U.S. patent application number 13/680840 was filed with the patent office on 2013-03-28 for utilizing optical bypass links in a communication network.
This patent application is currently assigned to ROCKSTAR CONSORTIUM US LP. The applicant listed for this patent is Rockstar Consortium US LP. Invention is credited to Donald Fedyk, Indermohan Monga, Bruce Schofield.
Application Number | 20130077960 13/680840 |
Document ID | / |
Family ID | 41799391 |
Filed Date | 2013-03-28 |
United States Patent
Application |
20130077960 |
Kind Code |
A1 |
Monga; Indermohan ; et
al. |
March 28, 2013 |
UTILIZING OPTICAL BYPASS LINKS IN A COMMUNICATION NETWORK
Abstract
Optical By-Pass (OBP) links may be created by adding wavelengths
between nodes on the network. The OBP may extend between any pair
of nodes on the network. Intermediate nodes on the OBP are
transient nodes and simply forward traffic optically. An OBP
extends between a pair of nodes and, unlike express links, is
created in such a manner that it does not affect the previous
allocation of resources on the network. This enables capacity to be
added between pairs of nodes on the network to alleviate congestion
at a portion of the network, without changing other traffic
patterns on the network. This enables inclusion of an OBP to be
deterministic and of linear impact on the network. The OBP links
may be statically provisioned or created on demand. Optionally, the
OBP links may be crated to coincide with PBB-TE tunnels on the
network.
Inventors: |
Monga; Indermohan; (Acton,
MA) ; Fedyk; Donald; (Groton, MA) ; Schofield;
Bruce; (Tyngsboro, MA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Rockstar Consortium US LP; |
Richardson |
TX |
US |
|
|
Assignee: |
ROCKSTAR CONSORTIUM US LP
Richardson
TX
|
Family ID: |
41799391 |
Appl. No.: |
13/680840 |
Filed: |
November 19, 2012 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12413150 |
Mar 27, 2009 |
8315159 |
|
|
13680840 |
|
|
|
|
61191712 |
Sep 11, 2008 |
|
|
|
Current U.S.
Class: |
398/5 |
Current CPC
Class: |
H04J 14/0246 20130101;
H04J 14/0257 20130101; H04J 14/0267 20130101; H04B 10/038 20130101;
H04J 14/0269 20130101; H04Q 11/0062 20130101; H04Q 2011/0081
20130101; H04Q 2011/0073 20130101; H04J 14/0268 20130101; H04J
14/0284 20130101 |
Class at
Publication: |
398/5 |
International
Class: |
H04B 10/038 20060101
H04B010/038 |
Claims
1. A method of routing traffic in a communication network
comprising a plurality of network nodes interconnected by a
plurality of links, the communication network being configured to
route data over links of the network according to a routing
protocol, the communication network comprising a hop-by-hop path
between a first node and a second node via a plurality of links and
at least one intermediate node, the method comprising: configuring
a bypass link between the first node and the second node, the
bypass link being configured to carry traffic between the first
node and second node without routing the traffic at the at least
one intermediate node; and diverting selected traffic to the bypass
link from the hop-by-hop path between the first node and the second
node in at least one of the first and second nodes responsive to
configuration of the bypass link, without modifying routing of
traffic at network nodes other than the first and second nodes
responsive to configuration of the bypass link.
2. The method of claim 1, wherein network nodes other than the
first and second nodes are not aware of the bypass link between the
first and second nodes.
3. The method of claim 1, wherein network nodes other than the
first and second nodes are not informed of configuration of the
bypass link.
4. The method of claim 1, wherein network nodes other than the
first and second nodes are informed of configuration of the bypass
link, but do not modify their routing of traffic in response to
configuration of the bypass link.
5. The method of claim 4, wherein routing functions at network
nodes other than the first and second nodes associate a cost with
the bypass link that is high enough to prevent network nodes other
than the first and second nodes from adjusting forwarding tables
responsive to configuration of the bypass link.
6. The method of claim 1, wherein the bypass link follows a path
through at least one intermediate node, but traffic is neither
added nor dropped at the at least one intermediate node.
7. The method of claim 1, wherein the bypass link is configured to
switch traffic at the at least one intermediate node without
processing data units of the traffic.
8. The method of claim 1, wherein the bypass link is an optical
bypass link.
9. The method of claim 8, wherein the optical bypass link is
configured to optically switch traffic at the at least one
intermediate node without processing data units of the traffic.
10. The method of claim 1, wherein at least one of the first and
second nodes is configured to aggregate traffic onto the bypass
link to alleviate congestion on the hop-by-hop path between the
first node and the second node.
11. The method of claim 1, wherein the first and second nodes are
configured to divert traffic onto the bypass link by providing
switching functions at the first and second nodes, the switching
functions being provided outside of routing/forwarding functions of
the first and second nodes and being configured to selectively
divert traffic onto the bypass link.
12. The method of claim 11, wherein the switching functions
comprise sorting functions configured to selectively divert
similarly addressed traffic over the bypass link.
13. The method of claim 12, wherein the sorting functions apply
policies local to the first and second nodes.
14. The method of claim 13, wherein the policies are applied on a
per-flow basis.
15. The method of claim 11, wherein the switching functions
comprise a hop-by-hop path queue for data units to be transmitted
over the hop-by-hop path and a bypass link queue for data units to
be transmitted over the bypass link, and the switching functions
are responsive to monitored occupancies of the hop-by-hop path
queue and the bypass link queue to selectively divert traffic to
the bypass link.
16. The method of claim 1, wherein the bypass link is statically
provisioned.
17. The method of claim 1, wherein the bypass link is dynamically
configured based on traffic on the hop-by-hop path between the
first node and the second node.
18. The method of claim 1, wherein the network implements IEEE
802.1Qay and the bypass link corresponds to a Provider Backbone
Bridge Traffic Engineering (PBB-TE) tunnel.
19. The method of claim 18, wherein the bypass link is configured
in response to advertisement of a PBB-TE tunnel when bandwidth
advertised for the PBB-TE tunnel exceeds a predetermined
bandwidth.
20. The method of claim 18, wherein the bypass link is configured
when traffic monitored on the PBB-TE tunnel exceeds a predetermined
amount.
21. The method of claim 1, comprising adjusting a loop prevention
mechanism at the second node to enable diverted traffic received
over the bypass link to be accepted by the second node.
22. The method of claim 1, comprising responding to failure of the
bypass link by pushing traffic onto the hop-by-hop path between the
first node and the second node.
23. The method of claim 1, comprising responding to failure of the
bypass link by configuring a second bypass link, the second bypass
link being diversely routed relative to the failed bypass link.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation of co-pending U.S. patent
application Ser. No. 12/413,150 filed on Mar. 27, 2009, entitled
"UTILIZING OPTICAL BYPASS LINKS IN A COMMUNICATION NETWORK," which
claims priority to U.S. Provisional Patent Application Ser. No.
61/191,712 filed Sep. 11, 2008, entitled "ETHERNET PBB-TE OPTICAL
SHORT-CUTS," which are hereby incorporated herein by reference in
their entirety.
TECHNICAL FIELD
[0002] The present invention relates to communication networks and,
more particularly, to a method and apparatus for utilizing optical
bypass links in a communication network.
BACKGROUND
[0003] Data communication networks may include switches and routers
coupled together to receive and forward data between each other.
These devices will be referred to herein as "network elements." A
network element is generally not a consumer of the data, but rather
is used to receive and forward data so that the data may pass
through the network. Data is communicated through a network by
enabling the network elements to pass protocol data units, such as
frames or packets, between each other over communication links. A
particular protocol data unit may be handled by multiple network
elements and cross multiple communication links as it travels
between its source and its destination over the network.
[0004] Long haul networks typically use optical links to transport
data between the network elements. When a network element receives
optical signals on an optical link, the network element may either
convert the optical signals to electrical signals for processing,
or may use an Optical Add Drop Multiplexers (OADM) to directly
switch the optical signals from one optical fiber to another
optical fiber. Optical cards are generally less expensive than
cards which convert the optical signals to electrical signals, and
thus it is often preferable to handle data optically when possible.
OADMs may be statically configured or may be reconfigurable
(ROADMs), so that the manner in which the node operates may be
remotely changed.
[0005] The demand for optical resources may not be optimally met by
an initial communication network design. Hence, particular links or
areas of the network may experience congestion. Likewise, over
time, congestion may develop as traffic patterns change. To
alleviate congestion, it is common to create an express link to add
a wavelength and, hence bandwidth, between a particular set of
nodes. The express link may go one hop on the network or may be set
up to extend through multiple hops on the network. Where the
express link goes through an intermediate node, the intermediate
node will optically forward the traffic using an OADM and treat the
traffic as transient traffic. Express links are typically manually
configured and provisioned, and then optically signaled to cause
the nodes on the network to add the wavelength(s) for the express
link.
[0006] FIG. 1 shows an example communication network 10. In this
example, traffic that is flowing between nodes E and F will follow
the top path (E, G, H, I, F) while traffic flowing between node A
and F will follow the bottom path (A, B, C, D, F). If an express
link 12 is added, as shown in FIG. 2, the new express link will be
advertised by the network routing system so that all nodes will
update the topology to reflect the new link. As the topology
changes, this will cause the nodes to recalculate the paths through
the network. For example, as shown in FIG. 2, when an express link
is added between nodes A-D, this may cause traffic between nodes E
and F to switch to follow path (E, A, D, F). Thus, when an express
link is added to the network to alleviate congestion, it may in
fact cause additional traffic to be re-routed toward the area of
the network that is already experiencing congestion. One reason for
this is that IP traffic and MPLS traffic will see the express link
as a single hop in the routing tables, which may make the path over
the express link shorter and, hence, preferable to another path
through the network. This may cause some of the traffic to be
diverted to traverse the newly added express link. Accordingly,
rather than helping alleviate congestion, the addition of the
express link may draw additional traffic to a congested area of the
network.
[0007] FIG. 3 shows an example long haul network that may be
implemented over a large geographic area, such as the United
States. In FIG. 3, it will be assumed that the network has a high
volume of traffic between Salt Lake City and St. Louis. To
alleviate this congestion, as shown in FIG. 4, an express link may
be created to carry traffic directly between Salt Lake City and St.
Louis. Once this link is added, the other nodes on the network will
recognize the new link, which will change the other traffic
patterns. For example, inclusion of the new link can cause traffic
on other links to increase dramatically, such as the link between
Chicago and St. Louis, while causing the utilization of other links
to decrease substantially. Indeed, it has been found that adding a
link can affect many links of the network, even those which are
geographically remote from the new express link.
[0008] Accordingly, although adding an express link can alleviate
congestion, it also causes all of the traffic patterns on the
network to change which can result in the creation of new
congestion points. The congestion points may be located near the
new link or very far away from the link on the network.
[0009] Thus, optimization of the network becomes an iterative
process, in which as new links are added, the traffic patterns are
adjusted to reveal new hot-spots, which then must be alleviated
through the addition of other links. This process may be iterated
several times. Additionally, it is possible that the resultant
network design may not be the optimal network design, because each
link is added serially. For example, when a first express link is
added to alleviate one region of congestion, and then a second
express link is added to alleviate a new area of congestion caused
by addition of the first express link, the second express link may
make the first link superfluous. Stated differently it may be that
the first express link is no longer required because addition of
the second express link may cause traffic to be re-routed away from
the area of the first express link to thereby obviate the efficacy
of the first express link. This interdependency of the various
traffic flows and link loading makes addition of network capacity
difficult to implement efficiently and effectively. Accordingly, it
would be advantageous to provide a way to more effectively utilize
express links to alleviate congestion in the network.
SUMMARY OF THE INVENTION
[0010] Optical By-Pass (OBP) links may be created by adding
wavelengths between nodes on the network. The OBP may extend
between any pair of nodes on the network. Intermediate nodes on the
OBP are transient nodes and simply forward traffic optically. An
OBP extends between a pair of nodes and, unlike express links, is
created in such a manner that it does not affect the previous
allocation of resources on the network. This enables capacity to be
added between pairs of nodes on the network to alleviate congestion
at a portion of the network, without changing other traffic
patterns on the network. This enables inclusion of an OBP to be
deterministic and of linear impact on the network. The OBP links
may be statically provisioned or created on demand. Optionally, the
OBP links may be crated to coincide with PBB-TE tunnels on the
network.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] Aspects of the present invention are pointed out with
particularity in the appended claims. The present invention is
illustrated by way of example in the following drawings in which
like references indicate similar elements. The following drawings
disclose various embodiments of the present invention for purposes
of illustration only and are not intended to limit the scope of the
invention. For purposes of clarity, not every component may be
labeled in every figure. In the figures:
[0012] FIG. 1 is a functional block diagram of an example
communication network and showing routes through the network in the
absence of an express link;
[0013] FIG. 2 is a functional block diagram of the example
communication network of FIG. 1 and showing how routes through the
network change with the addition of an express link;
[0014] FIG. 3 is a functional block diagram of an example long haul
communication network showing the traffic loading on the links
before inclusion of an express link;
[0015] FIG. 4 is a functional block diagram of the example long
haul communication network of FIG. 1 showing an hypothetical
example change in traffic patterns caused by inclusion of an
express link;
[0016] FIG. 5 is a functional block diagram of the example long
haul communication network of FIG. 1 after inclusion of an optical
bypass (OBP) link according to an embodiment of the invention;
[0017] FIG. 6 is a functional block diagram of a network element
configured to implement OBP links according to an embodiment of the
invention;
[0018] FIG. 7A is a functional block diagram of an example
communication network showing OPB links; and
[0019] FIG. 7B is a table showing PBB-TE tunnels implemented on the
communication network of FIG. 7A.
DETAILED DESCRIPTION
[0020] FIG. 3 shows an example communication network and example
traffic loading on the links of the communication network. The
number on the links represent a dimensionless traffic volume on the
links on the network. The higher the number the greater amount of
traffic flowing on the links of the network. FIG. 5 shows the same
communication network as FIG. 3, with the same traffic patterns,
but in which an optical bypass (OBP) link has been created between
Salt Lake City and Memphis. In the example shown in FIG. 5, the
loading of the links other than the affected links between Salt
Lake City and Denver, between Denver and St Louis, and between St.
Louis and Memphis, are unchanged. Specifically, when comparing the
network link utilization of the example network of FIG. 3 with the
network link utilization of the example network of FIG. 5, it is
clear that the inclusion of the OBP link did not affect the traffic
patterns on other links on the network. This directly flows from
the fact that the OBP link was created in such a manner as to not
affect resource allocation on the network. For example, in one
embodiment, the OBP link is not advertised via the network routing
system to the other nodes and, hence, the other nodes are not able
to route traffic over the OBP link. Since the other nodes don't
know about the OBP link they cannot route traffic over it. Hence,
the other traffic patterns on the network are unaffected by the
inclusion of the optical bypass link.
[0021] The end nodes, Salt Lake City and Memphis, may place traffic
onto the OBP link or may transmit traffic hop-by-hop through the
network. Selection of which link to utilize is localized to the end
node and is independent of other nodes on the network. The
intermediate nodes, Denver and St. Louis, optically forward traffic
on the OBP link as transient traffic. Thus, the intermediate nodes
do not add or drop traffic from the OBP link but rather simply
optically switch the traffic to cause the traffic on the OBP link
to follow the OBP link through the network. Suitable signaling may
be used to cause the end nodes to switch traffic onto the OBP link
and to enable the intermediate nodes to set up the correct
forwarding state in their data planes, e.g. to configure their
ROADM to optically forward traffic on the OBP link. OBP links may
be set up in the same manner that express links are currently set
up on optical networks, with the exception that they are not
advertised by any of the nodes or are otherwise created so as to
not affect resource allocation, so that use of the OBP link, from a
network traffic perspective, is confined to the end nodes.
[0022] As shown in FIG. 5, the concept of an optical bypass is
created where the traffic path follows the packet layer's
forwarding topology, but avoids packet forwarding in one or more of
the intermediate nodes. In this example, the intermediate nodes do
not perform packet forwarding since they simply optically switch
traffic on the OBP through the network.
[0023] The communication network will typically run a routing
protocol such as Open Shortest Path First (OSPF) or Intermediate
System to Intermediate System (IS-IS). The routing protocol enables
the nodes to learn the topology of the network. The nodes can then
use the knowledge of the topology to calculate routes through the
network. As the topology of the network changes, the nodes will
adjust the routes used to carry traffic through the network to
accommodate the new topology.
[0024] According to an embodiment of the invention, creation of OBP
links does not affect the routing view of the network so that the
routing system has a fixed view of resources on the network that is
not dependent on the OBP links. For example, the OBP links may not
be advertised using routing Link State Advertisements (LSAs) so
that the nodes on the network do not include the OBP links in their
routing topology. Alternatively, the OBP links may be advertised in
such a way that they do not affect the routing system view of the
available network resources. Similarly, the network may maintain
Ethernet adjacencies and paths, and likewise may maintain IP
adjacencies and paths. Creation of OBP links should either not be
advertised or advertised in such a manner that these paths and
adjacencies are not affected by creation of the OBP links.
[0025] Although the inclusion of OBP links does not affect
allocation of network resources, the end devices that connect to
the OBP links are configured to aggregate traffic onto the OBP link
so that the end node can use the OBP links to alleviate congestion
on the network along the path serviced by the OBP links. The end
nodes are able to use the OBP link, because of inclusion of a
switching function which exists outside of the routing/forwarding
function, so that traffic may be selectively diverted onto the OBP
link by the end nodes. This enables the nodes on the ends of the
OBP link to forward traffic to each other over the OBP link to
alleviate congestion on the hop-by-hop links.
[0026] In one embodiment, the OBP links are not advertised via the
routing system so that they are not included in the nodes' network
topology databases. This enables the end nodes to use the OBP links
without having establishment of the OBP links affect other traffic
patterns on the network. In another embodiment, the OBP links are
advertised via the routing system so that the other nodes know
about the existence of the OBP links, but the cost of the OBP links
is set very high so that the nodes will not route traffic on the
OBP links.
[0027] OBP links may be statically provisioned or dynamically
created based on traffic heuristics, explicit requests, or through
other provisioning systems. For example, a network operator may
determine that the network loading between a pair of nodes is too
high, and statically provision an OBP link between the nodes to
alleviate congestion in that area of the network. Alternatively,
the nodes on the network may receive a request to establish a
connection such as a PBB-TE tunnel and determine that optimal
routing of the connection will overload a portion of the network.
Rather than routing the connection along a sub-optimal path through
the network, the nodes may establish an OBP link to correspond to
the connection to alleviate the congestion on the route.
[0028] Once an OBP link has been created, the source node may
aggregate traffic from engineered paths or other flows that are to
be forwarded over a path through the network which would cause the
traffic to pass through the destination that corresponds with the
end point of the OBP link. The decision as to what traffic should
be placed on the OBP link is localized to the node at the end of
the OBP link, since all other nodes along the OBP link will simply
optically forward traffic and won't handle the traffic Likewise,
other nodes on the network will have limited visibility as to the
existence of the OBP link and, hence, won't route traffic over the
OBP link. The end node may place any traffic onto the OBP link that
will ultimately flow through the end point of the OBP link. Traffic
of this nature may be aggregated and placed onto the OBP link for
transportation on the network.
[0029] For example, in a network implementing 802.1Qay, traffic on
a PBB-TE tunnel may be placed on an OBP link corresponding to the
tunnel. By establishing an OBP link to correspond to the PBB-TE
tunnel, and then causing traffic from the PBB-TE tunnel to be
placed the OBP link, a tight coupling between layer 2 and layer 1
may be created on the network. Thus, according to an embodiment,
when a PBB-TE path is advertised, an OBP link may be created to
follow the PBB-TE path where the advertised bandwidth is
sufficiently large to justify establishment of a dedicated optical
path (i.e. one or more dedicated wavelengths) through the
underlying optical network.
[0030] OBP links are created by having dynamic optical switches at
each or some of the routing/Ethernet switching nodes of a network
to either drop/add traffic or to by-pass transient traffic.
Signaling or manual configuration may be used to notify the
transient hops to by-pass the assigned wavelength and to notify the
destination node to drop the wavelength associated with the
OBP.
[0031] The nodes on the network create OBP links where the
aggregate traffic from that node to the other node on the network
justifies creation of the OBP link. The ingress node will aggregate
all traffic destined for a particular node and use the OBP link for
that traffic. Thus, there is a tight coupling between the optical
physical layer (layer 1) and the Ethernet layer (layer 2).
Specifically, a dedicated physical path (layer 1 path) is created
for particular flow of layer 2 traffic. Accordingly, cross-layer
restoration becomes possible in the event of a failure on one of
the layers. For example, by allowing spare capacity to exist in
both layer 1 and layer 2, either layer can recover from
failure/fault in the other layer. Thus, a higher degree of
restoration optimization is possible. So, if an OBP link fails, it
can be either restored by pushing the traffic through the multi-hop
layer 2 network or by creating a diversely routed new OBP link.
Conversely, when a layer 2 fault occurs, it can be restored either
by another layer 2 path or by creating a new OBP link.
[0032] Since the response times for using these different recovery
mechanisms may differ between the layers, the several recovery
mechanisms may be used together to first quickly restore
connectivity and, then later, to obtain a more efficient solution
to the failure. For example, if an OBP link failed, it may be
quicker to route all traffic from the OBP link over the hop-by-hop
path through the network to quickly restore connectivity for the
traffic. A new OBP link may then be determined to replace the
failed OBP link. Once the new OBP link has been signaled on the
network and has been established, the traffic may once again be
directed to the new OBP link rather than the layer 2 network.
[0033] FIG. 7A shows an example network in which OBP links have
been implemented to correspond with PBB-TE tunnels extending
through the network. FIG. 7B shows a table of the OBP links that
may be created to correspond to the PBB-TE tunnels. In the example
shown in FIG. 7A, it will be assumed that there is a PBB-TE tunnel
from Los Angeles to Seattle, for example. When the traffic on the
PBB-TE tunnel exceeds a particular amount, it may be advantageous
to establish an OBP link from Los Angeles to Seattle to carry the
traffic that will be forwarded on the PBB-TE tunnel. Since traffic
on the PBB-TE tunnel will always enter the network at Los Angeles
and exit at Seattle, use of an OBP link to enable the intermediate
nodes to directly switch traffic on the link is more efficient and
less costly than requiring the intermediate nodes to inspect and
route the traffic at each hop. By not advertising the extra
capacity on the OBP path, the other network elements will not
divert traffic to that area of the network so that addition of the
OBP link to the network will not affect other traffic patterns on
the network. This allows selective placement of OBP links to be
used to alleviate congestion in the network without modifying
traffic patterns on the network as a whole.
[0034] Implementing OBP links that coincide with PBB-TE tunnels
also enables monitoring functions to follow the decisions of the
data through the network. Specifically, the OBP link may be
monitored to determine the presence of a fault on the OBP link.
Since the data on the OBP link will flow through the same set of
nodes as the PBB-TE tunnel, the presence of a fault on the OBP link
will typically coincide with a fault on the PBB-TE tunnel. Thus,
unlike MPLS or IP traffic that may take different paths through the
network, traffic on the PBB-TE tunnel and an OBP link that follows
the PBB-TE tunnel may be expected to have similar fault
characteristics. This simplifies handling of fault and re-routing
of traffic. In a normal IP or MPLS context, by contrast, the IP
traffic may experience a fault on the network whereas the traffic
on the OBP link would not. Thus, monitoring functions associated
with the OBP link are simplified in the PBB-TE context where the
OBP link are established to coincide with the PBB-TE tunnels.
[0035] FIG. 6 shows an example network element 60 that may be used
to implement an endpoint of an OBP link according to an embodiment
of the invention. As shown in FIG. 6, the network element includes
an encapsulation function 62 to encapsulate IP/Ethernet customer
traffic 64. Frequently, traffic received from a customer will be
encapsulated for transportation on the provider network. Provider
Backbone Bridging specified in IEEE 802.1ah, enables MAC-in-MAC
encapsulation to take place so that Ethernet frames may be
forwarded based on provider MAC address space (B-MAC) rather than
customer MAC address (C-MAC) space. Provider Backbone
Bridging--Traffic Engineering (PBB-TE or PBT) specified by 802.1Qay
provides a similar encapsulation function, but enables forwarding
on the network to be based on the provider destination MAC address
(B-DA) and VLAN ID (B-VID). This enables traffic engineered tunnels
to be created on the network. The encapsulation function 62 may
perform either of these types of encapsulation processes or,
alternatively, may implement another form of encapsulation.
Encapsulated traffic is sent to a forwarding function 66.
[0036] Customer traffic is only encapsulated at the ingress to the
provider network. Once encapsulated, the network elements on the
network will use the outer MAC header to forward the traffic on the
network. Thus, the network element 60 is not be required to
encapsulate all traffic but rather only performs this function on
customer traffic 64. Where the network element receives traffic 68
from another network element on the network, that traffic 68 will
have already been encapsulated by another network element at the
ingress to the network. Traffic of this nature may be passed
directly to the forwarding function 66 without being passed through
the encapsulation function.
[0037] The forwarding function 66, according to an embodiment of
the invention, operates under the control of a sorting function 70,
to either forward traffic over the hop-by-hop links 72 or over
Optical Bypass (OBP) links 74A and 74B. A queuing function 76 (or
individual queuing functions 76A, 76B, 76C) holds the traffic from
the forwarding function to enable the traffic to be placed onto the
appropriate link. Typically, a separate queuing function would be
implemented for each of the links, although the particular way in
which the traffic is handled and processed for transmission on the
intended link is dependent on the particular implementation.
[0038] Normally, a network element will have a forwarding function,
i.e. a forwarding information base, that is programmed into
hardware on the network element. The forwarding hardware is
commonly referred to as the network element data plane. A control
plane of the network element implements a routing function 81 that
determines how traffic is to be forwarded on the network. For
example, the routing function will receive routing advertisements
from other nodes on the network, generate a network topology
database, calculate routes through the network, and otherwise
determine how traffic is to be forwarded on the network. The
routing function will program this information into the data plane
to specify how traffic is to be forwarded using the hop-by-hop
links on the network.
[0039] According to an embodiment of the invention, the sorting
function 70 enables the data plane to selectively output similarly
addressed traffic over an OBP link or the hop-by-hop links without
requiring knowledge of the OBP links to affect allocation of
resources by the routing function 81. In this manner, the sorting
function can utilize the OBP links, so that additional capacity can
be added to the network without changing the visible topology of
the network from a routing perspective.
[0040] The sorting function 70 operates according to policies 78
which enable particular traffic to be selectively directed over the
hop-by-hop links 72 or over the OBP links 74. The policies are
local to the node and enable the node to decide how flows of
traffic should be sorted. The policies may be, for example, that
all traffic on a particular PBB-TE path should be directed to a
particular one of the OBP links 74. Preferably the policies 78 are
implemented on a per-flow basis rather than on a per-packet/frame
basis, so that packets/frames that belong to a particular flow
either all will be directed to the hop-by-hop links or all will be
directed to the OBP links. This enables packets/frames within the
flow to remain in order as they traverse the network.
[0041] A monitoring function 80 may monitor the queuing function to
keep track of the amount of traffic queued for the bypass links 74
and hop-by-hop links 72. The monitoring function 80 will typically
be a control plane function and enable the control plane to monitor
how the data plane is behaving. When the monitoring function
determines that the queuing function 76 has excess data queued at
one of the links, it may modify the policies 78 used by the sorting
function to divert traffic to the other type of link. Thus, for
example, if the queuing function 76C serving the hop-by-hop links
72 has excess traffic stored for transmission on the hop-by-hop
links, the monitoring function 80 may modify the policies 78 to
cause additional flows of traffic to be moved from the hop-by-hop
links to one of the OBP links 74. This may help to alleviate
congestion on the hop-by-hop links 72. Similarly, if there is
excess traffic on one of the OBP links 74, the monitoring function
80 will detect the excess traffic built in the queuing function and
modify one of the policies 78 to shift some of the traffic from the
OBP link 74 to another OBP link or to the hop-by-hop link 72.
[0042] Referring FIG. 7, assume that the network element 60 has
been implemented in Los Angeles. In this example, it will be
assumed that the bypass link 74A extends from Los Angeles through
San Francisco and Portland Oreg. to Seattle, that bypass link 74B
extends from Los Angeles to San Francisco, and that the hop-by-hop
links also extend from Los Angeles to San Francisco. The network
element 60, in this instance, will have three options to forward
traffic to Seattle. First, it can use bypass link 74A to forward
traffic on an OBP directly to Seattle. Second, the network element
60 may use bypass link 74B to forward the traffic to San Francisco,
at which point the traffic may be forwarded hop-by-hop the rest of
the way to Seattle. Third, the network element can use normal
hop-by-hop forwarding to forward the traffic to Seattle. In this
example it has been assumed that the hop-by-hop forwarding will
cause traffic to flow north through San Francisco. The hop-by-hop
forwarding may cause the traffic to take a different route,
however, such as through Phoenix, Denver, and Great Falls Mont.,
before getting to Seattle, without affecting the operation of this
example.
[0043] The sorting function 70 will have policies 78 specifying
which flows of traffic should be forwarded over links 74A, 74B, and
72. For example, the default policy may be that all flows which are
destined for Seattle should be aggregated and forwarded over the
bypass link 74A, to enable that traffic to be optically switched
en-route to Seattle. This keeps traffic off the other hop-by-hop
links by eliminating the need for the traffic to be processed by
each intermediate node. If, however, the control plane monitoring
function detects that the queue 76A is too long, the monitoring
function may look at the queue 76B for the OBP link 74B to
determine whether there is room on the OBP link 74B for the
traffic. If so, some of the traffic may be shifted from OBP link
74A to OBP link 74B. Alternatively, the traffic may be shifted to
the hop-by-hop links 72 to enable the traffic to be forwarded in a
normal manner on the network.
[0044] When the network element receives traffic, it will process
the traffic in a normal manner regardless of whether the traffic
was received on a bypass link or on a hop-by-hop link.
Specifically, the traffic that is placed onto the OBP link is not
modified in any way and, thus, when the traffic is received by a
network element off an OBP link, the network element may process
the traffic as if it had been received off a normal hop-by-hop
link. This stems from the fact that the traffic is directed, at the
ingress to the OBP link, by a sorting function which does not
require the frame headers or IP packet headers to be modified.
Thus, the traffic on the OBP link will appear, to the egress node,
the same as if it had been forwarded to the egress node over the
hop-by-hop links.
[0045] The one modification that may be required by the egress
node, depending on the implementation, is in connection with loop
suppression. In some implementations the nodes on the network may
be required to implement a loop prevention mechanism such as
Reverse Path Forwarding Check (RPFC). RPFC is a check that may be
performed by a network element to ensure that traffic has arrived
at the node from the correct port. If traffic from a particular
node arrives at an incorrect port, the network element may assume
that a loop has occurred and drop the traffic. Since the traffic on
an OBP link may arrive at a different port than it would if it had
followed normal hop-by-hop forwarding, the loop prevention
mechanism may need to be adjusted to enable the traffic that is
received over the bypass link to be accepted by the egress
node.
[0046] The functions described above may be implemented as a set of
program instructions that are stored in a computer readable memory
and executed on one or more processors on the computer platform.
However, it will be apparent to a skilled artisan that all logic
described herein can be embodied using discrete components,
integrated circuitry such as an Application Specific Integrated
Circuit (ASIC), programmable logic used in conjunction with a
programmable logic device such as a Field Programmable Gate Array
(FPGA) or microprocessor, a state machine, or any other device
including any combination thereof. Programmable logic can be fixed
temporarily or permanently in a tangible medium such as a read-only
memory chip, a computer memory, a disk, or other storage medium.
All such embodiments are intended to fall within the scope of the
present invention.
[0047] It should be understood that various changes and
modifications of the embodiments shown in the drawings and
described in the specification may be made within the spirit and
scope of the present invention. Accordingly, it is intended that
all matter contained in the above description and shown in the
accompanying drawings be interpreted in an illustrative and not in
a limiting sense. The invention is limited only as defined in the
following claims and the equivalents thereto.
* * * * *