Telecommunication network reliability Dr. Chidung LAC

1
Telecommunication network reliability Dr.
Chidung LAC
2
Outline

• Dependability and survivability
• Availability assessment
• Recovery mechanisms
• Physical layer
• ATM network
• IP level
• From random non-intentional to provocated
  failures

3
Dependability and survivability (1/2)
Quality of Service
Service
Accessibility
Security
Ease of use
Billing
continuity
Availability
Performances
Survivability
Degraded
Nominal
state
state
4
Dependability and survivability (2/2)
Unavailability
Failures
Congestion
Reliability
Traffic engineering
Network
System
Offered
Grade of
Maintenance
elements
Maintainability
capacity
traffic
service
support
Abnormal
situations
Man-made events
Natural disasters
(Security)
Dependability
(Safety)
5
Availability assessment methodology

• Identification of elements contributing to the
  services unavailability (equipments, cable
  infrastructure, OAM errors, )
• Block diagram representation (series, parallel)
  of these elements taking into account network
  architecture (redundancy, )
• Allocation of a failure rate l and a repair time
  MTTR to each element
• Estimated values from equipments vendors (l)
• Use of field data (l, MTTR)
• Service unavailability and failure rate
  assessment

6
Availability assessment example (1/2)
ADM
WDM
WDM
WDM
WDM
ADM
ADM
WDM
WDM
Amplifier
Optical distribution frame
Optical fiber
7
Availability assessment example (2/2)

• ADM block diagram representation
• Unavailability and failure rate calculation
• UBoard i lBoard i MTTRBoard i
• USerial system S UBoard j
• UParallel system P UBoard k ? UADM
• lSerial system S li
• lParallel system l1 U2 l2 U1 ? lADM

8
Pre-WDM networks protection

• Automatic protection switching

9
Pre-WDM self-healing rings (1/2)

• Unidirectional rings

10
Pre-WDM self-healing rings (2/2)

• Bidirectional rings

11
Pre-WDM meshed networks restoration
2
1
3
1
2
3
Failure
4
5
6
4
5
6
7
8
9
8
7
9
Around the failure
1
2
3
2
1
3
Failure
Failure
4
5
6
4
5
6
7
8
9
8
7
9
Local-destination
End-to-end
12
WDM rings protection
2-fiber bidirectional rings
13
WDM meshed networks (1/2)
1 2
1 2
(l
)
l
l

(l
)
1
2
1
Failure
1
l
l

1
2
1
2
3
l
2
l
1
4
5
6
Failure
5 6
5 6
(l
)
(l
)
1
1
Link protection
Dedicated approach
1 2
1 2
1 2
1 2
(l
)
(l
)
(l
)
(l
)
1
1
1
1
l
Failure
1
2
2
1
1
3
3
l
1
or
l
1
l
1
4
4
5
5
6
6
Failure
5 6
5 6
5 6
5 6
(l
)
(l
)
(l
)
1
1
1
(l
)
1
Shared approach
14
WDM meshed networks (2/2)
Path protection
15
ATM network (1/2)

• ATM v/s SDH-SONET
• Bandwidth on-demand v/s no dynamic bandwidth
  control
• Bursty traffic v/s switched services, private
  lines
• Logical v/s physical path structure
• Variations of VPs and VCs capabilities (from 0
  to the physical link rate) v/s fixed capacities
• Non-hierarchy v/s hierarchy path capacity
• Separation of capacity allocation and physical
  route assignment for VPs and VCs
• OAM bandwidth with allocation on demand ? delay
  reduction for restoration message exchange,
  quicker detection of system degradation
• Faster detection of soft failure

16
ATM network (2/2)

• Automatic protection switching ? VPs or VCs used
  as links (protection units)
• 11 connection switching by the receiver side
  node (from working to backup link)
• 11 switching done both by the transmitter
  side and the receiver side nodes
• Self-healing network
• Distributed control restoration scheme
• More effective resource utilization
• Dynamic planned scheme flooding algorithm to
  locate restoration routes
• Preplanned scheme predetermination of the
  optimal alternate route and pre-assignement of a
  backup VP to each VP

17
IP level

• Recovery mechanisms proposed for MultiProtocol
  Label Switching
• Architecture used
• Integrated Services with Resource Reservation
  Protocol (signaling protocol)
• Differentiated Services
• Main recovery modes
• Link protection 11 and 11
• Path protection
• Restoration/rerouting on-demand establishment
  of Label Switched Paths

18
From random non-intentional to provocated failures

• Telcos poor experience in the domain
• Example of assessment difficulty how to apply
  mathematical failure rate distribution ?
• Possible directions
• Infrastructures protection security
  reinforcement
• Logical intrusion main challenge with actuals
  networks interfunctioning (operators networks,
  Internet, )
• Behavior modelling human factors applied to
  volontary acts
• Some initiatives
• Critical Infrastructure Protection ? cascading
  effects due to interdependencies among different
  domains
• IEEE DIREN02, New York, June 2002

19
To probe further

• IEEE Communications Issue on Survivable
  communication networks, August 1999
• IEEE Communications Issue on Resilience in
  communication networks, January 2002
• IEEE Network Issue on Network and service
  security, Nov/Dec 2002
• ITU-T, Rec. E.800, Terms and definitions related
  to QoS and network performance including
  dependability
• Network reliability steering committee, Alliance
  for telecommunications industry solutions,
  http//www.atis.org
• T1A1.2 WG on Network Survivability Performance,
  Reliability/availability framework for IP-based
  networks and services, Draft technical report,
  July 2001