NOVATEUR PUBLICATIONS

International Journal Of Research Publications In Engineering And Technology [IJRPET]

ISSN: 2454-7875
VOLUME 2, ISSUE 12, Dec. -2016
ANALYSIS OF MECHANISMS FOR TOLERATING MULTIPLE LINK
FAILURES IN MPLS NETWORK
PROF. DIPTI SONAWANE
G. H. Raisoni College of Engineering, North Maharashtra University, Jalgaon, 425001, India
[email protected]
MR. MANGESH T. CHAUDHARI
G. H. Raisoni College of Engineering, North Maharashtra University, Jalgaon, 425001, India
[email protected]

ABSTRACT: The MPLS domain can be divided into MPLS

Multiprotocol Label Switching (MPLS) is core and MPLS edge. The MPLS core consists of nodes or
switching network and provides significant benefits router that can be capable of forwarding IP packets and
by fast forwarding packets. MPLS is scalable network attach label to it within a MPLS network, while the edge
and it is useful for end-to-end quality of service consists of nodes neighboring both MPLS capable and
(QoS), it also enabling efficient utilization of existing incapable nodes for IP packet forwarding. The nodes in
network resources. In MPLS, there is no admission the MPLS domain are called as LSRs (Label Switch
control for nodes and it is connection-oriented Routers). The nodes in the core are called transit LSRs
network which makes network more reliable. For and the nodes in the MPLS edge are called LERs (Label
MPLS network, failure can be occur at any point of Edge Routers). If a LER is the first node in the path for a
time if the network link is overloading with traffic or packet travelling through the MPLS domain this node is
node leave network. If the link failure occur in the called the ingress LER, if it is the last node in a path it's
MPLS network then there is need to establish a new called the egress LER. This depends on the direction of
label switched path (LSP) and then forward the traffic flow in the network, one node can therefore be
packets to the newly established LSP. The both ingress and egress LER depending on which flow is
forwarding of failed link traffic to different or considered in the network. The terms upstream and
backup path this may leads LSP get more congested. downstream routers are also used to indicate in which
Here some mechanisms used for to tolerate these order the routers are forwarding the traffic flow. If a LSR
link failures in MPLS network. The main focus to is upstream from another LSR, traffic is passed through
analyze the various mechanisms used for tolerates that LSR before the other (downstream). A schematic
the link failure in MPLS based on the Quality of view of the MPLS domain is shown as follows.
Service (QoS) parameters. The expected result from
this thesis, the network should maintain
connectivity after multiple failures without causing
congestion.
KEYWORDS: MPLS, LSP, QoS, etc.

INTRODUCTION:
Multiprotocol Label Switching (MPLS) is an
improved method for forwarding Internet Protocol (IP)
packets through a network using information contained
in labels. The labels are inserted between the Layer 3
(network) header and the Layer 2 (data link layer) Figure 1: MPLS Architecture
header, so it is also called 2.5 layer networks. Nowadays
IP based networks uses MPLS as backbone network for FORWARDING EQUIVALENCE CLASS (FEC):
fast forwarding and switching of IP packets. Also Frame In MPLS network, all IP packets that are
Relay (FR) and Asynchronous Transfer Mode (ATM) forwarded over the same path and treated in the same
networks have many disadvantages in the management manner belong to the same FEC. The traffic flows that
operation of large networks such as cost, security, are aggregated in MPLS are called an FEC. There should
scalability and flexibility; this can be overcome in MPLS be a FEC to assign any unlabeled incoming packet into a
network. group that will become MPLS labeled packets. MPLS FEC
membership is not strictly based on shortest path first

95 | P a g e
 NOVATEUR PUBLICATIONS
International Journal Of Research Publications In Engineering And Technology [IJRPET]
ISSN: 2454-7875
VOLUME 2, ISSUE 12, Dec. -2016
(SPF) destination address calculations as in IP, but can The label is 20 bit, the 3 bits for experimental
be determined based on other parameters such as which defines the class of service and Explicit Congestion
packet source, DiffServ code points (DSCP) and some Notification (ECN) bits for alert when there is congestion
QoS parameters found in the network, transport and in the MPLS network then this bit is set otherwise the bit
application headers. If the classification is based just on is not set. Third field for label stack bit if it set then there
the destination IP address, then the resulting FECs are of is label in the label stack. The last field is Time to Live
medium-granularity. If the FEC classification is based (TTL) which indicates the total time taken by an IP
solely on the egress LSR, this creates coarse-granularity packet to travel in the MPLS network.
FECs.
MPLS SIGNALING PROTOCOLS:
LABEL SWITCHED PATH (LSP): Signaling is a way in which routers exchange
When an IP packet traverses through a MPLS relevant information. Four methods have been specified
domain, it follows a predetermined path depending on for label distribution.
the FEC to which it was assigned by the ingress LER. The A. Label Distribution Protocol (LDP)
path the packet follows through the MPLS domain is B. Constrained Routing with LDP (CR-LDP)
called the Label Switched Path (LSP). LSPs are C. Resource Reservation Protocol extension for MPLS
unidirectional so to build a duplex communication two (RSVP-TE)
LSPs are needed. When various Layer 3 packets are D. Distributing labels with Border Gateway Protocol
entering the Ingress LSR, they are classified into a FEC. (BGP)
Once the packets are classified, they are forward to
respective LSP for this FEC. An LSP may carry more than A. LABEL DISTRIBUTION PROTOCOL (LDP):
one FEC. LDP is designed for the purpose of distributing
The packet is forwarded based on the MPLS labels, LDP works like "hop-by-hop" forwarding. It
information in the MPLS header and the interface that always selects the same physical path that conventional
the packet arrived on, which is used as an index in table IP routing would select. Thus LDP does not support
lookups. There are three basic types of operations that Traffic Engineering (TE). The motivation behind setting
can be applied to IP packet such as Push the label stack, up an LSP that follows the same path as conventional IP
Swap the top label with a new label and Pop the label instead of just using conventional IP routing was
stack. originally to speed up the forwarding in routers. In
conventional IP routing the next hop for each packet is
MPLS HEADER: found by a longest match prefix lookup on the IP header
From this MPLS architecture, the forwarding in in the routing table. These lookup could in some cases
MPLS is done by using the label in the MPLS header. where the routing tables were large be time consuming
Therefore the MPLS header has to be inserted into and it was thought that data forwarding with label
packages that are to be routed in the MPLS domain. For switching instead of IP lookups would speed up data
data link layer switching technologies like ATM and FR, forwarding. Because of the recent development in
the MPLS header is inserted in the native label field for routing technology, LDP is not much used for label
that protocol. In the case where the Layer 2 technology distribution today. There is however an extension to the
does not support a native label field, the MPLS header original LDP protocol that brings new functionality for
must be inserted between the Layer 2 and Layer 3 the LDP protocol called CR-LDP.
headers. This MPLS header is 32 bits long and is often
called the "shim" header. The MPLS header contains four B. CONSTRAINED ROUTING WITH LDP (CR-LDP):
fields such as. CR-LDP is an extension of LDP to support
constraint based routed LSPs. The term constraint
implies that for each set of nodes there exists a set of
constraint that must be satisfied for the link or links
between two nodes to be chosen for an LSP. An example
of a constraint is to find a path that needs a specific
amount of bandwidth. LSRs that use CR-LDP to exchange
label and FEC mapping information are called LDP peers;
they exchange this information by forming a LDP
session. There are four categories of LDP messages:
Figure 2: MPLS Header

96 | P a g e
 NOVATEUR PUBLICATIONS
International Journal Of Research Publications In Engineering And Technology [IJRPET]
ISSN: 2454-7875
VOLUME 2, ISSUE 12, Dec. -2016
1. Discovery messages announce and maintain the 1. FIXED FILTER (FF): In fixed filter a distinct
presence of an LSR in an MPLS domain. This message is reservation is made for traffic from each sender. This
periodically sent as a Hello message through a UDP port reservation cannot be shared by other senders.
with the multicast address of all routers on this subnet.
2. Session message is sent to establish, maintain and 2. SHARED EXPLICIT (SE): Allows a receiver to
delete sessions between LDP peers. explicitly specify the senders to be included in a
3. Advertisement messages create, change and delete reservation. There is a single reservation on a link for all
label mappings for FECs. the senders listed.
4. Notification Messages provides status, diagnostic and
error information. 3. WILDCARD FILTER (WF): With the Wildcard Filter
The last three message types are transported over TCP. (WF) reservation style, a single shared reservation is
CR-LDP makes hard state reservations which means that used for all senders to a session. The total reservation on
reserved resources has to be removed explicitly. a link remains the same regardless of the number of
senders. A single multipoint-to-point label switched path
C. RESOURCE RESERVATION PROTOCOL WITH is created for all senders to the session. On links that
TRAFFIC ENGINEERING (RSVP-TE): senders to the session share, a single label value is
The Resource Reservation Protocol with Traffic allocated to the session. If there is only one sender, the
Engineering (RSVP-TE) is an extension of RSVP that LSP looks like a normal point-to-point connection. When
utilizes the RSVP mechanisms to establish LSPs, multiple senders are present, a multipoint-to-point LSP
distribute labels and perform other label-related duties (a reversed tree) is created.
that satisfies the requirements for traffic engineering.
RSVP-TE supports both strict and loose routed LSPs that D. DISTRIBUTING LABELS WITH BORDER GATEWAY
do not have to follow conventional IP routing, giving PROTOCOL (BGP):
support also for traffic engineering. RSVP-TE is soft state The Border Gateway Protocol (BGP) can also be
protocol, which means that when a path has been setup used for label distribution. BGP is a routing protocol
by RSVP-TE it has to be continually updated to keep the used between different autonomous systems to
recourses reserved. RSVP-TE is a receiver-oriented exchange routing information. The update messages in
protocol, which means that requests for reservation are BGP that are used to distribute BGP routes can
made from the receiver end of the path. When RSVP-TE additionally carry the appropriate MPLS labels that are
is used for LSP setup the ingress router starts by sending mapped to the same BGP route. The label mapping
a PATH message on the path were a LSP shall be setup. information for a particular route is piggybacked in the
Each transit router on that path has to check if it has the same BGP update message that is used to distribute the
possibility to set up the requested LSP. If the requested route itself.
LSP is rejected an error message is returned upstream
until it reaches the ingress router. Otherwise the path RELATED WORK:
message is sent to the next transit router in the path In paper [1], depicts that for protecting a link
until it reaches the egress router. Then the egress router failures use a MPLS Fast Re-route (FRR) mechanism. FRR
sends a RESV message back through the path that the mechanism has two approaches such as link based or
PATH message travelled. local and other one is path based approach. In path
In the RESV message downstream routers based restoration approach, if single link failure occur
includes the label that they want the adjacent upstream then there need to re-route entire flow in the network.
router to use for the LSP that’s being setup. No While in link based or local approach, backup path
reservations are made in the routers until the RESV created for each link and if link failed then this link based
message is returned. A Record-Route-Object (RRO) is restoration only replace this failed link with backup path
included in both the PATH and RESV messages. In the without changing the rest of the route. This paper
PATH message the RRO is used to record each LSR and in depicts that the main objective is to maintain the
which order each LSR is visited. This list is then sent in connectivity after multiple failures without causing
the RESV message so that the each upstream router up to congestion. For distributing state information or routing
the ingress LSR receives this list. use some routing protocol such as Open Shortest Path
First (OSPF) and also reconfiguring backup paths use
RESERVATION STYLES: some distributed algorithm.
Each LSP can be reserved with a specific reservation In paper [2], depicts that FRR mechanism is
style. There are three types of reservation styles as beneficial over a link based or local based restoration.
97 | P a g e
 NOVATEUR PUBLICATIONS
International Journal Of Research Publications In Engineering And Technology [IJRPET]
ISSN: 2454-7875
VOLUME 2, ISSUE 12, Dec. -2016
This paper addresses the hybrid combination of p-cycles In paper [4], describe that Primary and backup
and FRR mechanism. While using only FRR backup paths paths in MPLS fast reroute (FRR) may be recognized as
are planned for each network link, the hybrid scheme shortest paths according to the administrative link costs
selects backup paths embedded within a set of p-cycles; of the IP control plane, or as explicitly calculated
this is based on holistic view of network performance arbitrary paths. The main objective is the maximum link
that is selecting the LSP which is less congested or less utilization for a set of considered failure scenarios is
traffic available on that LSP. This FRR protection is minimized. From this comparison shows that multiple
special case of p-cycle scheme because p-cycle scheme is explicit primary and backup paths allow lower maximum
a set of cycles are defined over the whole network such link utilization than unique explicit paths and unique
that each link is either on-cycle link or a straddling link primary and backup paths satisfying IP routing
(i.e., a link that connects two nodes on the same cycle but constraints may lead to higher maximum link utilization
is not itself part of the cycle). Hamiltonian p-cycle that is the use of explicit path layouts may increase the
created for whole network for used to protect all links. number of backup paths. Thus, a considerable
This scheme uses backup paths along a set of pre- improvement of the resource efficiency usage in
configured p-cycles that can be selected using design protected MPLS networks as compared to the simple
methodologies that consider the overall network setup of primary and backup paths with the IP control
performance. The benefits of the hybrid scheme increase plane can be obtained for the price of increased control
with the density of the network; hence adopting a p- plane complexity required for establishing optimized
cycle design is an attractive alternative for MPLS explicit paths and load balancing.
network operators.
In paper [3], depicts that several techniques
which are based on the IPFRR framework. These
techniques mainly focus on repairing paths rather than
mechanisms for fast failure detection. We propose a
routing technique, recursive Loop-Free Alternates
(rLFAs), to alleviate packet loss due to transient link
failures. This technique guarantees full repair coverage
for single link failures. This paper evaluates the Figure 4: Link Detour after Failure
performance of proposed system by simulations and also
shows that the incurred overheads that are pre- In paper [5], depicts that Fast-reroute
computed alternate paths and the failure-state Maximum mechanism especially for establishing backup path while
Link Utilization (MLU) are minimal. Several approaches link failures, but it is not effective for multiple failures
based on IP Fast Re-Route (IPFRR) in which alternate frequently occurring in backbone networks. Here
paths are pre-computed for fast re-route in presence of consider a protocol to reconfigure impacted backup
failures, have been proposed to alleviate (reduce) packet paths after a link failure, improving survivability from a
loss rate due to failures. The main objective of fast re- subsequent failure. Backbone network, router-to-router
route is to prevent packets from being dropped due to links carry the traffic of multiple end-to-end connections.
failures. If link failure occurs then all the connections traversing it
that failure link also fails. the main focus is on recovering
end-to- end connections using path protection
techniques. Although path protection is efficient in
resource utilization, it has the disadvantages of higher
complexity, poor scalability and large recovery times
requires. In link protection using MPLS fast reroute is to
pre-compute alternate paths to handle dual-link failures,
they are more complex. Because a first link failure may
affect the backup path of a other link, the pre-computed
backup paths for each link would have to consider all
possible combinations of failures of other links.
This paper also addresses, cross-layer
reconfiguration technique is used to improve
Figure 3: Comparison of Link Protection survivability from a subsequent link failure occur in the
MPLS network. Here uses OSPF-TE and RSVP and is a

98 | P a g e
 NOVATEUR PUBLICATIONS
International Journal Of Research Publications In Engineering And Technology [IJRPET]
ISSN: 2454-7875
VOLUME 2, ISSUE 12, Dec. -2016
natural extension to the MPLS fast-reroute. The main alternative path is dynamically recomputed after a fault
focus is that each node running a simple reconfiguration is detected. For both techniques, the alternative path can
algorithm independently. Further we can deal with be either global or local.
multiple concurrent failures in a scalable and adaptive The main focus is combine the protection-
manner by exploiting the capability of Layer 3 protocols switching algorithm with the rerouting algorithm and
(OSPF) to disseminate (i.e. spread information) the the choice of the algorithm is based on the performance
backup path information for a failed link, so as to criteria such as Fault recovery time, Packet loss, Packet
reconfigure other backup paths. reordering and Multiple faults. The challenge is to find
an efficient way to combine the two algorithms in order
to obtain a third one that would perform well in all four
these criteria.
In [8] depict that Multiprotocol Label Switching
(MPLS) technology enables configuration of end-to-end
virtual connections in communication networks,
especially in networks without connection-oriented
capabilities. Labeled packets can be sent over the
connections and forwarded according to the labels over
called as Label Switched Paths (LSPs). MPLS is able to
detect network failures locally and thus a failure-
detecting router can quickly switch all packets from
Figure 5: Backup path Reconfiguration failing primary LSP path to a backup LSP path just after a
failure is detected. This is called fast reroute (FRR)
In [6] depict that fast rerouting has been a key capability and the failure-detecting router is the called
component for providing service continuity to end users. point of local repair (PLR).
This paper focuses on improving current mechanisms This paper focuses on compact node-link
for Reliable and Fast Rerouting (RFR). These RFR formulations for MPLS fast reroute optimal single path
mechanisms are able to significantly reduce average layout. Also proposes mathematical formulations for
delay due to path restoration while eliminating packet MPLS fast reroute local protection mechanisms. The
disorder for traffic in MPLS networks for a protected comparison one-to-one (called detour) local protection
LSP. However, critical services or hard real time will be and many-to-one (backup) local protection mechanisms
affected by packet losses and, for TCP traffic, lost packets with respect to minimized maximum link utilization.
trigger retransmission requests; hence the gains due to In [9] depict that the author consider the two
the decrease in restoration time may become negligible. recovery possibilities for the alternative or backup LSP
The poor performance and degraded service delivery such as pre-established and dynamic recovery i.e.
will be experienced and QoS parameters will be rerouting. The objective is to provide a path protection
seriously affected during the restoration period. The mechanism in MPLS networks. The Haskin’s proposal
main factors that affect the performance of fast rerouting scheme uses a fault notification mechanism (FIS) to send
mechanisms are packet loss, traffic recovery delay i.e. the information about the occurrence of a fault to a
Full Restoration Time and packet disorder. The main responsible node in order to take the appropriate action
objective of this study is to provide and guarantee QoS to that failure such as in ingress LSR is notified to switch
for critical traffic carried by protected LSPs in MPLS traffic from the protected path to the alternative path.
networks and that not all LSPs are protected. The Haskin’s proposes method based on FRR mechanism
In [7] depict that Multiprotocol Label Switching and rerouting mechanism i.e. dynamic routing. This
is an architecture developed to combine the dynamic mechanism uses FRR with reversing backup for link
nature of IP routing protocols and the efficiency of label failure environment in MPLS network. This mechanism
switching. There is an issue in network such that it must beneficial for reducing the packet loss and but there is
support the real-time services or multimedia need for packet reordering. It is totally based on the FRR
applications even in the presence of node or link failures. mechanism in that it uses local mechanism for it path
MPLS employs two basic techniques for network recovery when there is link failure.
protection from such failures as first protection In [10] depict that the authors consider the two
switching where a pre-computed alternative path is set recovery possibilities for the alternative LSP such as pre-
up for every flow and second is rerouting where an established or FRR mechanism and dynamic recovery or
re-routing. The objective is to provide a path protection
99 | P a g e
 NOVATEUR PUBLICATIONS
International Journal Of Research Publications In Engineering And Technology [IJRPET]
ISSN: 2454-7875
VOLUME 2, ISSUE 12, Dec. -2016
mechanism in MPLS networks. This scheme uses a fault REFERENCES:
notification mechanism (FIS) to convey information 1) Rakesh K. Sinha, Funda Ergun, Kostas N. Oikonomou,
about the occurrence of a fault to a responsible node in K. K. Ramakrishnan, “Network Design for Tolerating
order to take the any action against the link failure. In Multiple Link Failures Using Fast Re-Route (FRR)”,
the case of using the pre-established alternative LSP or IEEE , 2014.
backup path, the traffic entering the domain is directly 2) Chang Cao, George N. Rouskas, “Hybrid FRR/p-Cycle
diverted to the pre-established alternative LSP by the MPLS Link Protection Design”, IEEE Communications
ingress LSR after the arrival of the notification signal. Society subject matter experts for publication in the
This method provides better resource utilization IEEE Globecom , 2011.
in than network than Haskin’s scheme because the 3) Suksant Sae Lor, Redouane Ali, Raul Landa, and
length of the protection path used during the recovery Miguel Rio, “Recursive Loop- Free Alternates for Full
period is less than that of Haskin’s proposal. However, Protection Against Transient Link Failures, IEEE ,
the traffic that is in transit during the interval of time 2010.
between the detection of the fault detected and the time 4) Micha Piro, Artur Tomaszewski, Cezary Zukowski,
the fault notification signal reaches the ingress LSR will David Hock, Matthias Hartman, Michael Menth,
be dropped by the alert LSR. Moreover, those packets “Optimized IP-Based vs. Explicit Paths for One-to-One
that were circulating on the failed link at the time of the Backup in MPLS Fast Reroute, ”, IEEE, 14th
failure will also be lost. When the dynamic method is International Telecommunications Network Strategy
applied, as it takes much longer to establish the and Planning Symposium(NETWORKS
alternative LSP, and the amount of dropped packets is 2010),Warsaw, Poland, sept 2010.
larger than with the pre-established alternative LSP or 5) Ajay Todimala, K. K. Ramakrishnan and Rakesh K.
backup path. Resource utilization is more efficient than Sinha, “Cross-layer Reconfiguration for Surviving
other scheme because updated network information is Multiple-link Failures in Backbone Networks”, IEEE,
used. This scheme also provides more flexibility in the 2009.
establishment of a new alternative LSP or backup path. 6) Lemma Hundessa and Jordi Domingo-Pascual,
“Reliable and Fast Rerouting Mechanism for a
CONCLUSION: Protected Label Switched Path”, Department d’
From this survey analysis of various recovery Arquitectura de Compotators Universitat
mechanisms of MPLS based on some performance Polit`ecnica de Catalunya (UPC) C/ Jordi Girona 1-3,
parameters. The parameters consider for analysis such 08034. Barcelona, Spain.
as resource requirement, fault recovery time, packet loss 7) Maria Hadjiona, Chryssis Georgiou and Vasos
ratio, packet re-ordering, complexity, optimal path Vassiliou, “A Hybrid Fault-Tolerant Algorithm for
option selection. The analysis can be done through any MPLS Networks”, Department of Computer Science
simulation tool such as Network Simulator version 2 University of Cyprus.
(NS2), NS3 etc. 8) Cezary ˙ Zukowski, Artur Tomaszewski, Michał
Pióro, David Hock, Matthias Hartmann and Michael
Menth, “Compact node-link formulations for the
optimal single path MPLS Fast Reroute layout”,
ADVANCES IN ELECTRONICS AND
TELECOMMUNICATIONS, VOL. 2, NO. 3, SEPTEMBER
2011.
9) D. Haskin and R. Krishnan, “A Method for Setting an
Alternative Label Switched Paths to Handle Fast
Reroute”, Work in progress, Internet draft <draft-
haskin-mpls-fast-reroute-05.txt>, November 2000.
10) S. Makam, V. Sharma, K. Owens, and C. Huang.
Protection/Restoration of MPLS Networks. Work in
progress, Internet draft <draft-makam-Mpls
protection- 00.txt>, October 1999.
11) Lemma Hundessa Gonfa , “ENHANCED FAST
REROUTING MECHANISMS FOR PROTECTED
Figure 6: Comparative Analysis of various MPLS based TRAFFIC IN MPLS NETWORKS”, UPC. Universitat
Recovery mechanisms Polit‘ecnica de Catalunya, 2011.
100 | P a g e
 NOVATEUR PUBLICATIONS
International Journal Of Research Publications In Engineering And Technology [IJRPET]
ISSN: 2454-7875
VOLUME 2, ISSUE 12, Dec. -2016
12) Johan Martin Olof Petersson,"MPLS Based Recovery
Mechanisms”, Master Thesis, 2005.
13) Rozita Yunos, Siti Arpah Ahmad, Noorhayati
Mohamed Noor, Raihana Md Saidi, Zarina Zaino,
“Analysis of Routing Protocols of VoIP VPN over MPLS
Network”, 2013 IEEE Conference on Systems,
Process & Control (ICSPC2013), 13 - 15 December
2013.
14) Md. Arifur Rahman, Ahmedul Haque Kabir, K. A. M.
Lutfullah, M. Zahedul Hassan, M. R. Amin,
“Performance Analysis and the Study of the behavior
of MPLS Protocols”, Proceedings of the International
Conference on Computer and Communication
Engineering 2008.
15) http://www.cisco.com/c/en/us/td/docs/ios/12_0s
/feature/guide/gslnh29.html.

101 | P a g e