1. Overcomming Link/Router Failure In MPLS Networks
Yuval Hava
Pini Halperin

2. Introduction MPLS networks overview MPLS in the use of path recovery
Methods of recovery
Results

3. Ethernet, Frame Relay, ATM, PPP…
MPLS – Terminology
FEC – Forwarding Equivalence Classes
LER – Label Edge Router
LSR – Label Switching Router
LSP – Label Switching Path
LDP – Label Distribution Protocol IP
MPLS
Ethernet, Frame Relay, ATM, PPP…
Physical Layer

4. MPLS – Terminology MPLS Network
הסבר איך הרשת עובדת מבחינת כל המרכיבים ( בכלליות ) עם החבילות.
בהתחלה משתמש קצה מנסה להגיע למשתמש אחר
הLER מוסיף לחבילה לייבל
החבילה נעה ברשת MPLS עד שתגיע לLER ששמנו יוצאת
ביציאה LER מוריד את התוספת של החבילה

5. MPLS
Header Format
MPLS ‘Shim’ Headers (1-n) n
••• 1
Layer 2 Header
(eg. PPP, 802.3)
Network Layer Header
and Packet (eg. IP)
4 Bytes
Label Stack
Entry Format
Label
Exp. S
TTL
| |
| |
|- 1 -|
| |
Label: Label Value, 20 bits (0-16 reserved)
Exp.: Experimental / Class of Service, 3 bits
S: Bottom of Stack, 1 bit (1 = last entry in label stack)
TTL: Time to Live, 8 bits
MPLS header is inserted between layer 3 and layer 2 headers

6. MPLS – FEC FEC - forward equivalence class
LER classifies incoming IP traffic, relating it to the appropriate label by the FEC priority.
We establish routing paths (A…Z), and we call them forward equivalence classes, or FECs.
The FEC “A” paths are the highest-quality paths, and the FEC “Z” paths are the lowest-quality paths.

7. MPLS – FEC LSP IP1 IP2 IP1 IP2 IP1 #L1 IP2 IP1 #L2 IP2 IP1 #L3 IP2 LER
LSR
LSR
LER
LSP
IP1
IP2
IP1
IP2
IP1
#L1
IP2
IP1
#L2
IP2
IP1
#L3
IP2
FEC – חבילות שמסלולן ברשת הוא זהה יוגדרו כשייכות לאותו ה-FEC
Packets are destined for different address prefixes, but can be
mapped to common path
FEC - A subset of packets that are all treated the same way by a router

8. MPLS – LSR LSR – Label Switching Router
When packets leave the LER, they are destined for the LSR, then there examined for the presence of labels.
The LSR looks to its forwarding table.
LIB (label information base), the LSR will swap labels according to instructions in LIB table.

9. MPLS – LSR Example R1 X A B R4 D R3
None 20
Label Stack R2 Z R5

10. MPLS – LSR Example R1 X A B R4 D R3 20
600
Label Stack R2 Z R5

11. MPLS – LSR Example R1 X A B R4 D R3
600
Label Stack R2 Z R5

12. MPLS – LSP LSP – Label Switching Path
MPLS domain a path is setup for a given packet to travel based on an FEC
Two options for route selection for a particular FEC :
Hop by hop routing
Explicit routing
Hop-by-Hop routing
This method allows each LSR to independently choose the next hop for each FEC. It is similar to that currently used in IP networks. The LSR uses any available routing protocols like Open Shortest Path First (OSPF).
Within an MPLS domain a path is setup for a given packet to travel based on an FEC. The LSP setup for an FEC is unidirectional in nature. The return traffic must take another LSP. There are two different options available to select the LSP for a particular FEC.

13. MPLS – LSP
Explicit Routing (Source Routing) is a very powerful technique
With pure datagram routing, overhead of carrying complete explicit route is prohibitive
MPLS allows explicit route to be carried only at the time the LSP is setup, and not with each packet
MPLS makes explicit routing practical
In an explicitly routed LSP
LSP next hop is not chosen by the local node
Selected by a single node, usually the ingress
The sequence of LSRs may be chosen by
Configuration (e.g., by an operator or by a centralized server)
Selected dynamically by Ingress or Egress LSR
MPLS explicit routing much more efficient than the alternative of IP source routing
In MPLS the explicit route needs to be specified at the time when labels are assigned, but explicit route does not have to be specified with each IP packet. This makes MPLS explicit routing much more efficient than the alternative of IP source routing.

14. MPLS – LSP Example Consider the next parameters : LSR1 -> LSR7
Ingress
Egress
Consider the MPLS domain shown in Figure 2-7. Without explicit path routing, the tunnel is created hop by hop along the following path:
LSR 1 -> LSR 3 -> LSR 4 -> LSR 7
Suppose LSR 5 and LSR 6 are underused and LSR 4 is overused. In this case you might choose to configure the following explicit path because it forwards the data better than the hop-by-hop path:
LSR 1 -> LSR 3 -> LSR 5 -> LSR 6 -> LSR 7
physical connection
hop by hop LSP
explicit routing LSP
Consider the next parameters :
LSR1 -> LSR7
LSR5, LSR6 are underused
LSR4 is overused

15. MPLS Example of hop by hop path: R1 R2 R3 R4 Packet P
R1 analyzes P’s dest. And sets L1 to P
L1 is being pushed into MPLS label stack in P’s header
L1 represents LSP <R1,R2,R3,R4>

16. MPLS Example of hop by hop path: R1 R2 R3 R4 R21 R22 R23
R2 determines that P must pass through green tunnel
R2 pushes a new label: L2
During tunneling through R2, R21, R22, R23, R3, stack has depth 2
L2 represents LSP <R2, R21, R22, R23, R3>

17. MPLS Example of hop by hop path: R1 R2 R3 R4 R21 R22 R23
R3 finds out it is the final hop of L2 LSP and pops L2 from the stack
R3 discovers L1 on top of the stack and forwards P to the next hop for L1 LSP: R4

18. MPLS
Why use MPLS ?
IP-based forwarding is too slow for large traffic loads.
In MPLS, the lookup requires only one access to the forwarding table
Scalability: Label switching allowing large number of IP addresses with one or few labels
Route control
(exist in IP-based forwarding but is too messy)
Route control קיים בניתוב IP אבל הוא לא הכי נוח לשימוש:
המקור חייב לדעת מראש את מסלול הניתוב ליעד
תוספת הניתוב יכולה להיות ארוכה- overhead
לא כל הרשתות תומכות בזה

19. MPLS
With the needs of real-time, high priority, and mission critical application services, network reliability and survivability have become important issues in the Internet
Network failure is critical to these applications
IP based recovery mechanism may take a long time (10s of seconds) - may result in a large amount of packet loss
MPLS based protection: a protected LSP makes traffic travel through it at the same service quality regardless of any failures
הסבר על הקריטיות וחשיבות השיקום המהיר

20. MPLS – LSP Failure Recovery
All methods pre-establish a backup path
Quick recovery
2 common types of recovery:
Global repair
The ingress LSR establishes the backup path to the whole LSP
Local repair
Each LSR along the path establishes the backup path to the next hop
החלוקה ל-local ןל- global זהה לחלוקת שיטות הניתוב: hop by hop ו- explicit

21. RSVP-TE
RSVP defines a 'session' to be a data flow with a particular destination and transport-layer protocol
RSVP-TE – an extension to RSVP, defines resource reservation for IP systems
RFC 4090 efficiently uses RSVP-TE to establish backup label-switched path (LSP) tunnels for the local repair of LSP tunnels
RSVP – שיטה המגדירה session ברשת המכיל מידע מוגדר לגבי איכות התקשורת אל היעד
RSVP-TE – הרחבה המיועדת לרשתות IP
באמצעות RSVP-TE ניתן להגדיר בקלות מסלולי גיבוי של LSPs.

22. RSVP-TE – backup LSPs Meets the needs of real-time applications
Traffic should be redirected onto backup LSP tunnels in 10s of milliseconds
Can be satisfied by computing and signaling backup LSP tunnels in advance of failure and by re-directing traffic as close to the failure point as possible
The time for redirection includes no path computation and no signaling delays, including delays to propagate failure notification between label-switched routers (LSRs).
This extension will meet the needs of real-time applications such as voice over IP, for which user traffic should be redirected onto backup LSP tunnels in 10s of milliseconds. This timing requirement can be satisfied by computing and signaling backup LSP tunnels in advance of failure and by re-directing traffic as close to the failure point as possible. In this way, the time for redirection includes no path computation and no signaling delays, including delays to propagate failure notification between label-switched routers (LSRs).

23. RSVP-TE – backup LSPs One-to-one backup
Protected LSP: [R1-R2-R3-R4-R5]
R1's Backup: [R1-R6-R7-R8-R3]
R2's Backup: [R2-R7-R8-R4]
R3's Backup: [R3-R8-R9-R5]
R4's Backup: [R4-R9-R5] R6 R7 R8 R9
Detour – partial one-to-one backup LSP
For instance, if the link [R2-R3] fails , R2 will switch traffic received from R1 onto the protected LSP along link [R2-R7], using the label received when R2 created the detour. When R4 receives traffic with the label provided for R2's detour, R4 will switch that traffic onto link [R4-R5], using the label received from R5 for the protected LSP.
At no point does the depth of the label stack increase as a result of the detour.
While R2 is using its detour, traffic will take the path [R1-R2-R7-R8-R4-R5]
For N nodes, there could be as many as (N - 1) detours
Protected LSP: [R1->R2->R3->R4->R5]
R1's Backup: [R1->R6->R7->R8->R3]
R2's Backup: [R2->R7->R8->R4]
R3's Backup: [R3->R8->R9->R5]
R4's Backup: [R4->R9->R5]
We refer to a partial one-to-one backup LSP [R2->R7->R8->R4] as a detour.
To minimize the number of LSPs in the network, it is desirable to merge a detour back to its protected LSP, when feasible.
When a detour LSP intersects its protected LSP at an LSR with the same outgoing interface, it will be merged.
When a failure occurs along the protected LSP, the PLR redirects traffic onto the local detour.
For instance, if the link [R2->R3] fails in Example 1, R2 will switch traffic received from R1 onto the protected LSP along link [R2->R7], using the label received when R2 created the detour.
When R4 receives traffic with the label provided for R2's detour, R4 will switch that traffic onto link [R4-R5], using the label received from R5 for the protected LSP.
At no point does the depth of the label stack increase as a result of the detour. While R2 is using its detour, traffic will take the path [R1->R2->R7->R8->R4->R5].

24. RSVP-TE – backup LSPs Facility backup
Protected LSP 1: [R1-R2-R3-R4-R5] Protected LSP 2: [R8-R2-R3-R4] Protected LSP 3: [R2-R3-R4-R9]
Bypass LSP Tunnel: [R2-R6-R7-R4] R6 R7 R9
For instance, if link [R2-R3] fails , R2 will switch onto link [R2-R6]. The label will be switched for one which will be understood by R4 to indicate the protected LSP, and the bypass tunnel's label will then be pushed onto the label-stack of the redirected packets.
R4 will pop the bypass tunnel's label and examine the label underneath to determine the protected LSP that the packet is to follow.
When R2 is using the bypass tunnel for protected LSP 1, the traffic takes the path [R1-R2-R6-R7-R4-R5]; the bypass tunnel is the connection between R2 and R4.
There could be as many as (N-1) bypass tunnels to fully protect an LSP that traverses N nodes
R2 has built a bypass tunnel that protects against the failure of link [R2->R3] and node [R3]. The doubled lines represent this tunnel.
This technique provides a scalability improvement, in that the same bypass tunnel can also be used to protect LSPs from any of R1, R2, or R8 to any of R4, R5, or R9.
The example describes three different protected LSPs that are using the same bypass tunnel for protection.
As with the one-to-one method, there could be as many as (N-1) bypass tunnels to fully protect an LSP that traverses N nodes.
However, each of those bypass tunnels could protect a set of LSPs.
When a failure occurs along a protected LSP, the PLR redirects traffic into the appropriate bypass tunnel.
For instance, if link [R2->R3] fails in Example 2, R2 will switch traffic received from R1 on the protected LSP onto link [R2->R6]. The label will be switched for one which will be understood by R4 to indicate the protected LSP, and the bypass tunnel's label will then be pushed onto the label- stack of the redirected packets. If penultimate-hop-popping is used, the merge point in Example 2, R4, will receive the redirected packet with a label indicating the protected LSP that the packet is to follow. If penultimate-hop-popping is not used, R4 will pop the bypass tunnel's label and examine the label underneath to determine the protected LSP that the packet is to follow. When R2 is using the bypass tunnel for protected LSP 1, the traffic takes the path [R1->R2->R6->R7->R4->R5]; the bypass tunnel is the connection between R2 and R4.

25. Pre-Qualify
IETF’s recovery mechanisms have not considered optimal backup path for the recovery of an LSP in the occurrence of a network failure
Pre-qualified recovery path – establishing an optimal backup LSP during the working LSP setup time
Has a drawback: as time goes by, the network status changes
The pre-qualified recovery path may not be optimal at the time of failure
IETF הגדירו בגדול שני סוגים להגנה על LSP:
Protection switching
Rerouting model
אבל לא מבטיחים שמסלול הגיבוי יהיה אופטימאלי לסוג התעבורה, כמו המסלול המקורי

26. Efficient Pre-Qualify
During setup, each LSR calculates the pre-qualified recovery path to the next hop
Whenever each LSR receives routing update message (and information of current network parameters), the qualified recovery path is also updated immediately
When a fault occurs, the LSR establishes the recovery path using constraint-based LDP or sends FIS to LER
If LER cannot establish a recovery path, it notifies to network manager
Efficient pre-qualify מבוסס על שיטת explicit LSP אך עם local repair
כאשר LSR מזהה נפילה, הוא ישר מודיע על הקמת recovery path ע"י constraint-based LDP.
(constraint-based LDP – פרוטוקול להעברת הודעות "מפורשות" ליצירת ניתובים ברשת)
אם לא, אז שולח FIS – fault indication signal

27. Efficient Pre-Qualify
Simulation results:
Figure 5 shows the result of simulation. Due to using
he same backup path(LSR , in Figure 4), both the
umber of packet loss and re-ordering has no difference
ntil the amount of traffic of working and backup path
eaches 5Mb. However, the packet loss of new
roposed method is much smaller than that of an existing
ne, from the traffic is 6Mb. This result is caused that
ew proposed pre-qualified method uses the optimal
ackup path(LSR ), because congestion has
occurred after protection LSP establishment. In case of
packet re-ordering, the result of simulation has shown the
same trend.

28. Hundessa Fast rerouting MPLS
Upgraded D.Haskin method for backwards LSP.
Backwards LSP :
When a failure is detected the traffic is sent backwards to the ingress LSR using the pre-established LSP.
From the ingress LSR the data will now sent through the recovery path .
Drawback in this method is the delay involved in detecting the first packet plush the delay of the subsequent packets .
backwards path
Ingress
Egress
recovery path

29. Hundessa Fast rerouting MPLS
Hundessa proposed a better way to overcome D.Haskin method with respect to RTT delay and packet disorder .
When a fault is detected by LSR :
Each LSR on backward start storing incoming packets in local buffer
Last packet before initiating storing is tagged to indentify way back
Each LSR on the backward send back its stored packets when he received its tagged packet
We use one of the Exp field bits in MPLS header to avoid overheads
Ingress LSR sent its stored packet with all the new packets from the backwards LSR’s through the alternative LSP
Ingress
Egress

30. Hundessa Fast rerouting MPLS
A simulation was made in a network simulator for MPLS called MNS

31. FRR with PBT Fast ReRoute with Pre-establised Bypass Tunnels
Establishes bypass tunnels rather than backup paths
A tunnel back up all protected LSPs, not a particular one
Max-Flow-Min-Cut is adopted to find the necessary links through which all paths between LSRi and LSRj must pass
To protect both link failure and node failure, bypass tunnels are established around the next hop to the next-next hop

32. FRR with PBT Fast ReRoute with Pre-establised Bypass Tunnels
Shortest augmenting path algorithm:
1. Set the residual bandwidth of every link as the link bandwidth
Identify the shortest path from one LSR to the other.
If the path does not exit, then end the algorithm.
3. Discover the minimum residual bandwidth R in the path, and decrease the residual bandwidth by R for each link in the path. If the residual bandwidth of a link is zero, then set this link to be disconnected.
4. Store the path and go to step 2.

33. FRR with PBT Fast ReRoute with Pre-establised Bypass Tunnels
Each LSR establishes the bypass tunnels in the network initial state.
The bypass tunnels to the next-next hop are set up with the shortest augmenting path algorithm
By that, we establish the least amount of tunnels between 2 LSRs
PBT-D
Disjoint bypass tunnel for every link
Disjoint algorithm can reduce the searching time for the shortest path

34. Reroute Methods comparison
The simulation topology

35. Reroute Methods comparison
Packet loss vs Transmission rates (bps)
Pre-qualify has more packet loss since it reroutes the packets after failure

36. Reroute Methods comparison
CRR – the number of packets received without link failure divided by the number of packets received with link failure