performance or to
and validate analytical approximations. Simulation can, however,
computationally costly and may not always provide
insights. An appropriate approach to a given network
evaluation problem may involve a hybrid combination of
techniques, simulation, and empirical methods

As a general rule, traffic engineering concepts and mechanisms
be sufficiently specific and well defined to address
requirements, but simultaneously flexible and extensible
accommodate unforeseen future demands

1.2.

The scope of this document is intra-domain traffic engineering;
is, traffic engineering within a given autonomous system in
Internet. This document will discuss concepts pertaining to intra
domain traffic control, including such issues as routing control
micro and macro resource allocation, and the control
problems that arise consequently

This document will describe and characterize techniques already
use or in advanced development for Internet traffic engineering.
way these techniques fit together will be discussed and scenarios
which they are useful will be identified

While this document considers various intra-domain
engineering approaches, it focuses more on traffic engineering
MPLS. Traffic engineering based upon manipulation of IGP metrics
not addressed in detail. This topic may be addressed by
working group document(s).

Awduche, et. al. Informational [Page 7]

RFC 3272 Overview and Principles of Internet TE May 2002

Although the emphasis is on intra-domain traffic engineering,
Section 7.0, an overview of the high level considerations
to inter-domain traffic engineering will be provided. Inter-
Internet traffic engineering is crucial to the
enhancement of the global Internet infrastructure

Whenever possible, relevant requirements from existing IETF
and other sources will be incorporated by reference

1.3

This subsection provides terminology which is useful for
traffic engineering. The definitions presented apply to
document. These terms may have other meanings elsewhere

- Baseline analysis
A study conducted to serve as a baseline for comparison
the actual behavior of the network

- Busy hour
A one hour period within a specified interval of
(typically 24 hours) in which the traffic load in a
or sub-network is greatest

- Bottleneck
A network element whose input traffic rate tends to
greater than its output rate

- Congestion
A state of a network resource in which the traffic
on the resource exceeds its output capacity over an
of time

- Congestion avoidance
An approach to congestion management that attempts
obviate the occurrence of congestion

- Congestion control
An approach to congestion management that attempts to
congestion problems that have already occurred

- Constraint-based routing
A class of routing protocols that take specified
attributes, network constraints, and policy constraints
account when making routing decisions. Constraint-
routing is applicable to traffic aggregates as well
flows. It is a generalization of QoS routing

Awduche, et. al. Informational [Page 8]

RFC 3272 Overview and Principles of Internet TE May 2002

- Demand side congestion management
A congestion management scheme that addresses
problems by regulating or conditioning offered load

- Effective bandwidth
The minimum amount of bandwidth that can be assigned to
flow or traffic aggregate in order to deliver '
service quality' to the flow or traffic aggregate

- Egress traffic
Traffic exiting a network or network element

- Hot-spot
A network element or subsystem which is in a state
congestion

- Ingress traffic
Traffic entering a network or network element

- Inter-domain traffic
Traffic that originates in one Autonomous system
terminates in another

- Loss network
A network that does not provide adequate buffering
traffic, so that traffic entering a busy resource within
network will be dropped rather than queued

- Metric
A parameter defined in terms of standard units
measurement

- Measurement Methodology
A repeatable measurement technique used to derive one
more metrics of interest

- Network Survivability
The capability to provide a prescribed level of QoS
existing services after a given number of failures
within the network

- Offline traffic engineering
A traffic engineering system that exists outside of
network

Awduche, et. al. Informational [Page 9]

RFC 3272 Overview and Principles of Internet TE May 2002

- Online traffic engineering
A traffic engineering system that exists within the network
typically implemented on or as adjuncts to
network elements

- Performance measures
Metrics that provide quantitative or qualitative measures
the performance of systems or subsystems of interest

- Performance management
A systematic approach to improving effectiveness in
accomplishment of specific networking goals related
performance improvement

- Performance Metric
A performance parameter defined in terms of standard
of measurement

- Provisioning
The process of assigning or configuring network resources
meet certain requests

- QoS routing
Class of routing systems that selects paths to be used by
flow based on the QoS requirements of the flow

- Service Level Agreement
A contract between a provider and a customer that
specific levels of performance and reliability at a
cost

- Stability
An operational state in which a network does not
in a disruptive manner from one mode to another mode

- Supply side congestion management
A congestion management scheme that provisions
network resources to address existing and/or
congestion problems

- Transit traffic
Traffic whose origin and destination are both outside of
network under consideration

- Traffic characteristic
A description of the temporal behavior or a description
the attributes of a given traffic flow or traffic aggregate

Awduche, et. al. Informational [Page 10]

RFC 3272 Overview and Principles of Internet TE May 2002

- Traffic engineering system
A collection of objects, mechanisms, and protocols that
used conjunctively to accomplish traffic
objectives

- Traffic flow
A stream of packets between two end-points that can
characterized in a certain way. A micro-flow has a
specific definition: A micro-flow is a stream of
with the same source and destination addresses, source
destination ports, and protocol ID

- Traffic intensity
A measure of traffic loading with respect to a
capacity over a specified period of time. In
telephony systems, traffic intensity is measured in units
Erlang

- Traffic matrix
A representation of the traffic demand between a set
origin and destination abstract nodes. An abstract node
consist of one or more network elements

- Traffic monitoring
The process of observing traffic characteristics at a
point in a network and collecting the traffic
for analysis and further action

- Traffic trunk
An aggregation of traffic flows belonging to the same
which are forwarded through a common path. A traffic
may be characterized by an ingress and egress node, and
set of attributes which determine its
characteristics and requirements from the network

2.0

The Internet has quickly evolved into a very critical
infrastructure, supporting significant economic, educational,
social activities. Simultaneously, the delivery of
communications services has become very competitive and end-users
demanding very high quality service from their service providers
Consequently, performance optimization of large scale IP networks
especially public Internet backbones, have become an
problem. Network performance requirements are multi-dimensional
complex, and sometimes contradictory; making the traffic
problem very challenging

Awduche, et. al. Informational [Page 11]

RFC 3272 Overview and Principles of Internet TE May 2002

The network must convey IP packets from ingress nodes to egress
efficiently, expeditiously, and economically. Furthermore, in
multiclass service environment (e.g., Diffserv capable networks),
resource sharing parameters of the network must be
determined and configured according to prevailing policies
service models to resolve resource contention issues arising
mutual interference between packets traversing through the network
Thus, consideration must be given to resolving competition
network resources between traffic streams belonging to the
service class (intra-class contention resolution) and traffic
belonging to different classes (inter-class contention resolution).

2.1 Context of Internet Traffic

The context of Internet traffic engineering pertains to the
where traffic engineering is used. A traffic engineering
establishes appropriate rules to resolve traffic performance
occurring in a specific context. The context of Internet
engineering includes

(1) A network context defining the universe of discourse, and
particular the situations in which the traffic
problems occur. The network context includes
structure, network policies, network characteristics
network constraints, network quality attributes, and
optimization criteria

(2) A problem context defining the general and concrete
that traffic engineering addresses. The problem
includes identification, abstraction of relevant features
representation, formulation, specification of
requirements on the solution space, and specification of
desirable features of acceptable solutions

(3) A solution context suggesting how to address the
identified by the problem context. The solution
includes analysis, evaluation of alternatives, prescription
and resolution

(4) An implementation and operational context in which
solutions are methodologically instantiated.
implementation and operational context includes planning
organization, and execution

The context of Internet traffic engineering and the different
scenarios are discussed in the following subsections

Awduche, et. al. Informational [Page 12]

RFC 3272 Overview and Principles of Internet TE May 2002

2.2 Network

IP networks range in size from small clusters of routers
within a given location, to thousands of interconnected routers
switches, and other components distributed all over the world

Conceptually, at the most basic level of abstraction, an IP
can be represented as a distributed dynamical system consisting of
(1) a set of interconnected resources which provide
services for IP traffic subject to certain constraints, (2) a
system representing the offered load to be transported through
network, and (3) a response system consisting of network processes
protocols, and related mechanisms which facilitate the movement
traffic through the network [see also AWD2].

The network elements and resources may have specific
restricting the manner in which the demand is handled. Additionally
network resources may be equipped with traffic control
superintending the way in which the demand is serviced.
control mechanisms may, for example, be used to control
packet processing activities within a given resource,
contention for access to the resource by different packets,
regulate traffic behavior through the resource. A
management and provisioning system may allow the settings of
traffic control mechanisms to be manipulated by external or
entities in order to exercise control over the way in which
network elements respond to internal and external stimuli

The details of how the network provides transport services
packets are specified in the policies of the network
and are installed through network configuration management and
based provisioning systems. Generally, the types of
provided by the network also depends upon the technology
characteristics of the network elements and protocols, the
service and utility models, and the ability of the
administrators to translate policies into network configurations

Contemporary Internet networks have three
characteristics: (1) they provide real-time services, (2) they
become mission critical, and (3) their operating environments
very dynamic. The dynamic characteristics of IP networks can
attributed in part to fluctuations in demand, to the
between various network protocols and processes, to the
evolution of the infrastructure which demands the constant
of new technologies and new network elements, and to transient
persistent impairments which occur within the system

Awduche, et. al. Informational [Page 13]

RFC 3272 Overview and Principles of Internet TE May 2002

Packets contend for the use of network resources as they are
through the network. A network resource is considered to
congested if the arrival rate of packets exceed the output
of the resource over an interval of time. Congestion may result
some of the arrival packets being delayed or even dropped

Congestion increases transit delays, delay variation, packet loss
and reduces the predictability of network services. Clearly
congestion is a highly undesirable phenomenon

Combating congestion at a reasonable cost is a major objective
Internet traffic engineering

Efficient sharing of network resources by multiple traffic streams
a basic economic premise for packet switched networks in general
for the Internet in particular. A fundamental challenge in
operation, especially in a large scale public IP network, is
increase the efficiency of resource utilization while minimizing
possibility of congestion

Increasingly, the Internet will have to function in the presence
different classes of traffic with different service requirements
The advent of Differentiated Services [RFC-2475] makes
requirement particularly acute. Thus, packets may be grouped
behavior aggregates such that each behavior aggregate may have
common set of behavioral characteristics or a common set of
requirements. In practice, the delivery requirements of a
set of packets may be specified explicitly or implicitly. Two of
most important traffic delivery requirements are capacity
and QoS constraints

Capacity constraints can be expressed statistically as peak rates
mean rates, burst sizes, or as some deterministic notion of
bandwidth. QoS requirements can be expressed in terms of (1)
integrity constraints such as packet loss and (2) in terms
temporal constraints such as timing restrictions for the delivery
each packet (delay) and timing restrictions for the delivery
consecutive packets belonging to the same traffic stream (
variation).

2.3 Problem

Fundamental problems exist in association with the operation of
network described by the simple model of the previous subsection
This subsection reviews the problem context in relation to
traffic engineering function

Awduche, et. al. Informational [Page 14]

RFC 3272 Overview and Principles of Internet TE May 2002

The identification, abstraction, representation, and measurement
network features relevant to traffic engineering is a
issue

One particularly important class of problems concerns how
explicitly formulate the problems that traffic engineering
to solve, how to identify the requirements on the solution space,
to specify the desirable features of good solutions, how to
solve the problems, and how to measure and characterize
effectiveness of the solutions

Another class of problems concerns how to measure and
relevant network state parameters. Effective traffic
relies on a good estimate of the offered traffic load as well as
view of the underlying topology and associated resource constraints
A network-wide view of the topology is also a must for
planning

Still another class of problems concerns how to characterize
state of the network and how to evaluate its performance under
variety of scenarios. The performance evaluation problem is two
fold. One aspect of this problem relates to the evaluation of
system level performance of the network. The other aspect relates
the evaluation of the resource level performance, which
attention to the performance analysis of individual
resources. In this memo, we refer to the system
characteristics of the network as the "macro-states" and the
level characteristics as the "micro-states." The system
characteristics are also known as the emergent properties of
network as noted earlier. Correspondingly, we shall refer to
traffic engineering schemes dealing with network
optimization at the systems level as "macro-TE" and the schemes
optimize at the individual resource level as "micro-TE."
certain circumstances, the system level performance can be
from the resource level performance using appropriate rules
composition, depending upon the particular performance measures
interest

Another fundamental class of problems concerns how to
optimize network performance. Performance optimization may
translating solutions to specific traffic engineering problems
network configurations. Optimization may also entail some degree
resource management control, routing control, and/or
augmentation

Awduche, et. al. Informational [Page 15]

RFC 3272 Overview and Principles of Internet TE May 2002

As noted previously, congestion is an undesirable phenomena
operational networks. Therefore, the next subsection addresses
issue of congestion and its ramifications within the problem
of Internet traffic engineering

2.3.1 Congestion and its

Congestion is one of the most significant problems in an
IP context. A network element is said to be congested if
experiences sustained overload over an interval of time.
almost always results in degradation of service quality to end users
Congestion control schemes can include demand side policies
supply side policies. Demand side policies may restrict access
congested resources and/or dynamically regulate the demand
alleviate the overload situation. Supply side policies may expand
augment network capacity to better accommodate offered traffic
Supply side policies may also re-allocate network resources
redistributing traffic over the infrastructure.
redistribution and resource re-allocation serve to increase
'effective capacity' seen by the demand

The emphasis of this memo is primarily on congestion
schemes falling within the scope of the network, rather than
congestion management systems dependent upon sensitivity
adaptivity from end-systems. That is, the aspects that
considered in this memo with respect to congestion management
those solutions that can be provided by control entities operating
the network and by the actions of network administrators and
operations systems

2.4 Solution

The solution context for Internet traffic engineering
analysis, evaluation of alternatives, and choice between
courses of action. Generally the solution context is predicated
making reasonable inferences about the current or future state of
network, and subsequently making appropriate decisions that
involve a preference between alternative sets of action.
specifically, the solution context demands reasonable estimates
traffic workload, characterization of network state,
solutions to traffic engineering problems which may be implicitly
explicitly formulated, and possibly instantiating a set of
actions. Control actions may involve the manipulation of
associated with routing, control over tactical capacity acquisition
and control over the traffic management functions

The following list of instruments may be applicable to the
context of Internet traffic engineering

Awduche, et. al. Informational [Page 16]

RFC 3272 Overview and Principles of Internet TE May 2002

(1) A set of policies, objectives, and requirements (which
be context dependent) for network performance evaluation
performance optimization

(2) A collection of online and possibly offline tools
mechanisms for measurement, characterization, modeling,
control of Internet traffic and control over the
and allocation of network resources, as well as control
the mapping or distribution of traffic onto
infrastructure

(3) A set of constraints on the operating environment,
network protocols, and the traffic engineering
itself

(4) A set of quantitative and qualitative techniques
methodologies for abstracting, formulating, and
traffic engineering problems

(5) A set of administrative control parameters which may
manipulated through a Configuration Management (CM) system
The CM system itself may include a configuration
subsystem, a configuration repository, a
accounting subsystem, and a configuration
subsystem

(6) A set of guidelines for network performance evaluation
performance optimization, and performance improvement

Derivation of traffic characteristics through measurement and/
estimation is very useful within the realm of the solution space
traffic engineering. Traffic estimates can be derived from
subscription information, traffic projections, traffic models,
from actual empirical measurements. The empirical measurements
be performed at the traffic aggregate level or at the flow level
order to derive traffic statistics at various levels of detail
Measurements at the flow level or on small traffic aggregates may
performed at edge nodes, where traffic enters and leaves the network
Measurements at large traffic aggregate levels may be
within the core of the network where potentially numerous
flows may be in transit concurrently

To conduct performance studies and to support planning of
and future networks, a routing analysis may be performed to
the path(s) the routing protocols will choose for various
demands, and to ascertain the utilization of network resources
traffic is routed through the network. The routing analysis
capture the selection of paths through the network, the assignment

Awduche, et. al. Informational [Page 17]

RFC 3272 Overview and Principles of Internet TE May 2002

traffic across multiple feasible routes, and the multiplexing of
traffic over traffic trunks (if such constructs exists) and over
underlying network infrastructure. A network topology model is
necessity for routing analysis. A network topology model may
extracted from network architecture documents, from network designs
from information contained in router configuration files,
routing databases, from routing tables, or from automated tools
discover and depict network topology information.
information may also be derived from servers that monitor
state, and from servers that perform provisioning functions

Routing in operational IP networks can be administratively
at various levels of abstraction including the manipulation of
attributes and manipulation of IGP metrics. For path
technologies such as MPLS, routing can be further controlled by
manipulation of relevant traffic engineering parameters,
parameters, and administrative policy constraints. Within
context of MPLS, the path of an explicit label switched path (LSP
can be computed and established in various ways including: (1)
manually, (2) automatically online using constraint-based
processes implemented on label switching routers, and (3)
automatically offline using constraint-based routing
implemented on external traffic engineering support systems

2.4.1 Combating the Congestion

Minimizing congestion is a significant aspect of Internet
engineering. This subsection gives an overview of the
approaches that have been used or proposed to combat
problems

Congestion management policies can be categorized based upon
following criteria (see e.g., [YARE95] for a more detailed
of congestion control schemes): (1) Response time scale which can
characterized as long, medium, or short; (2) reactive
preventive which relates to congestion control and
avoidance; and (3) supply side versus demand side
management schemes. These aspects are discussed in the
paragraphs

(1) Congestion Management based on Response Time

- Long (weeks to months): Capacity planning works over a
long time scale to expand network capacity based on estimates
forecasts of future traffic demand and traffic distribution.
router and link provisioning take time and are generally expensive
these upgrades are typically carried out in the weeks-to-months
even years time scale

Awduche, et. al. Informational [Page 18]

RFC 3272 Overview and Principles of Internet TE May 2002

- Medium (minutes to days): Several control policies fall within
medium time scale category. Examples include: (1) Adjusting
and/or BGP parameters to route traffic away or towards
segments of the network; (2) Setting up and/or adjusting
explicitly routed label switched paths (ER-LSPs) in MPLS networks
route some traffic trunks away from possibly congested resources
towards possibly more favorable routes; (3) re-configuring
logical topology of the network to make it correlate more
with the spatial traffic distribution using for example
underlying path-oriented technology such as MPLS LSPs, ATM PVCs,
optical channel trails. Many of these adaptive medium time
response schemes rely on a measurement system that monitors
in traffic distribution, traffic shifts, and network
utilization and subsequently provides feedback to the online and/
offline traffic engineering mechanisms and tools which employ
feedback information to trigger certain control actions to
within the network. The traffic engineering mechanisms and tools
be implemented in a distributed fashion or in a centralized fashion
and may have a hierarchical structure or a flat structure.
comparative merits of distributed and centralized control
for networks are well known. A centralized scheme may have
visibility into the network state and may produce potentially
optimal solutions. However, centralized schemes are prone to
points of failure and may not scale as well as distributed schemes
Moreover, the information utilized by a centralized scheme may
stale and may not reflect the actual state of the network. It is
an objective of this memo to make a recommendation
distributed and centralized schemes. This is a choice that
administrators must make based on their specific needs

- Short (picoseconds to minutes): This category includes packet
processing functions and events on the order of several round
times. It includes router mechanisms such as passive and
buffer management. These mechanisms are used to control
and/or signal congestion to end systems so that they can
regulate the rate at which traffic is injected into the network.
of the most popular active queue management schemes, especially
TCP traffic, is Random Early Detection (RED) [FLJA93], which
congestion avoidance by controlling the average queue size.
congestion (but before the queue is filled), the RED scheme
arriving packets to "mark" according to a probabilistic
which takes into account the average queue size. For a router
does not utilize explicit congestion notification (ECN) see e.g.,
[FLOY94], the marked packets can simply be dropped to signal
inception of congestion to end systems. On the other hand, if
router supports ECN, then it can set the ECN field in the
header. Several variations of RED have been proposed to
different drop precedence levels in multi-class environments [RFC

Awduche, et. al. Informational [Page 19]

RFC 3272 Overview and Principles of Internet TE May 2002

2597], e.g., RED with In and Out (RIO) and Weighted RED. There
general consensus that RED provides congestion avoidance
which is not worse than traditional Tail-Drop (TD) queue
(drop arriving packets only when the queue is full). Importantly
however, RED reduces the possibility of global synchronization
improves fairness among different TCP sessions. However, RED
itself can not prevent congestion and unfairness caused by
unresponsive to RED, e.g., UDP traffic and some misbehaved
connections. Other schemes have been proposed to improve
performance and fairness in the presence of unresponsive traffic
Some of these schemes were proposed as theoretical frameworks and
typically not available in existing commercial products. Two
schemes are Longest Queue Drop (LQD) and Dynamic Soft
with Random Drop (RND) [SLDC98].

(2) Congestion Management: Reactive versus Preventive

- Reactive: reactive (recovery) congestion management policies
to existing congestion problems to improve it. All the
described in the long and medium time scales above can be
as being reactive especially if the policies are based on
and identifying existing congestion problems, and on the
of relevant actions to ease a situation

- Preventive: preventive (predictive/avoidance) policies
proactive action to prevent congestion based on estimates
predictions of future potential congestion problems. Some of
policies described in the long and medium time scales fall into
category. They do not necessarily respond immediately to
congestion problems. Instead forecasts of traffic demand
workload distribution are considered and action may be taken
prevent potential congestion problems in the future. The
described in the short time scale (e.g., RED and its variations, ECN
LQD, and RND) are also used for congestion avoidance since
or marking packets before queues actually overflow would
corresponding TCP sources to slow down

(3) Congestion Management: Supply Side versus Demand Side

- Supply side: supply side congestion management policies
the effective capacity available to traffic in order to control
obviate congestion. This can be accomplished by augmenting capacity
Another way to accomplish this is to minimize congestion by having
relatively balanced distribution of traffic over the network.
example, capacity planning should aim to provide a physical
and associated link bandwidths that match estimated traffic
and traffic distribution based on forecasting (subject to
and other constraints). However, if actual traffic distribution

Awduche, et. al. Informational [Page 20]

RFC 3272 Overview and Principles of Internet TE May 2002

not match the topology derived from capacity panning (due
forecasting errors or facility constraints for example), then
traffic can be mapped onto the existing topology using
control mechanisms, using path oriented technologies (e.g., MPLS
and optical channel trails) to modify the logical topology, or
using some other load redistribution mechanisms

- Demand side: demand side congestion management policies control
regulate the offered traffic to alleviate congestion problems.
example, some of the short time scale mechanisms described
(such as RED and its variations, ECN, LQD, and RND) as well
policing and rate shaping mechanisms attempt to regulate the
load in various ways. Tariffs may also be applied as a demand
instrument. To date, however, tariffs have not been used as a
of demand side congestion management within the Internet

In summary, a variety of mechanisms can be used to address
problems in IP networks. These mechanisms may operate at
time-scales

2.5 Implementation and Operational

The operational context of Internet traffic engineering
characterized by constant change which occur at multiple levels
abstraction. The implementation context demands effective planning
organization, and execution. The planning aspects may
determining prior sets of actions to achieve desired objectives
Organizing involves arranging and assigning responsibility to
various components of the traffic engineering system and
the activities to accomplish the desired TE objectives.
involves measuring and applying corrective or perfective actions
attain and maintain desired TE goals

3.0 Traffic Engineering Process Model(s

This section describes a generic process model that captures the
level practical aspects of Internet traffic engineering in
operational context. The process model is described as a sequence
actions that a traffic engineer, or more generally a
engineering system, must perform to optimize the performance of
operational network (see also [RFC-2702, AWD2]). The process
described here represents the broad activities common to most
engineering methodologies although the details regarding how
engineering is executed may differ from network to network.
process model may be enacted explicitly or implicitly, by
automaton and/or by a human

Awduche, et. al. Informational [Page 21]

RFC 3272 Overview and Principles of Internet TE May 2002

The traffic engineering process model is iterative [AWD2]. The
phases of the process model described below are repeated continually

The first phase of the TE process model is to define the
control policies that govern the operation of the network.
policies may depend upon many factors including the
business model, the network cost structure, the
constraints, the utility model, and optimization criteria

The second phase of the process model is a feedback
involving the acquisition of measurement data from the
network. If empirical data is not readily available from
network, then synthetic workloads may be used instead which
either the prevailing or the expected workload of the network
Synthetic workloads may be derived by estimation or
using prior empirical data. Their derivation may also be
using mathematical models of traffic characteristics or other means

The third phase of the process model is to analyze the network
and to characterize traffic workload. Performance analysis may
proactive and/or reactive. Proactive performance analysis
potential problems that do not exist, but could manifest in
future. Reactive performance analysis identifies existing problems
determines their cause through diagnosis, and evaluates
approaches to remedy the problem, if necessary. A number
quantitative and qualitative techniques may be used in the
process, including modeling based analysis and simulation.
analysis phase of the process model may involve investigating
concentration and distribution of traffic across the network
relevant subsets of the network, identifying the characteristics
the offered traffic workload, identifying existing or
bottlenecks, and identifying network pathologies such as
link placement, single points of failures, etc. Network
may result from many factors including inferior network architecture
inferior network design, and configuration problems. A
matrix may be constructed as part of the analysis process.
analysis may also be descriptive or prescriptive

The fourth phase of the TE process model is the
optimization of the network. The performance optimization
involves a decision process which selects and implements a set
actions from a set of alternatives. Optimization actions may
the use of appropriate techniques to either control the
traffic or to control the distribution of traffic across the network
Optimization actions may also involve adding additional links
increasing link capacity, deploying additional hardware such
routers and switches, systematically adjusting parameters
with routing such as IGP metrics and BGP attributes, and

Awduche, et. al. Informational [Page 22]

RFC 3272 Overview and Principles of Internet TE May 2002

traffic management parameters. Network performance optimization
also involve starting a network planning process to improve
network architecture, network design, network capacity,
technology, and the configuration of network elements to
current and future growth

3.1 Components of the Traffic Engineering Process

The key components of the traffic engineering process model include
measurement subsystem, a modeling and analysis subsystem, and
optimization subsystem. The following subsections examine
components as they apply to the traffic engineering process model

3.2

Measurement is crucial to the traffic