Request for Comments: 4397 Independent Consultant
Category: Informational A. Farrel
Old Dog Consulting
February 2006
A Lexicography for the Interpretation of Generalized Multiprotocol
Label Switching (GMPLS) Terminology within the Context of the
ITU-T’s Automatically Switched Optical Network (ASON) Architecture
Status of This Memo
This memo provides information for the Internet community. It does
not specify an Internet standard of any kind. Distribution of this
memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (2006).
Abstract
Generalized Multiprotocol Label Switching (GMPLS) has been developed
by the IETF to facilitate the establishment of Label Switched Paths
(LSPs) in a variety of data plane technologies and across several
architectural models. The ITU-T has specified an architecture for
the control of Automatically Switched Optical Networks (ASON).
This document provides a lexicography for the interpretation of GMPLS
terminology within the context of the ASON architecture.
It is important to note that GMPLS is applicable in a wider set of
contexts than just ASON. The definitions presented in this document
do not provide exclusive or complete interpretations of GMPLS
concepts. This document simply allows the GMPLS terms to be applied
within the ASON context.
Table of Contents
1. Introduction ....................................................3
2. Terminology .....................................................3
2.1. GMPLS Terminology Sources ..................................3
2.2. ASON Terminology Sources ...................................4
2.3. Common Terminology Sources .................................4
3. Lexicography ....................................................4
3.1. Network Presences ..........................................4
3.2. Resources ..................................................5
3.3. Layers .....................................................6
3.4. Labels .....................................................7
3.5. Data Links .................................................7
3.6. Link Interfaces ............................................8
3.7. Connections ................................................9
3.8. Switching, Termination, and Adaptation Capabilities .......10
3.9. TE Links and FAs ..........................................11
3.10. TE Domains ...............................................13
3.11. Component Links and Bundles ..............................13
3.12. Regions ..................................................14
4. Guidance on the Application of this Lexicography ...............14
5. Management Considerations ......................................15
6. Security Considerations ........................................15
7. Acknowledgements ...............................................15
8. Normative References ...........................................16
9. Informative References .........................................16
1. Introduction
Generalized Multiprotocol Label Switching (GMPLS) has been developed
by the IETF to facilitate the establishment of Label Switched Paths
(LSPs) in a variety of data plane technologies such as Packet
Switching Capable (PSC), Layer Two Switching Capable (L2SC), Time
Division Multiplexing (TDM), Lambda Switching Capable (LSC), and
Fiber Switching Capable (FSC).
The ITU-T has specified an architecture for the control of
Automatically Switched Optical Networks (ASON). This architecture
forms the basis of many Recommendations within the ITU-T.
Because the GMPLS and ASON architectures were developed by different
people in different standards bodies, and because the architectures
have very different historic backgrounds (the Internet, and transport
networks respectively), the terminology used is different.
This document provides a lexicography for the interpretation of GMPLS
terminology within the context of the ASON architecture. This allows
GMPLS documents to be generally understood by those familiar with
ASON Recommendations. The definitions presented in this document do
not provide exclusive or complete interpretations of the GMPLS
concepts.
2. Terminology
Throughout this document, angle brackets ("<" and ">") are used to
indicate the context in which a term applies. For example, "<Data
Plane>" as part of a description of a term means that the term
applies within the data plane.
2.1. GMPLS Terminology Sources
GMPLS terminology is principally defined in [RFC3945]. Other
documents provide further key definitions including [RFC4201],
[RFC4202], [RFC4204], and [RFC4206].
The reader is recommended to become familiar with these other
documents before attempting to use this document to provide a more
general mapping between GMPLS and ASON.
For details of GMPLS signaling, please refer to [RFC3471] and
[RFC3473]. For details of GMPLS routing, please refer to [RFC4203]
and [RFC4205].
2.2. ASON Terminology Sources
The ASON architecture is specified in ITU-T Recommendation G.8080
[G-8080]. This is developed from generic functional architectures
and requirements specified in [G-805], [G-807], and [G-872]. The
ASON terminology is defined in several Recommendations in the ASON
family such as [G-8080], [G-8081], [G-7713], [G-7714], and [G-7715].
The reader must be familiar with these documents before attempting to
apply the lexicography set out in this document.
2.3. Common Terminology Sources
The work in this document builds on the shared view of ASON
requirements and requirements expressed in [RFC4139], [RFC4258], and
[RFC4394].
3. Lexicography
3.1. Network Presences
3.1.1. GMPLS Terms
Transport node <Data Plane> is a logical network device that is
capable of originating and/or terminating of a data flow and/or
switching it on the route to its destination.
Controller <Control Plane> is a logical entity that models all
control plane intelligence (routing, traffic engineering (TE), and
signaling protocols, path computation, etc.). A single controller
can manage one or more transport nodes. Separate functions (such
as routing and signaling) may be hosted at distinct sites and
hence could be considered as separate logical entities referred
to, for example, as the routing controller, the signaling
controller, etc.
Label Switching Router (LSR) <Control & Data Planes> is a logical
combination of a transport node and the controller that manages
the transport node. Many implementations of LSRs collocate all
control plane and data plane functions associated with a transport
node within a single physical presence making the term LSR
concrete rather than logical.
In some instances, the term LSR may be applied more loosely to
indicate just the transport node or just the controller function
dependent on the context.
Node <Control & Data Planes> is a synonym for an LSR.
Control plane network <Control Plane> is an IP network used for
delivery of control plane (protocol) messages exchanged by
controllers.
3.1.2. ASON Terms
A GMPLS transport node is an ASON network element.
A GMPLS controller is the set of ASON functional components
controlling a given ASON network element (or partition of a network
element). In ASON, this set of functional components may exist in
one place or multiple places.
A GMPLS node is the combination of an ASON network element (or
partition of a network element) and its associated control
components.
The GMPLS control plane network is the ASON Signaling Communication
Network (SCN). Note that both routing and signaling exchanges are
carried by the SCN.
3.2. Resources
3.2.1. GMPLS Terms
Non-packet-based resource <Data Plane> is a channel of a certain
bandwidth that could be allocated in a network data plane of a
particular technology for the purpose of user traffic delivery.
Examples of non-packet-based resources are timeslots, lambda
channels, etc.
Packet-based resource <Data Plane> is an abstraction hiding the means
related to the delivery of traffic with particular parameters
(most importantly, bandwidth) with particular quality of service
(QoS) over PSC media. Examples of packet-based resources are
forwarding queues, schedulers, etc.
Layer Resource (Resource) <Data Plane>. A non-packet-based data
plane technology may yield resources in different network layers
(see section 3.3). For example, some TDM devices can operate with
VC-12 timeslots, some with VC-4 timeslots, and some with VC4-4c
timeslots. There are also multiple layers of packet-based
resources (i.e., one per label in the label stack). Therefore, we
define layer resource (or simply resource) irrespective of the
underlying data plane technology as a basic data plane construct.
It is defined by a combination of a particular data encoding type
and a switching/terminating bandwidth granularity. Examples of
layer resources are: PSC1, PSC4, ATM VP, ATM VC, Ethernet, VC-12,
VC-4, Lambda 10G, and Lambda 40G.
These three definitions give rise to the concept of Resource Type.
Although not a formal term, this is useful shorthand to identify how
and where a resource can be used dependent on the switching type,
data encoding type, and switching/terminating bandwidth granularity
(see section 3.8).
All other descriptions provided in this memo are tightly bound to the
resource.
3.2.2. ASON Terms
ASON terms for resource:
- In the context of link discovery and resource management
(allocation, binding into cross-connects, etc.), a GMPLS resource
is one end of a link connection.
- In the context of routing, path computation, and signaling, a GMPLS
resource is a link connection or trail termination.
Resource type is identified by a client CI (Characteristics
Information) that could be carried by the resource.
3.3. Layers
3.3.1. GMPLS Terms
Layer <Data Plane> is a set of resources of the same type that could
be used for establishing a connection or used for connectionless
data delivery.
Note. In GMPLS, the existence of non-blocking switching function in
a transport node in a particular layer is modeled explicitly as one
of the functions of the link interfaces connecting the transport node
to its data links.
A GMPLS layer is not the same as a GMPLS region. See section 3.12.
3.3.2. ASON Terms
A GMPLS layer is an ASON layer network.
3.4. Labels
3.4.1. GMPLS Terms
Label <Control Plane> is an abstraction that provides an identifier
for use in the control plane in order to identify a transport
plane resource.
3.4.2. ASON Terms
A GMPLS label is the portion of an ASON SNP name that follows the
SNPP name.
3.5. Data Links
3.5.1. GMPLS Terms
Unidirectional data link end <Data Plane> is a set of resources that
belong to the same layer and that could be allocated for the
transfer of traffic in that layer from a particular transport node
to the same neighboring transport node in the same direction. A
unidirectional data link end is connected to a transport node by
one or more link interfaces (see section 3.6).
Bidirectional data link end <Data Plane> is an association of two
unidirectional data link ends that exist in the same layer and
that could be used for the transfer of traffic in that layer
between a particular transport node and the same neighbor in both
directions. A bidirectional data link end is connected to a
transport node by one or more link interfaces (see section 3.6).
Unidirectional data link <Data Plane> is an association of two
unidirectional data link ends that exist in the same layer, that
are connected to two transport nodes adjacent in that layer, and
that could be used for the transfer of traffic between the two
transport nodes in one direction.
Bidirectional data link <Data Plane> is an association of two
bidirectional data link ends that exist in the same layer, that
are connected to two transport nodes adjacent in that layer, and
that could be used for the transfer of traffic between the two
transport nodes in both directions.
3.5.2. ASON Terms
A GMPLS unidirectional data link end is a collection of connection
points from the same client layer that are supported by a single
trail termination (access point).
A GMPLS data link is an ASON link supported by a single server trail.
3.6. Link Interfaces
3.6.1. GMPLS Terms
Unidirectional link interface <Data Plane> is an abstraction that
connects a transport node to a unidirectional data link end and
represents (hides) the data plane intelligence like switching,
termination, and adaptation in one direction. In GMPLS, link
interfaces are often referred to as "GMPLS interfaces" and it
should be understood that these are data plane interfaces and the
term does not refer to the ability of a control plane interface to
handle GMPLS protocols.
A single unidirectional data link end could be connected to a
transport node by multiple link interfaces with one of them, for
example, realizing switching function, while others realize the
function of termination/adaptation.
Bidirectional link interface <Data Plane> is an association of two or
more unidirectional link interfaces that connects a transport node
to a bidirectional data link end and represents the data plane
intelligence like switching, termination, and adaptation in both
directions.
Link interface type <Data Plane> is identified by the function the
interface provides. There are three distinct functions --
switching, termination, and adaptation; hence, there are three
types of link interface. Thus, when a Wavelength Division
Multiplexing (WDM) link can do switching for some lambda channels,
and termination and TDM OC48 adaptation for some other lambda
channels, we say that the link is connected to the transport node
by three interfaces each of a separate type: switching,
termination, and adaptation.
3.6.2. ASON Terms
A GMPLS interface is the set of trail termination and adaptation
functions between one or more server layer trails and a specific
client layer subnetwork (which commonly is a matrix in a network
element).
The GMPLS interface type may be identified by the ASON adapted client
layer, or by the terminated server layer, or a combination of the
two, depending on the context. In some cases, a GMPLS interface
comprises a set of ASON trail termination/adaptation functions, for
which some connection points are bound to trail terminations and
others to matrices.
3.7. Connections
3.7.1. GMPLS Terms
In GMPLS a connection is known as a Label Switched Path (LSP).
Unidirectional LSP (connection) <Data Plane> is a single resource or
a set of cross-connected resources of a particular layer that
could deliver traffic in that layer between a pair of transport
nodes in one direction.
Unidirectional LSP (connection) <Control Plane> is the signaling
state necessary to maintain a unidirectional data plane LSP.
Bidirectional LSP (connection) <Data Plane> is an association of two
unidirectional LSPs (connections) that could simultaneously
deliver traffic in a particular layer between a pair of transport
nodes in opposite directions.
In the context of GMPLS, both unidirectional constituents of a
bidirectional LSP (connection) take identical paths in terms of
data links, are provisioned concurrently, and require a single
(shared) control state.
Bidirectional LSP (connection) <Control Plane> is the signaling state
necessary to maintain a bidirectional data plane LSP.
LSP (connection) segment <Data Plane> is a single resource or a set
of cross-connected resources that constitutes a segment of an LSP
(connection).
3.7.2. ASON Terms
A GMPLS LSP (connection) is an ASON network connection.
A GMPLS LSP segment is an ASON serial compound link connection.
3.8. Switching, Termination, and Adaptation Capabilities
3.8.1. GMPLS Terms
Switching capability <Data Plane> is a property (and defines a type)
of a link interface that connects a particular data link to a
transport node. This property/type characterizes the interface’s
ability to cooperate with other link interfaces connecting data
links within the same layer to the same transport node for the
purpose of binding resources into cross-connects. Switching
capability is advertised as an attribute of the TE link local end
associated with the link interface.
Termination capability <Data Plane> is a property of a link interface
that connects a particular data link to a transport node. This
property characterizes the interface’s ability to terminate
connections within the layer that the data link belongs to.
Adaptation capability <Data Plane> is a property of a link interface
that connects a particular data link to a transport node. This
property characterizes the interface’s ability to perform a
nesting function -- to use a locally terminated connection that
belongs to one layer as a data link for some other layer.
The need for advertisement of adaptation and termination capabilities
within GMPLS has been recognized, and work is in progress to
determine how these will be advertised. It is likely that they will
be advertised as a single combined attribute, or as separate
attributes of the TE link local end associated with the link
interface.
3.8.2. ASON Terms
In ASON applications:
The GMPLS switching capability is a property of an ASON link end
representing its association with a matrix.
The GMPLS termination capability is a property of an ASON link end
representing potential binding to a termination point.
The GMPLS adaptation capability is a property of an ASON link end
representing potential adaptation to/from a client layer network.
3.9. TE Links and FAs
3.9.1. GMPLS Terms
TE link end <Control Plane> is a grouping for the purpose of
advertising and routing of resources of a particular layer.
Such a grouping allows for decoupling of path selection from
resource assignment. Specifically, a path could be selected in a
centralized way in terms of TE link ends, while the resource
assignment (resource reservation and label allocation) could be
performed in a distributed way during the connection setup. A TE
link end may reflect zero, one or more data link ends in the data
plane. A TE link end is associated with exactly one layer.
TE link <Control Plane> is a grouping of two TE link ends associated
with two neighboring transport nodes in a particular layer.
In contrast to a data link, which provides network flexibility in
a particular layer and, therefore, is a "real" topological
element, a TE link is a logical routing element. For example, an
LSP path is computed in terms of TE links (or more precisely, in
terms of TE link ends), while the LSP is provisioned over (that
is, resources are allocated from) data links.
Virtual TE link is a TE link associated with zero data links.
TE link end advertising <Control Plane>. A controller managing a
particular transport node advertises local TE link ends. Any
controller in the TE domain makes a TE link available for its
local path computation if it receives consistent advertisements of
both TE link ends. Strictly speaking, there is no such thing as
TE link advertising -- only TE link end advertising. TE link end
advertising may contain information about multiple switching
capabilities. This, however, should not be interpreted as
advertising of a multi-layer TE link end, but rather as joint
advertisement of ends of multiple parallel TE links, each
representing resources in a separate layer. The advertisement may
contain attributes shared by all TE links in the group (for
example, protection capabilities, Shared Risk Link Groups (SRLGs),
etc.), separate information related to each TE link (for example,
switching capability, data encoding, unreserved bandwidth, etc.)
as well as information related to inter-layer relationships of the
advertised resources (for example, termination and adaptation
capabilities) should the control plane decide to use them as the
termination points of higher-layer data links. These higher-layer
data links, however, are not real yet -- they are abstract until
the underlying connections are established in the lower layers.
LSPs created in lower layers for the purpose of providing data
links (extra network flexibility) in higher layers are called
hierarchical connections or LSPs (H-LSPs), or simply hierarchies.
LSPs created for the purpose of providing data links in the same
layer are called stitching segments. H-LSPs and stitching
segments could, but do not have to, be advertised as TE links.
Naturally, if they are advertised as TE links (LSPs advertised as
TE links are often referred to as TE-LSPs), they are made
available for path computations performed on any controller within
the TE domain into which they are advertised. H-LSPs and
stitching segments could be advertised either individually or in
TE bundles. An H-LSP or a stitching segment could be advertised
as a TE link either into the same or a separate TE domain compared
to the one within which it was provisioned.
A set of H-LSPs that is created (or could be created) in a
particular layer to provide network flexibility (data links) in
other layers is called a Virtual Network Topology (VNT). A single
H-LSP could provide several (more than one) data links (each in a
different layer).
Forwarding Adjacency (FA) <Control Plane> is a TE link that does not
require a direct routing adjacency (peering) between the
controllers managing its ends in order to guarantee control plane
connectivity (a control channel) between the controllers. An
example of an FA is an H-LSP or stitching segment advertised as a
TE link into the same TE domain within which it was dynamically
provisioned. In such cases, the control plane connectivity
between the controllers at the ends of the H-LSP/stitching segment
is guaranteed by the concatenation of control channels
interconnecting the ends of each of its constituents. In
contrast, an H-LSP or stitching segment advertised as a TE link
into a TE domain (different than one where it was provisioned)
generally requires a direct routing adjacency to be established
within the TE domain where the TE link is advertised in order to
guarantee control plane connectivity between the TE link ends.
Therefore, is not an FA.
3.9.2. ASON Terms
The ITU term for a TE link end is Subnetwork Point (SNP) pool (SNPP).
The ITU term for a TE link is SNPP link.