BRKIPM-2008
Advanced Topics in IP Multicast Deployment
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 2
Agenda
Market Overview
PIM Configuration notes
Market Data Topologies
High Availability
Admission Control
Channel Changing
Receiver Tracking
Multicast in 802.11
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 3
Market Overview
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 4
Multicast Solutions
Finance (Trading, Market Data, Financial SP)
Tibco, Hoot-n-Holler, Data Systems
Enterprise Video and collaborative environments
Cisco TelePresence®, DMS, MP/WebEx Video Conferencing, Video Surveillance
Broadband (Entertainment)
Includes Cable, DSL, ETTH, LRE, Wireless
Broadcast TV / IP/TV, VOD, Connected Home
Service Provider (Transit Services)
Native v4 and v6
Label Switched Multicast (LSM)
Multicast VPNs (IP and MPLS-based)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 5
Multicast for IP/TV Delivery
Multicast
1. Efficiently Controls network traffic
2. Reduces server and CPU loads
3. Eliminates traffic redundancy
4. Makes Multipoint applications possible
Multicast Benefits
Increase Productivity and Save Cost
Generate New Revenue Stream
0
2
4
6
8
Ne
two
rk T
raff
ic
1 20 40 60 80 100# Clients
Multicast
Unicast
Distribute Information to Large Audiences
over an IP Network
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 6
Market Data Architectural ComponentsService
Distribution
Network
Stock Exchange
Data CenterTraders
Financial Service
Providers
A Feed
B Feed
Location A
Location B
Brokerage
Data Center
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 7
Mobile
Access anywhere
Seamless
Secure
Similar look and feel
Ease of use
Mobility Vision
Coffee Shop
Hotel
Plane Airport LoungeCar Train
Train Station
Customer
Intelligent Mobile Device
Web AppsTelepresence
eLearning &
Communication
Access
Footprint
Streaming
Media
Video
SurveillanceCommunication and Collaboration
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 8
PIM Configuration Notes
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 9
PIM Modes
Bidirectional PIM (PIM-Bidir)
Supports only Shared Trees
Sparse Mode (PIM-SM) (ASM)
Original (Classic)
Supports both Shared and Source Trees
Source Specific Multicast (PIM-SSM)
aka Single Source Multicast
Supports only Source Trees
No need for RPs, RP Failover, etc.
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 10
PIM Mode Selection Criteria
Use SSM
For One-to-Many applications
Eliminates need for RP Engineering
Greatly simplifies network
Data and Control Planes are decoupled
Use Bidir
For Many-to-Many | Few applications
Drastically reduces total (S,G) state in network
Data and Control Planes are decoupled
Use PIM-SM
For all other general purpose applications
SSM/BiDir and SM can all coexist in the same network at the same time and applications can be moved gracefully (groupwise) between modes
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 11
Intermittent Source Applications
Definition:
Applications with sources that temporarily stop sending for > 3 minutes
Impact:
(S,G) state times out within the network
Initial packets often lost during SPT switchover
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 12
Solutions to Intermittent Sources
PIM-Bidir or PIM-SSM
No data driven events
Periodic keepalives or heartbeats
S,G Expiry timer
Needs to be set on every router
Typically set to maintain state throughout entire trading day
ip pim sparse sg-expiry-timer <secs>
Available 12.2(18)SXE5, 12.2(18)SXF4, 12.2(35)SE
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 13
Static RPs
Hard-coded RP address
When used, must be configured on every router
All routers must have the same RP address (per group)
Anycast RP must be used for Redundancy
Configuration
ip pim rp-address <address> [group-list <acl>]
[override][bidir]
Optional group list specifies group range
Default: Range = 224.0.0.0/4
Override keyword ―overrides‖ Auto-RP information
Default: Auto-RP and BSR learned info takes precedence
Bidir keyword specifies this group range as PIM-Bidir
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 14
Why Use Auto-RP?
Auto-RP is a dynamic method for the network to learn RP to Group mapping information
This helps when:
RP address and group ranges change often
Your network has 100s or 1000s of routers and you want to simplify the config
There are several RPs for different applications
RPs maintained by different administrative groups
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 15
Auto-RP Configuration—Auto-RP Listener
Use global command (recommended)ip pim autorp listener
Added support for Auto-RP Environments
Modifies interface behavior:
Interface can be configured in SM
Interface always uses DM for ONLY Auto-RP groups
Only needed if Auto-RP is to be used
Available 12.3(4)T, 12.2(28)S, 12.1(13)E7
Use with interface command
ip pim sparse-mode Recommended
Prevents DM Flooding
No longer need:ip pim sparse-dense-mode
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 16
Dense Mode Fallback
Caused by loss of local RP information (in older IOS® releases)
Entry in Group-to-RP mapping cache times out
Can happen when:
All C-RPs fail
Auto-RP/BSR mechanism fails
Possibly a result of network congestion
Group is switched over to Dense mode
(*,G) mroute entry changes to Dense mode
Flags of (S,G) entries change from JT to T
(S,G) PIM Joins are no longer sent
State times out on upstream router—traffic flow stops
All existing PIM-SM SPTs are dropped!
Dense mode flooding begins if interfaces configured with:
ip pim sparse-dense-mode
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 17
Avoiding DM Fallback in Auto-RP/BSR
Use RP-of-last-resort
Assign local Loopback as RP-of-last-resort on each router
Example
ip pim rp-address <local_loopback> 10
access-list 10 deny 224.0.1.39
access-list 10 deny 224.0.1.40
access-list 10 permit any
Needed Prior to IOS 12.3(4)T, 12.2(33)SXH
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 18
Avoiding DM Fallback Automatically
New IOS global command
no ip pim dm-fallback
Totally prevents DM Fallback!
No DM Flooding (since all state remains in SM)
Default RP Address = 0.0.0.0 [nonexistent]
Used if all RPs fail
All SPTs remain active
Behavior is enabled by default if all interfaces are in sparse mode
Available 12.3(4)T, 12.2(33)SXH
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 19
Auto-RP Summary
Use Auto-RP Listener with Sparse Mode interfaces
With newer code you automatically get No Dense Mode Fallback
With older code that doesn‘t have:
no ip pim dm-fallback
Use RP of Last Resort to prevent loss of active S,G entries in case of RP failure
With ancient code that doesn‘t have AutoRPListener
Use sparse-dense interfaces
Use RP of Last Resort
Or Upgrade ;-)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 20
Anycast RP Overview
MSDP
RecRec RecRec
Src Src
SA SAA
RP1
10.1.1.1B
RP2
10.1.1.1
X
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 21
Anycast RP Overview
RecRec RecRec
SrcSrc
A
RP1
10.1.1.1B
RP2
10.1.1.1
X
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 22
Anycast RP Configuration
Interface loopback 0
ip address 10.0.0.1 255.255.255.255
ip pim sparse-mode
Interface loopback 1
ip address 10.0.0.2 255.255.255.255
!
ip msdp peer 10.0.0.3 connect-source loopback 1
ip msdp originator-id loopback 1
Interface loopback 0
ip address 10.0.0.1 255.255.255.255
ip pim sparse-mode
Interface loopback 1
ip address 10.0.0.3 255.255.255.255
!
ip msdp peer 10.0.0.2 connect-source loopback 1
ip msdp originator-id loopback 1
MSDPB
RP2
A
RP1
C D
ip pim rp-address 10.0.0.1 ip pim rp-address 10.0.0.1
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 23
Combining Auto-RP and Anycast-RP
Provides advantages of both methods
Rapid RP failover of Anycast RP
No DM Fallback
Configuration flexibility of Auto-RP
Ability to effectively disable undesired groups
Anycast RP and Auto-RP May Be Combined
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 24
Anycast RP with Auto-RP Configuration
Interface loopback 0
ip address 10.0.0.1 255.255.255.255
Interface loopback 1
ip address 10.0.0.2 255.255.255.255
!
ip pim send-rp-announce loopback 0 scope 32
ip pim send-rp-discovery loopback 1 scope 32
!
ip msdp peer 10.0.0.3 connect-source loopback
1
ip msdp originator-id loopback 1
Interface loopback 0
ip address 10.0.0.1 255.255.255.255
Interface loopback 1
ip address 10.0.0.3 255.255.255.255
!
ip pim send-rp-announce loopback 0 scope 32
ip pim send-rp-discovery loopback 1 scope 32
!
ip msdp peer 10.0.0.2 connect-source loopback 1
ip msdp originator-id loopback 1
MSDPB
RP2
A
RP1
ip multicast-routing
ip pim autorp-listener
no ip pim dm-fallback
C
ip multicast-routing
ip pim autorp-listener
no ip pim dm-fallback
D
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 25
Bidir PIM—Phantom RP
Question: Does a Bidir RP even have to physically exist?
Answer: No. It can just be a phantom address.
RP
Receiver 2
E1 (DF) E1 (DF) E1 (DF) E1 (DF)
E1 (DF) E1 (DF)
E0 (DF)
C
E0 E0 E0 E0
E0 E0
Source Receiver 1
F
D
E
A B
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 26
ip multicast-routing
!
interface Loopback0
ip address 1.1.1.1 255.255.255.252
ip pim sparse-mode
ip ospf network point-to-point
!
interface Ethernet0/0
ip address 10.1.1.1 255.255.255.0
ip pim sparse-mode
!
interface Ethernet1/0
ip address 10.1.2.1 255.255.255.0
ip pim sparse-mode
!
router ospf 11
network 1.1.1.0 0.0.0.3 area 0
network 10.1.1.0 0.0.0.255 area 0
network 10.1.2.0 0.0.0.255 area 0
!
ip pim bidir-enable
ip pim rp-address 1.1.1.2 bidir
Phantom RP on Point-to-Point Core
RP: 1.1.1.2
ip multicast-routing
!
interface Loopback0
ip address 1.1.1.1 255.255.255.248
ip pim sparse-mode
ip ospf network point-to-point
!
interface Ethernet0/0
ip address 10.1.1.2 255.255.255.0
ip pim sparse-mode
!
interface Ethernet1/0
ip address 10.1.2.2 255.255.255.0
ip pim sparse-mode
!
router ospf 11
network 1.1.1.0 0.0.0.7 area 0
network 10.1.1.0 0.0.0.255 area 0
network 10.1.2.0 0.0.0.255 area 0
!
ip pim bidir-enable
ip pim rp-address 1.1.1.2 bidir
SP
30 Bit Mask 29 Bit Mask
OSPF requires
P2P interfaces
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 27
Phantom RP with Auto-RP
ip pim send-rp-announce 1.1.1.2 scope 32 bidir
ip pim send-rp-discovery Loopback1 scope 32
interface Loopback0
ip address 1.1.1.1 255.255.255.252
ip pim sparse-mode
ip pim send-rp-announce <[int] | [ip-address]> scope [group-list] [bidir]
Previously, Auto-RP could only advertise IP address on interface (e.g. loopback) as RP
New option has been added—now we can advertise any address on a directly connected subnet
In example below, Phantom RP address is being advertised through Auto-RP; the source of the Mapping packets are the address on Loopback1
Available 12.4(7)T, 12.2(18)SXF4
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 28
Market Data Topologies
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 29
Brokerage
Content
ProviderContent
ProviderContent
Provider
Financial
Service
Provider
Brokerage Brokerage
Market Data Distribution—Interface
Static Forwarding
Static Service Levels—Cable Model
Dynamic Forwarding
Hybrid Design
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 30
Traditional MD Interface Requirements
Customers and Providers Prefer Lowest Common Denominator—Least Coordination
Providers are under contract to deliver stream
Each side wants to limit organizational liability and coordination—KISS
Ideal Scenario: Provider and Customer have separate multicast domains
Therefore:
Traffic is statically nailed up
No PIM Neighbors
No DR on edge
No PIM Joins
No shared RP
No MSDP
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 31
Market Data
Source Network
Customer
224.0.2.64
DestinationSource
10.2.2.2
e0
e1
interface Ethernet0
ip address 10.1.2.1 255.255.255.0
ip pim sparse-mode
ip igmp static-group 224.0.2.64
Virtual RP
interface Ethernet1
ip address 10.1.2.2 255.255.255.0
ip pim sparse-mode
ip pim rp-address 10.1.1.1
ip route 10.1.1.1 255.255.255.255 10.1.2.5
router ospf 11
network 10.1.0.0 0.0.255.255 area 0
redistribute static subnets
ip pim rp-address 10.1.1.1
MD Distribution—Virtual RP
MD feed is statically nailed up
Customer Edge router advertises RP address from upstream interface
Every router in customer network needs to know about the RP
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 32
Market Data
Source Network
Customer
224.0.2.64
DestinationSource
10.2.2.2
e0
e1
interface Ethernet0
ip address 10.1.2.1 255.255.255.0
ip pim sparse-mode
ip igmp static-group 224.0.2.64
RP
interface Ethernet1
ip address 10.1.2.2 255.255.255.0
ip pim sparse-mode
interface Loopback0
ip address 10.1.1.1 255.255.255.255
ip pim sparse-mode
ip pim rp-address 10.1.1.1
ip pim rp-address 10.1.1.1
MD Distribution—Edge Router Is RP
MD feed is statically nailed up
Customer Edge router is RP—so that it will accept a non-connected source
Every router in customer network needs to be know about the RP
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 33
Market Data
Source Network
Customer
224.0.2.64
DestinationSource
10.2.2.2
e0
e1
interface Ethernet0
ip address 10.1.2.1 255.255.255.0
ip pim sparse-mode
ip igmp static-group 224.0.2.64
RP
interface Ethernet1
ip address 10.1.2.2 255.255.255.0
ip pim sparse-mode
interface Loopback0
ip address 10.1.1.1 255.255.255.255
ip pim sparse-mode
ip pim rp-address 10.1.1.1
ip pim rp-address 10.1.1.1
MD Distribution—Edge Router Is RP
MD feed is statically nailed up
Customer Edge router is RP—so that it will accept a non-connected source
Every router in customer network needs to be know about the RP
This Method Will Not Work
with Future Versions of IOS
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 34
Market Data
Source Network
Customer
224.0.2.64
DestinationSource
10.2.2.2
e0
e1
RP
interface Ethernet1
ip address 10.1.2.2 255.255.255.0
ip pim dense-mode proxy-register list 100
access-list 100 permit ip any any
ip pim rp-address 10.1.1.1
interface Loopback0
ip address 10.1.1.1 255.255.255.255
ip pim sparse-mode
ip pim rp-address 10.1.1.1
interface Ethernet0
ip address 10.1.2.1 255.255.255.0
ip pim sparse-mode
ip igmp static-group 224.0.2.64
MD Distribution—Edge Router Proxy Registers to RP
MD feed is statically nailed up
Customer Edge router has dense-mode on IIF and proxy registers to RP
RP is configured inside customer network
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 35
Market Data
Source Network
Customer
224.0.2.64
DestinationSource
10.2.2.2
e0
e1
RP
interface Ethernet1
ip address 10.1.2.2 255.255.255.0
ip pim dense-mode
interface Loopback0
ip address 10.1.1.1 255.255.255.255
interface Loopback1
ip address 10.1.3.2 255.255.255.255
ip pim rp-address 10.1.1.1
ip msdp peer 10.1.3.1 connect-source Loopback1
ip msdp originator-id Loopback1
RP
interface Loopback0
ip address 10.1.1.1 255.255.255.255
interface Loopback1
ip address 10.1.3.1 255.255.255.255
ip pim rp-address 10.1.1.1
ip msdp peer 10.1.3.2 connect-source Loopback1
ip msdp originator-id Loopback1
MD Distribution—Edge Router Is RP and MSDP Peer
Customer Edge router is RP and MSDP peer of main customer RP
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 36
Market Data
Source Network
Customer
224.0.2.64
DestinationSource
10.2.2.2
e0
e1
RP
interface Ethernet1
ip address 10.1.2.2 255.255.255.0
ip pim dense-mode
interface Loopback0
ip address 10.1.1.1 255.255.255.255
interface Loopback1
ip address 10.1.3.2 255.255.255.255
ip pim rp-address 10.1.1.1
ip msdp peer 10.1.3.1 connect-source Loopback1
ip msdp originator-id Loopback1
RP
interface Loopback0
ip address 10.1.1.1 255.255.255.255
interface Loopback1
ip address 10.1.3.1 255.255.255.255
ip pim rp-address 10.1.1.1
ip msdp peer 10.1.3.2 connect-source Loopback1
ip msdp originator-id Loopback1
MD Distribution—Edge Router Is RP and MSDP Peer
Customer Edge router is RP and MSDP peer of main customer RP
Dense mode is required on the IIF so
that the A flag will be set and MSDP
will forward an SA
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 37
Static Forwarding—Cable Model
Basic Service
ip access-list standard basic-service
permit 239.192.1.0 0.0.0.255 ! Basic service channels
Premium Service
ip access-list standard premium-service
permit 239.192.1.0 0.0.0.255 ! Basic service channels
permit 239.192.2.0 0.0.0.255 ! Premium service channels
Premium Plus Service
ip access-list standard premium-plus-service
permit 239.192.1.0 0.0.0.255 ! Basic service channels
permit 239.192.2.0 0.0.0.255 ! Premium service channels
permit 239.192.3.0 0.0.0.255 ! Premium Plus service channels
Adapt Cable Model of Provisioning for Market Data by
qualifying multicast boundary with each of following:
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 38
interface Vlan6
ip igmp static-group 224.0.2.64
ip igmp static-group 224.0.2.65
ip igmp static-group 224.0.2.66
...
ip igmp static-group 224.0.2.80
Static Forwarding—Group Range Command
Subscribing dozens or hundreds of groups can be cumbersome with the static-group command:
The static group range command simplifies the config:
Available in 12.2(18)SXF5
class-map type multicast-flows
market-data group 224.0.2.64 to 224.0.2.80
interface Vlan6
ip igmp static-group class-map market-data
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 39
Advantages of Static Forwarding
Provider and Customer Have Separate Multicast Domains
Each is free to use any forwarding model, e.g. PIM-SM, PIM-SSM, PIM-Bidir
Each is responsible for their portion of the delivery model—clear demarcation
Simple, straight-forward
Has traditionally been first choice for FSP
Main Disadvantage
Customer is not able to control subscriptions and bandwidth usage of last mile dynamically
As data rates climb this is more of a issue
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 40
Dynamic Forwarding
Rising data rates and 24 hour trading are driving the requirement for dynamic subscriptions
Methods:
IGMP Membership Reports
PIM Joins—*,G for PIM-SM and PIM-Bidir
PIM Joins—S,G for PIM-SSM
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 41
Dynamic Service—Static Subscriptions with IGMP
Customers Want Ability to Nail Up Service
Existing Issues
ip igmp join-group <group>
Sends an IGMP report out the interface
Traffic gets punted to CPU
ip igmp static-group <group>
Adds interface to OIL
Does not send IGMP report out the interface
Workarounds
Separate router—Put IGMP join group on a dedicated router
Need Better Solution
ip igmp join-group <group> passive —under consideration
IGMP report will be sent but traffic will not be punted to CPU
IGMP host code on router will respond to queries
L flag will not be set?
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 42
Market Data
Source Network
Customer
224.0.31.20
DestinationSource
10.2.2.2
e0
e1
IGMP
IGMP
PIM
MD Distribution—Provider Wants IGMP Report
Assumes that hosts sit on edge of customer network or breaks multicast delivery model
Stretches the original design and purpose of IGMP
In deployment today
We can make this work dynamically today with a combination of:
ip igmp helper
ip igmp proxy-service
ip igmp mroute-proxy
Industry may want to recommend this model going forward
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 43
Market Data
Source Network
e0
IGMP
IGMP
interface Loopback1
ip address 10.3.3.3 255.255.255.0
ip pim sparse-mode
ip igmp helper-address 10.4.4.4
ip igmp proxy-service
ip igmp access-group filter-igmp-helper
ip igmp query-interval 9
interface Ethernet0
ip address 10.2.2.2 255.255.255.0
ip pim sparse-mode
ip igmp mroute-proxy Loopback1
ip pim rp-address 20.20.20.20
ip route 20.20.20.20 255.255.255.255 10.4.4.4
e1
Customer
e0loopback1
PIM
10.4.4.0/24
MD Distribution—igmp mroute-proxy
igmp proxy service and helper are configured on loopback
Downstream interface is configured with igmp mroute-proxy
Every router in customer network needs to be know about the virtual RP
Virtual RP: 20.20.20.20
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 44
Market Data
Source Network
e0
IGMP(*, 239.254.1.0), 00:00:01/00:02:55, RP 20.20.20.20, flags: SC
Incoming interface: FastEthernet1/15, RPF nbr 10.2.2.2, RPF-MFD
Outgoing interface list:
Vlan194, Forward/Sparse, 00:00:01/00:02:55, H
e1
Customer
e0loopback1
PIM
10.4.4.0/24
MD Distribution—igmp mroute-proxy detail
4. The first PIM *,G Join on e0 triggers an unsolicited IGMP report to be generated on the loopback1 interface
3. PIM *,G Join message is received on e0 interface and mroute state is created; the igmp mroute-proxy command on interface causes special internal flag to be added to mroute
2. PIM *,G Join message filters up towards virtual RP
1. Host sends IGMP report and creates mroute state
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 45
Market Data
Source Network
e0
IGMP
IGMPe1
Customer
e0loopback1
PIM
10.4.4.0/24
MD Distribution—igmp mroute-proxy detail
6. When the periodic IGMP query is run on loopback1 the igmp proxy-service command initiates a walk through the mroutetable looking for the mroute-proxy flag; an IGMP report is generated for each mroutewith the mroute-proxy flag set
As long as the mroute is kept alive with PIM joins the IGMP reports will be forwarded
5. The igmp helper command directs the IGMP report out the e1 interface
IGMP reports are dynamic—they are only sent when there is interest in the customer domain; however, the edge router does not respond to queries from provider router
Consideration:Seven times more IGMP messages than needed
Complex config
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 46
Dynamic Service—Dynamic Subscriptions with IGMP
Better Solution: IGMP Host Proxy
Global command:
ip igmp host-proxy <group acl>
Functionality:
1. When mroute state is created for any group defined by <acl>
The mroute-proxy flag will be set
An IGMP report is generated for the mroute and set out the IIF of the mroute
2. When an IGMP Query is received for a group defined by <acl>
If mroute state exists, generate IGMP report
3. When a PIM Prune is received for a group defined by <acl>
An IGMP leave message is generated
4. All modes of PIM should be supported—PIM-SM, PIM-Bidir and PIM-SSM; the groups must be defined on the router and the behavior will work appropriately—this solution should be compliant with RFC4605
Coming
Soon
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 47
Market Data
Source Network
Customer
224.0.2.64
DestinationSource
10.2.2.2
e0
e1
RP
RP
RP
MD Distribution—Other Options
Provider accepts PIM join
Sparse Mode
Provider must supply RP addr
Requires PIM Neighbor relationship
No RP on customer Side
One multicast domain
Source Specific Multicast
Provider must supply S,G info
Requires PIM Neighbor relationship
MSDP
Standard Interdomain Multicast
Requires peering relationship
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 48
Dynamic Forwarding—S,G PIM Joins
Works in situations that are ideal for SSM
No need to share RP info or use MSDP
Redundancy options:
Host Side
Host can join both primary and secondary servers—for both A and B streams
Host will need to arbitrate between primary and standby
Network/Server Side
Anycast Source—Hosts only join one server and network tracks server and forwards active stream
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 49
Market Data Design Whitepapers
Market Data Network Architecture (MDNA)
Trading Floor Architecture
Design Best Practices for Latency Optimization
IP Multicast Best Practices for Enterprise Customers
http://www.cisco.com/go/financial
A Set of Four Documents that Cover All Aspects of Network and Application Design for Market Data Distribution
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 50
High Availability
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 51
IP/TV Deployments Today
Two schools of thought in deployments today:
I think I need 50ms convergence
IPMulticast is fast enough
IP Multicast is UDP
The only acceptable loss is 0ms
How much is ―reasonable‖?
50ms ―requirement‖ is not a video requirement
Legacy telco voice requirement
Efforts for 50ms only cover a limited portion network events
Where to put the effort?
Make IPMulticast better?
Improve the transport?
Add layers of network complexity to improve core convergence?
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 52
Application Side Resiliency
FEC: Forward Error Correction
Compensate for statistical packet loss
Use existing FEC, e.g. for MPEG transport to overcome N msec (>= 50 msec) failures?
Cover loss of N[t] introduces delay > N[t]!
Retransmissions
Done e.g. with vendor IP/TV solutions—unicast retransmissions
Candidate large bursts of retransmissions!
Limit #retransmissions necessary
Multicast retransmissions (e.g. PGM ?)
No broadcast IP/TV solutions use this
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 53
Impact of Packet Loss on MPEG Stream
Compressed Digitized Video is sent as I, B, P Frames
I-frames: Contain full picture information
Transmit I frames approximately every 15 frames (GOP interval)
P-frames: Predicted from past I or P frames
B-frames: Use past and future I or P frames
I B B P B B P B B P B BI B B P B B P B B P B B
I-Frame Loss “Corrupts” P/B Frames for the Entire GOP
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 54
Service Availability Overview
IP Host Components Redundancy
Single transmission from Logical IP address
Anycast — Use closest instance
Prioritycast — Use best / preferred instance
Benefit over anycast: no synchronization of sources needed, operationally easier to predict which source is used
Signaling host to network for fast failover
RIPv2 as a simple signaling protocol
Normal Cisco IOS/IGP configuration used to inject these source server routes into the main IGP being used (OSPF/ISIS)
Dual Transmission with Path separation
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 55
Video Source Redundancy: Two Approaches
Primary Backup Live-Live/Hot-Hot
Two sources: one is active and src‘ing content, second is in standby mode (not src‘ing content)
Heartbeat mechanism used to communicate with each other
Two sources, both are active and src‘ing multicast into the network
No protocol between the two sources
Only one copy is on the network at any instant
Single multicast tree is built per the unicast routing table
Two copies of the multicast packets will be in the network at any instant
Two multicast trees on almost redundant infrastructure
Uses required bandwidth Uses 2X network bandwidth
Receiver‘s functionality simpler:
Aware of only one src, failover logic handled between sources
Receiver is smarter:
Is aware/configured with two feeds (s1,g1), (s2,g2) / (*,g1), (*,g2)
Joins both and receives both feeds
This approach requires the network to have fast IGP and PIM convergence
This approach does not require fast IGP and PIM convergence
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 56
Source Redundancy: Anycast/Prioritycast Signaling
Redundant sources or NMS announce Source Address via RIPv2
Per stream source announcement
Routers redistribute (with policy)into IGP
Easily done from IP/TV middleware (UDP)
No protocol machinery required—only periodic announce packets
Small periodicity for fast failure detection
All routers support RIPv2 (not deployed as IGP):
Allows secure constrained configuration on routers
Src
RIP (v2)
Report (UDP)
Router
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 57
Anycast-Based Load Balancing
1.1.1.1 1.1.1.1
IGMP Report IGMP Report
I Will Send Join
to the Nearest
1.1.1.1/32
I Will Send Join
to the Nearest
1.1.1.1/32
PIM Join PIM Join
Service Router 2Service Router 1
Agg RouterAgg Router
STB STB
Source 2Source 2
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 58
Encoder Failover Using Anycast
1.1.1.1 1.1.1.1
Service Router 2Service Router 1
Agg RouterAgg Router
STB STB
Source 2Source 2
IGP Recalc >>
PIM Join
X
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 59
Source RedundancyAnycast/Prioritycast Policies
Policies
Anycast: Clients connect to the closest instance of redundant IP address
Prioritycast: Clients connect to the highest-priority instance of the redundant IP address
Also used in other places
e.g. PIM-SM and Bidir-PIM RP redundancy
Policy simply determined by routing announcement and routing config
Anycast well understood
Prioritycast: Engineer metrics of announcements or use different prefix length
Src B
Secondary
10.2.3.4/32
Rcvr 2Rcvr 1
Src A
Primary
10.2.3.4/31
Example: Prioritycast with
Prefixlength Announcement
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 60
Source RedundancyAnycast/Prioritycast Benefits
Sub-second failover possible
Represent program channel as single (S,G)
SSM: single tree, no signaling; ASM: no RPT/SPT
Move instances ―freely‖ around the network
Most simply within IGP area
Regional to national encoder failover (BGP…)?
No vendor proprietary source sync proto required
Per program, not only per-source-device failover
Use different source address per program
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 61
FRR for Native IP Multicast/mLDP
Do not require RSVP-TE for general purpose multicast deployments
Sub 50 msec FRR possible to implement for PIM or mLDP
Make-before-break during convergence
Use of link-protection tunnels
Initial: one-hop RSVP-TE P2P tunnels
Future: NotVia IPFRR tunnels (no TE needed then)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 62
Multicast only Fast ReRoute - MoFRR
It is make-before-break solution
Multicast routing doesn‘t have to wait for unicastrouting to converge
An alternative to source redundancy, but:
Don‘t have to provision sources
Don‘t have to sync data streams
No duplicate data to multicast receivers
No repair tunnels
No new setup protocols
No forwarding/hardware changes
http://tools.ietf.org/html/draft-karan-mofrr-00
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 63
MoFRR - Concept Example
S
R
BJoin
Path
Data Path
Alt
Path
Alt Data Path
Wasted Bandwidth
Wasted Bandwidth
R
Not
1. D has ECMP path {BA, CA} to S
2. D sends join on RPF path through C
3. D can send alternate-join on BA path
4. A has 2 oifs leading to a single receiver
5. When RPF path is up, duplicates come to D
6. But D RPF fails on packets from B
7. If upstream of D there are
receivers, bandwidth is only
wasted from that point to D 8. When C fails or DC link fails, D makes
local decision to accept packets from B
9. Eventually unicast routing says B is new
RPF path
rpf Path (RPF Join)
Alt Join (Sent on Non-rpf)
Data Path
Interface in oif-list
Link Down or RPF-Failed Packet Drop
DD
AA
BB CC
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 64
Multicast Fast Convergence
IP multicast
All failures/topology changes are corrected byre-converging the trees
Re-convergence time is sum of:
Failure detection time (only for failure cases)
Unicast routing re-convergence time
~ #Multicast-trees x PIM re-convergence time
Possible
~ minimum of 200 msec initial
~ 500 ... 4000 trees convergence/sec (perf)
Same behavior with mLDP
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 65
Multicast Node Protectionwith p2p Backup Tunnels
If router with fan-out of N fails, N-times as much backup bandwidth as otherwise is needed
Provisioning issue depending on topology!
Some ideas to use multipoint backup to resolve this, but…
Recommendation? Rely on Node HA instead!!
S(ource)Rcvr1
R1 R2 R3
R4 R5 R6
Rcvr2
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 66
Lin
e C
ard
Lin
e C
ard
Lin
e C
ard
Lin
e C
ard
AC
TIV
E
STA
ND
BY
Failure
AC
TIV
E
Periodic PIM Joins
GENID PIM Hello
Triggered PIM Joins
Multicast HA for SSM: Triggered PIM Join(s)
Active Route Processor receives periodic PIM Joins in steady-state
Active Route Processor fails
Standby Route Processor takes over
PIM Hello with GENID is sent out
Triggers adjacent PIM neighbors to resend PIM Joins refreshing state of distribution tree(s) preventing them from timing out
How Triggered PIM Join(s)
Work When Active Route
Processor Fails:
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 67
Multi-Topology (MT)-Technologyand IP Multicast
… When not all traffic should flow on the same paths
Interdomain: Incongruent routing
BGP SAFI2 (MBGP)
Intradomain: Incongruent routing workarounds
Static mroutes
Multiple IGP processes (tricky)
Do Not Use: DVMRP—in process of being removed from Cisco IOS
Intradomain: Multi-Topology-Routing
Multicast and Unicast solution in Cisco IOS (c7600); multiple topologies for unicast, one additional for multicast (today)
Intradomain: MT-technology for multicast
Subset of MTR: Only the routing component, sufficient for incongruent routing for IP multicast; currently: MT-ISIS in Cisco IOS-XR for IP multicast
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 68
Auto Multicast without explicit Tunnels
Tunnel through non-multicast enabled network segment
draft-ietf-mboned-auto-multicast-09
GRE or UDP encap
Relay uses well known ‗anycast‘ address
Difference to IPSec, L2TPv3, MobileIP, …
Simple and targeted to problem
Consideration for NAT (UDP)
Ease implemented in applications (PC/STB) (UDP)
Variety of target deployment cases
Relay in HAG—provide native multicast in home
Gateway in core-SP—non-multicast Access-SP
Access-SP to Home—non-multicast DSL
In-Home only—e.g. multicast WLAN issues
Non
multicast
Multicast
Capable
AMT Gateway
AMT Relay
AMT Tunnel
Non-
Multicast
HAGNAT
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 69
Video Topology
Multicast Topology
Voice Topology
Cisco IOS IPv4 Multi-Topology Routing (MTR)
Define traffic-class specific topologies across a contiguous subsection of the network
Individual links can belong to multiple topologies
Base Topology
Start with a Base Topology
Includes All Routers and All Links
Full Solution with Both MT-Technology Routing and Forwarding
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 70
Live-Live
Live-Live—Spatial Separation
Two separate paths through network; can engineer manually (or with RSVP-TE P2MP )
Use of two topologies (MTR)
―Naturally‖ diverse/split networks work well (SP cores, likely access networks too), especially with ECMP
Target to provide ―zero loss‖ by merging copies based on sequence number
Live-Live—Temporal Separation
In application device—delay one copy—need to know maximum network outage
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 71
What Is Live-live (with Path Diversity)?
Why bother?
Only resiliency solution in the network that that can be driven to provide zero packet loss under any single failure in the network—without introducing more than path propagation delay (latency)!
Much more interesting for multicast than unicast
Individual unicast packet flow typically for just one receiver
Individual multicast flow (superbowl) for N(large) receivers!
Path diversity in the network
Lots of alternatives: VRF-lite, routing tricks, RSVP-TE, L2 VLAN
Multi-topology routing considered most simple/flexible approach!
Standard solution in finance market data networks
Legacy: Path diversity through use of two networks!
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 72
Cable Industry Example
Path separation does not necessarily mean separate parts of network!
Carrying copies counterclockwise in rings allows single ring redundancy to provide live-live guarantee; less expensive network
Target in cable industry (previously used non-IP SONET rings!)
IP live-live not necessarily end-to-end (STB), but towards Edge-QAM (RH*)—merging traffic for non-IP delivery over digital cable
With path separation in IP network and per-packet merge in those devices solution can target zero packet loss instead of just sub 50msec
STBs
STBs
HFC1
HFC2RH1b
RH1b
RH1a
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 73
cFRRPIM/mLDP Break before Make
RPF change on C from A to C:
1.Receive RPF change from IGP
2.Send prunes to A
3.Change RPF to B
4.Send joins to B
Same methodology, different terminology in mLDP
RPF == ingres label binding
Some more details (not discussed)
A B
S(ource)
Cost: 10
C
Cost: 12
R(eceiver)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 74
cFRRPIM/mLDP Make before Break
1. Receive RPF change from unicast
2. Send joins to A
3. Wait for right time to go to 4.
Until upstream is forwarding traffic
4. Change RPF to A
5. Send prunes to B
Should only do Make-before-Break when old path (B) is known to still forward traffic after 1.
Path via B failed but protected
Path to A better, recovered
Not: path via B fails, unprotected
Make before Break could cause more interruption than Break before Make !
A B
S(ource)
Cost: 10
C
Cost: 12
R(eceiver)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 75
Multipath for IP Multicast
In unicast, multipath selection happens during packet forwarding
In multicast, multipath selection happens during RPF-selection for PIM join!
Multipath selection happening whenever route from RPF-lookup has more than one path (like from IGP equal cost multipath)
Also needs to be enabled
Source
(S)
Receiver (D)
IP Packet
R1
R2 R3
?
PIM Join
e.g. (S,G)
RPF Selection
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 76
Cisco IOS IPv4 Multipath (1)
Classic Cisco IOS IPv4 multicast: ~ Cisco IOS 11.0 (or earlier)
ip multicast multipath
Disabled by default: RPF-selection selects the highest IP address PIM neighbor amongst the nexthop of the available (equal cost) paths
RFC-2362 compliant (old PIM-SM RFC); new PIM-SM RFC-4601 does not specify multipath behavior anymore
If enabled: per-source (and per-RP for (*,G)) multipath:
multipath_classic(A.B.C.D, Npaths) = (A ^ B ^ C ^ D) % Npaths
A.B.C.D is the IP address of the source or RP
Npaths is the number of paths available
Result is the path to choose 0..(Npaths-1)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 77
Cisco IOS IPv4 Multipath (2)
Need to enable consistently on all (downstream) routers in LANs to avoid inconsistent RPF-selection on LAN
Inconsistent selection would cause asserts on LANs, resulting in some duplicate/delayed packets!
Algorithm is polarizing
Can be good and bad
Algorithm does not include group-address:
Consider a single Video server is sending many video flows via SSM: (S,G1) … (S,G100); no load-splitting will happen because source is always the same
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 78
Cisco IOS IPv4 Per (S,G) Improvements for ECMP
Added two per (S,G) ECMP alternatives to IPv4 IP multicastip multicast multipath [ s-g-hash [ basic | next-hop-based]]
Basic: polarizing/predictable—but per (S,G):
(S XOR G % Nlinks)
Next-hop-based: stable/non-polarizingHash(S,G, Nbr-i) = bsr_hash(bsr_hash(S,G), Nbr-i ))
Select Nbr-i | max({ Hash(S,G,Nbr-i) | NBr-i })
Nbr-i is the IP address of the next-hop of a path; Bsr_hash is the hash function also used in the BSR protocol in PIM (creates random number out of its two parameters)
Algorithm select the one neighbor for which the Hash(S,G,Nbr-i) is highest
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 79
Next-Hop Load-Split Algorithm
Needed to have non-polarizing algorithm and non-assert-causing!
Router-local hash to cause non-polarization would cause assert issue!
Also would like stability under re-convergence:
Re-convergence causes interruption! More in multicast than unicast; when loosing/adding an ECMP path, traffic on unaffected paths should not need to re-converge!
Polarizing algorithm is not-stable: change in number of ECMP path changes ―modulo‖ of algorithm, reshuffling large percentage of flows unnecessarily!
Hash algorithm taken from BSR/RFC, better than XOR for this purpose
R4
R1 R2 R3
If Link to R1 fails, R4
Re-Converges Both Red
Trees toward R2 and R3
without Affecting the
Orange and Blue Trees that
Already Used those Two
Next Hops
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 80
Admission Control
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 81
Static vs. dynamic trees
1. ―Broadcast Video‖Dynamic IGMP forward up to DSLAM
DSL link can only carry required program!
static forwarding into DSLAM
Fear of join latency
History (ATM-DSLAM)
2. ―Switched Digital Video‖Allow oversubscription of PE-
AGG/DSLAM link
3. ―Real Multicast‖dynamic tree building full path
Source
Home
Gateway
DSLAM
PE-AGGSta
tic (
PIM
) tr
ee
(1)
Sta
tic (
PIM
) tr
ee
(2)
IGM
P jo
ins
PIM
jo
ins
(3)
IGM
P jo
ins
IGM
P jo
ins
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 82
Switched Digital VideoWhy oversubscription of access links makes sense
Switched Digital VideoConsider 500…1000 users on DSLAM
Consider 300 available TV programs
Monitor customer behavior – what is being watched ?Example (derived from actual MSO measurements)
Some 50 TV programs almost always watched (big channels)
Out of remaining 220 TV programs never than ¼ watched
Never need more bandwidth than ~ 125 channels!
Dynamic joining towards core ?Todays offered content << #users aggregated -> worst case traffic will always flow.
More a provisioning issue – and when content expands well beyond current cable-TV models
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 83
Admission control
Congestion must be avoided
Inelastic: TV traffic can not throttle upon congestion
One flow too many disturbs all flows
Need to do per TV-flow admission control
Router-links
Router local CLI solution
Strategic solution: RSVP
Already used for unicast VoD
Can only share bandwidth between unicast and multicast with RSVP
Broadband access (DSL link, Cable)
Issues with L2 equipment (eg: DSLAM)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 84
3. Fair sharing of bandwidth
1GE
Multicast Call Admission Control
250-500 users
per DLAM
DSLAM
DSLAM
DSLAM
Example CAC use:
4. 250 Mbps for each CP
250 Mbps Internet/etc
1. Three CPs
10GE
Content
Provider 1
Content
Provider 2
Content
Provider 3
Content
Providers
Service
ProviderPaying
Customers
2. Different BW:
- MPEG2 SDTV: 4 Mbps
- MPEG2 HDTV: 18 Mbps
- MPEG4 SDTV: 1.6 Mbps
- MPEG4 HDTV: 6 Mbps
MPEG4 SDTV
MPEG2 SDTV
MPEG4 SDTVMPEG2 HDTV
MPEG4 SDTV
MPEG2 SDTV
MPEG4 SDTVMPEG2 HDTV
MPEG4 SDTV
MPEG2 SDTV
MPEG4 SDTVMPEG2 HDTV
5. Simply add global costs
PE
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 85
Broadband link access, admission control
DSL
link
BRAS No IGMP snooping (replication) on DSLAM
PE-AGG access/admission control on PE-AGG link affects only single subscriber == equivalent to do access/admission control on DSL link.
Or BRAS (if traffic not native but via PPPoE tunnel
IGMP snooping on DSLAM
PE-AGG stopping multicast traffic on PE-AGG link will affect all subscriber. Only DSLAM can control DSL link multicast traffic
IP Multicast extensions to ANCP(Access Node Control Protocol)
Work in IETF
In IGMP snooping on DSLAM, before forwarding, request authorization from ANCP server.
Allow ANCP server to download access control list to DSLAM.
Similar model as defined in DOCSIS 3.0
CMTS controls CM
ANCP
PE-AGG
DSLAM
PE-AGG
link
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 86
Channel Changing
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 87
Join Latency
Static forwarding (to PE-AGG, or DSLAM) To avoid join latency
Sometimes other reasons too (policy, …)
Bogus ?Hop-by-hop Join latency (PIM/IGMP) very low, eg: individual < 100 msec …
Joins stop at first router/switch in tree that already forwards tree
Probability for joins to go beyond PE-AGG very low !If you zap to a channel and it takes ¼ sec more: You are the first guy watching this channel in a vicinity of eg: 50,000 people. Are you sure you want to watch this lame program ?
ImportantTotal channel zapping performance of system – Primetime TV full hour or (often synchronized) commercial breaks.
Join latency during bursts might be worse than on average. (DSLAM performance)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 88
IGMPv2 leave latencyObsolete problem
Congesting issues due to IGMPv2 leave latency when only admission control mechanism is:
DSL link fits only N TV programs …and subscriber can only have N STB.
Example:
4Mbps DSL link, 3.5 Mbps MPEG2
Can only receive one TV channel at a time
Leave latency on channel change complex (triggers IGMP queries from router/DSLAM) and long (spec default: 2 seconds)
Resolved with IGMPv3/MLDv2
Ability for explicit tracking (vendor specific)
Can immediately stop forwarding upon leaves
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 89
Channel ChangingGOP size and channel changing
GOP size of N seconds causes channel change latency USER EXPERIENCE>= N seconds
Can not start decoding before next I-frame
Need/should-have channel change acceleration for GOP sizes > 0.5 sec ?
Many codec dependencies:
How much bandwidth is saved in different codecs by raising GOP size but keep the quality.
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 90
Video Quality Experience
Three functions (currently): Video Quality monitoring, FEC/ARQ support for DSL links, Fast Channel change
Uses standards RTP/RTCP, FEC extensions.
Fast channel channel by RTCP ―retransmission‖ triggered resend of missing GOP packets from VQE (cached on VQE).
STBHome
GatewayDSLAM
PE-AGG
Core Distribution
regional
Aggregation Home NetAccess
ASERVER VQE
multicast
Unicast/
(multicast)
control
VQE Client
library
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 91
Receiver Tracking
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 92
NewIGMP Explicit Tracking
Displays receiver IP address and interface for IGMPv2 hosts
Available on 6500 in 12.2(33)SXH
es1-7606-c3#show ip igmp snooping explicit-tracking vlan 301
Source/Group Interface Reporter Filter_mode
---------------------------------------------------------------------------
0.0.0.0/224.0.1.39 Vl301: 126.1.99.15 EXCLUDE
0.0.0.0/239.254.1.1 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.4 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.3 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.2 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.5 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.2 Vl301:Gi3/2 126.1.99.41 EXCLUDE
0.0.0.0/239.254.1.5 Vl301:Gi3/2 126.1.99.41 EXCLUDE
0.0.0.0/239.254.1.3 Vl301:Gi3/2 126.1.99.41 EXCLUDE
0.0.0.0/239.254.1.4 Vl301:Gi3/2 126.1.99.41 EXCLUDE
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 93
NewIGMP Explicit Tracking
Displays receiver IP address and interface for IGMPv2 hosts
Available on 6500 in 12.2(33)SXH
es1-7606-c3#show ip igmp snooping explicit-tracking vlan 301
Source/Group Interface Reporter Filter_mode
---------------------------------------------------------------------------
0.0.0.0/224.0.1.39 Vl301: 126.1.99.15 EXCLUDE
0.0.0.0/239.254.1.1 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.4 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.3 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.2 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.5 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.2 Vl301:Gi3/2 126.1.99.41 EXCLUDE
0.0.0.0/239.254.1.5 Vl301:Gi3/2 126.1.99.41 EXCLUDE
0.0.0.0/239.254.1.3 Vl301:Gi3/2 126.1.99.41 EXCLUDE
0.0.0.0/239.254.1.4 Vl301:Gi3/2 126.1.99.41 EXCLUDE
Host IP Addr
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 94
NewIGMP Explicit Tracking
Displays receiver IP address and interface for IGMPv2 hosts
Available on 6500 in 12.2(33)SXH
es1-7606-c3#show ip igmp snooping explicit-tracking vlan 301
Source/Group Interface Reporter Filter_mode
---------------------------------------------------------------------------
0.0.0.0/224.0.1.39 Vl301: 126.1.99.15 EXCLUDE
0.0.0.0/239.254.1.1 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.4 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.3 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.2 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.5 Vl301:Gi3/1 126.1.99.36 EXCLUDE
0.0.0.0/239.254.1.2 Vl301:Gi3/2 126.1.99.41 EXCLUDE
0.0.0.0/239.254.1.5 Vl301:Gi3/2 126.1.99.41 EXCLUDE
0.0.0.0/239.254.1.3 Vl301:Gi3/2 126.1.99.41 EXCLUDE
0.0.0.0/239.254.1.4 Vl301:Gi3/2 126.1.99.41 EXCLUDE
VLAN 301 Physical Int/Switchport
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 95
NewUser Subscriber Rates
Displays total traffic subscribed from one user
Combines information from:
show ip igmp snooping explicit-tracking vlan [vlan]
show ip mroute [group] active
Available on 6500 in 12.2(33)SXI
es1-7606-c3#show ip igmp snooping subscriber-rate 126.1.99.37
0.0.0.0/239.254.1.2 Vl301:Gi3/1 126.1.99.37 996 pps/366 kbps (1 sec)
0.0.0.0/239.254.1.1 Vl301:Gi3/1 126.1.99.37 996 pps/366 kbps (1 sec)
0.0.0.0/239.254.1.5 Vl301:Gi3/1 126.1.99.37 1000 pps/368 kbps (1 sec)
0.0.0.0/239.254.1.4 Vl301:Gi3/1 126.1.99.37 1000 pps/368 kbps (1 sec)
0.0.0.0/239.254.1.3 Vl301:Gi3/1 126.1.99.37 1000 pps/368 kbps (1 sec)
--------------------------
Total = 4992 pps/1836 kbps (1 sec)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 96
Multicast in 802.11
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 97
I can‘t do my job without wireless. It has to work.
―‖
Wireless is best-effort. I can‘t support a level 1 SLA.
―‖
Pull Toward Business Mobility
IT Lacks RF Resources and ExpertiseVS
ContinuedGrowth and Reliance
on Wi-Fi Devices
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 98
Multicast in 802.11
Driven by applications: iTunes streaming, IP/TV, Vocera, WebEX, etc.
Challenges:
Variable data rate
Frames admit increased loss due to interferences, collisions
Data distribution delayed as APs buffer multicast
Multicast unreliability
Multicast packets (UDP) are sent as broadcast packets
Do not use error correction: ―fire and forget‖
Sent at highest supported mandatory data rate: 11 MB for B/G, 24 MB for A
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 99
Background on Video and WiFi Multicast
Streaming video requirements
•Many video codecs such as MPEG-2 are intolerant of packet loss
•Loss of one packet impacts multiple video frames
•Since many frames are ―incremental‖ from previous frames
ex., MPEG-2 requires a PLR of < 0.5% or serious impairments occur
Native WiFi multicast is not a reliable service
•In WiFi, it is normal to see a Packet Error Rate (PER) of 1 or 2%
•For unicast, this is normally corrected by the use of ACKs.
If no ACK received, then re-send
•But for multicast there are no ACKs
Packet Loss Rate (PLR) = PER = 1 or 2% (or higher)
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 100
Physical Layer Enhancements
Performance parity with 100 Mbps fast Ethernet
Improved reliability
Backward compatibility with A/B/G
Improved immunity to noise
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 101
MAC Layer Enhancement: MC2UC
Application:
Broadcast video over WiFi at Hotspots
Issue:
Broadcast video is multicast on IP network
But multicast over WiFi is not reliable
Leads to poor video quality
Multicast to Unicast Solution:
Snoop IGMP request for video
Convert video to unicast on WiFi last hop
Transparent to the client
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 102
MAC Layer Enhancement: MC2UC
Multicast source
Wirednetwork
Multicast converted to unicast
WLC
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 103
VideoStream Multicast Delivery Solution
1
2
5.5
6
9
11
12
18
24
36
48
54
M0
M1
...
M14
M15
802.11
Data Rates
B/G
N
Video
Server
AP 1140
• IGMP state monitored for each client. Only
send video to clients requesting
• Multicast packets replicated at AP and sent to
individual clients at their data-rate
• Resource Reservation Control (RRC) used to
prevent channel oversubscription. Works in
conjunction with Voice CAC
• Stream Prioritization ensures important
videos take precedence over others
• SAP/SNMP error message created when
Channel Subscription violated
Technical Solution
Smooth, Reliable Video
• Video delivered reliably at
802.11n data rates
• Quality of Video protected in
varying channel load
conditions
• Prevents video flooding
• Prioritizes Business Video
over other video
Video Impact
Default 802.11B/G
mandatory data rates
Intelligence
in the AP
QoS Marked
on CAPWAP
From WLC
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 104
VideoStream Delivery Solution
Stream Prioritization • Identify specific Video Streams for preferential
QoS treatment
Resource Reservation Control
(RRC)
• Quality of Video Enforcement by denying client
when channel busy
• Video Bandwidth protection to prevent video from
consuming wifi channel
Multicast Direct • Sends multicast video stream as unicast directly
to client
• Video QoS promotion
• Enables use of 11n data rates and standards
packet error correction
Monitoring • Client alert for insufficient bandwidth
• SNMP trap for QoS/bandwidth problem
Roaming Support (existing)• Roaming with pre-built multicast flows
• Proxy IGMP join (cross controller roam)
IGMP snooping (existing)• Prevents video flooding
Feature Overview
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 105
Network Layer Enhancement
Improved multicast performance over wireless networks
Multicast packet replication occurs only at points in the network where it is required, saving wired network bandwidth
One Multicast Packet InCAPWAP Tunnels
One Multicast Packet InCAPWAP
Multicast Group
One CAPWAP Multicast
Packet Out
Three CAPWAP Unicast
Packets Out
Unicast Mechanism
Multicast Mechanism
Network Replicates
Packet as Needed
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 106
Multicast Mode Selection
Multicast mode and multicast group configured on WLC general interface
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 107
Questions ?
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 108
BRKIPM-2008 Recommended Reading
© 2010 Cisco and/or its affiliates. All rights reserved. Cisco PublicBRKIPM-2008 109
We value your feedback - don't forget to complete your online session evaluations after each session. Complete 4 session evaluations & the Overall Conference Evaluation (available from Thursday) to receive your Cisco Networkers 20th Anniversary t-shirt.
All surveys can be found on our onsite portal and mobile website: www.ciscoliveeurope.com/connect/mobi/login.ww
You can also access our mobile site and complete your evaluation from your mobile phone:
1. Scan the Access Code(See http://tinyurl.com/qrmelist for software,
alternatively type in the access URL)
2. Login
3. Complete and Submit the evaluation
Please complete your Session Survey