Changes to Default System Behavior
SmartEdge OS for SmartEdge Routers , Release 6.4.1

Contents

1Introduction

2

Changes to Default System Behavior in Release 6.4.1
2.1Changes to Default System Behavior in Release 6.4.1.3
2.2Changes to Default System Behavior in Release 6.4.1.2
2.3Changes to Default System Behavior in Release 6.4.1.1

Glossary

Reference List
Copyright

© Ericsson AB 2010-11. All rights reserved. No part of this document may be reproduced in any form without the written permission of the copyright owner.

Disclaimer

The contents of this document are subject to revision without notice due to continued progress in methodology, design and manufacturing. Ericsson shall have no liability for any error or damage of any kind resulting from the use of this document.

Trademark List
SmartEdge is a registered trademark of Telefonaktiebolaget L M Ericsson.

1   Introduction

This document describes the changes to the default system behavior introduced in Release 6.4.1 of the SmartEdge® OS for SmartEdge routers. For software installation and upgrade instructions, see Reference [1]. For details about the new and enhanced software features and the new hardware in this release, see Reference [2].

For current security notifications, visit the Ericsson E-business portal at https://ebusiness.ericsson.net.

2   Changes to Default System Behavior in Release 6.4.1

2.1   Changes to Default System Behavior in Release 6.4.1.3

This release introduced the following changes to default system behavior.

2.1.1   Maximum CLIPS Sessions Increased

The maximum number of Clientless IP Service Selection (CLIPS) sessions per parent circuit has increased to 32,000 from the previous maximum of 16,000. However, the actual number of CLIPS sessions allowed for a parent circuit may be restricted by the circuit limit of the line card.

2.2   Changes to Default System Behavior in Release 6.4.1.2

This release introduced the following changes to default system behavior.

2.2.1   Timestamps for Core Dump Files

In previous releases, core dump files were not timestamped. Since "last modified" date of the file updates after a switchover, it was challenging to determine the time of the crash that created the core dump file. In this release, core dump files in the /md folder include a timestamp as part of the file name.

2.3   Changes to Default System Behavior in Release 6.4.1.1

This release introduced the following changes to default system behavior.

2.3.1   Changes to the qos rate max Command Under a Hitless Access LAG

Changes to the structure and operation of the scheduling hierarchy for hitless access link groups introduced for the SmartEdge OS Release 6.4.1 have resulted in a change to the behavior of the qos rate maximum command configured at the link group level (that is, in link-group configuration mode). Previously, the resulting collective rate restriction applied only to link-group circuits that had an explicitly-configured or inherited PWFQ policy binding applied. In this release, the collective rate is enforced for all circuits (for example, 802.1Q PVCs, PPPOE sessions, and CLIPS subscribers) configured under the link group, regardless of whether they are subject to a PWFQ policy binding, an MDRR policy binding, or have no queuing policy binding configured or inherited.

The resulting rate enforcement is more intuitive and more consistent with the operation of the qos rate maximum command configured under a physical TM port and under economical access LAG. However, this changes the system behavior for a hitless access link group when some circuits have explicit PWFQ configuration and others are configured with default scheduling. Previously, only the egress traffic of circuits with PWFQ configuration (either explicit or inherited) were counted against the link-group-level rate; now, all the egress traffic of the circuits counts against the rate and is subject to queue tail-drop discard if that rate is exceeded. If you are upgrading an existing hitless access link-group configuration to SEOS 6.4.1.1 or later from an earlier release, the qos rate maximum setting to ensure that it is sufficient for all traffic carried on the link group. With this change, both PVC 1 and PVC 2, are subject to and counted against the rate limit of 500 Mbps enforced for the link group:

link-group main_group access
  encapsulation dot1q
  qos pwfq scheduling physical-port
  qos rate maximum 500000
  dot1q pvc 1 
   qos policy queuing pwfq_policy
  dot1q pvc 2  

Previously, only the traffic of 802.1Q PVC 1 would have been subject to and counted against the rate limit.

2.3.2   Changes to IPsec VPN

IPsec VPN changed in Release 6.4.1.1 in the following ways.

2.3.2.1   Maximum Number of On–Demand Tunnels Increased to 1,024

Up to 1,024 tunnels can be configured for an on-demand tunnel in tunnel configuration mode. In previous releases, the maximum number was 32. The default remains the same, at 8 tunnels.

2.3.2.2   show ike sa statistics tunnel Command Deprecated

The new show tunnel ipsec statistics ike command replaces the show ike sa statistics tunnel command. The new command provides options to:

2.3.2.3   Syntax Changes for debug ike card and debug ipsec card Commands

The card keyword replaces the asp keyword in the syntax for the following two commands in this release:

For both commands, the value you define for the level keyword is now a numeric value instead of a descriptive word.

Table 1    Values Defined for the level Keyword

New Value

Old Value

Meaning

0

emergency

Only emergency events

1

alert

Alert and more-severe events

2

critical

Critical and more-severe events

3

error

Error and more-severe events

4

warning

Warning and more-severe events

5

notice

Notice and more-severe events

6

informational

Informational and more-severe events

7

debug

All events, including debug events

2.3.2.4   L4L7d Module Responsible for Tunnel Processing

The L4L7d module now handles all IPsec tunnel processing. All IPsec tunnel–related interactions with the ISM, RIB, or RCM should be debugged in L4L7d. In previous releases, the tunnelD and tunnel-mgr modules handled IPsec tunnel processing.

2.3.3   BFD Support on PIM Interfaces

BFD is enabled on PIM interfaces and for each neighbor on the interface.

2.3.4   BGP Graceful Restart Support for Labeled Address Families

In previous releases, BGP graceful restart negotiation was not supported for labeled address families. In Release 6.4.1.1, BGP graceful restart negotiation is supported for all IPv4 and IPv6 address families (including labeled address families).

2.3.5   Support for IPv6 Path MTU Negotiation

In previous releases, the SmartEdge router dropped IPv6 packets that were too big for the router to process. In Release 6.4.1.1, all IPv6 packets are processed. You can use the new ipv6 path-mtu-discovery discovery-interval command (in global configuration mode) to globally enable IPv6 PMTU negotiation on the router and configure the timeout value used for aging PMTUs. Enabling IPv6 PMTU triggers the PMTU discovery process when the PMTU discovery timer expires.

2.3.6   BGP Minimum Route Advertisement Interval

Setting the BGP MRAI to 0 (with the advertisement-interval 0 command) may generate a higher number of BGP UPDATE messages and increase BGP CPU 25 usage.

2.3.7   BGP Fast-Reset Interval Enhancement

In previous releases, the fast-reset command interval could be configured in seconds only. In Release 6.4.1.1, the fast-reset command has been updated to include an optional milliseconds keyword. If you include the milliseconds keyword in the fast-reset command string, the interval is interpreted in milliseconds; without the milliseconds keyword, the interval is interpreted in seconds (preserving the backward compatibility of the SmartEdge OS).

2.3.8   SmartEdge BGF Support for IPv4 and IPv6 Payload Traffic

The media address and media port entry size are increased to accommodate both IPv4 and IPv6 addresses (the IPv6 address is 128 bits long, and the IPv4 address is 32 bits long), consuming additional memory on XC and PPA traffic cards.

2.3.9   Changes to DHCPv6 PD Prefix Assignment

In Release 6.3.1, the SmartEdge router supported the configuration of multiple DHCPv6 PD prefixes. In Release 6.4.1, only one DHCPv6 PD prefix is assigned by default unless you use the ipv6 delegated-prefix maximum command to modify the maximum number of DHCPv6 PD prefixes that can be assigned.

In Release 6.3.1, the maximum number of DHCPv6 PD prefixes that could be assigned was 100. In Release 6.4, the maximum number of DHCPv6 PD prefixes that can be assigned is five.

2.3.10   Changes to Traffic Management Support

Table 2 shows the changes to the MDRR priority threshold from the previous release to this release.

Table 2    Changes to the MDRR Priority Threshold

Default Queue Map (PD Priority > Queue)

MDRR Queue

Prior Release

6.4.1 Release

1q

2q

4q

8q

Priority

Strict

WRR

Priority

Strict

WRR

0-7

0

0

0

Queue 0

RT

RT

RT

RT

RT

nRT

 

1-7

1-2

1

Queue 1

RT

nRT

RT

nRT

nRT

nRT

   

4-6

2

Queue 2

RT

nRT

RT

nRT

nRT

nRT

   

7

3

Queue 3

RT

nRT

RT

nRT

nRT

nRT

     

4

Queue 4

RT

nRT

RT

nRT

nRT

nRT

     

5

Queue 5

RT

nRT

RT

nRT

nRT

nRT

     

6

Queue 6

RT

nRT

RT

nRT

nRT

nRT

     

7

Queue 7

RT

nRT

RT

nRT

nRT

nRT

2.3.11   New VSA

Cluster-Partition-ID RADIUS VSA is a new VSA.

2.3.12   MDRR Queues Available for 4-Port 10 Gigabit Ethernet Traffic Card

For the 4-port 10 Gigabit Ethernet traffic card, the maximum number of egress queues available per card for Modified Deficit Round-Robin (MDRR) scheduling, as documented in Reference [3], was 3,920 in previous releases. However, for Release 6.4.1, this number was reduced to 3,600. This value reflects the following possible combination of queues and circuits:

For applications that require a greater number of scheduling queues, Release 6.4.1 introduced PWFQ scheduling support for the 4-port 10 Gigabit Ethernet card, which supports a greater number of queues per card (up to 8 queues per circuit with a limit of 32,000 circuits). For more information, see the section "Hierarchical Scheduling in Virtual-port TM" in Reference [3].


Glossary

BFD
Bidirectional Forwarding Detection
 
BGF
border gateway function
 
BGP
Border Gateway Protocol
 
CLIPS
Clientless IP Service Selection
 
DHCPv6
Dynamic Host Configuration Protocol Version 6
 
IKE
Internet Key Exchange
 
IPsec
Internet Protocol Security
 
IPv4
Internet Protocol Version 4
 
IPv6
Internet Protocol Version 6
 
ISM
Interface and Circuit State Manager
 
MDRR
Modified Deficit Round-Robin
 
MRAI
Minimum Route Advertisement Interval
 
MTU
maximum transmission unit
 
PD
Prefix Delegation
 
PIM
Protocol Independent Multicast
 
PMTU
path maximum transmission unit (PMTU)
 
RCM
Router Configuration Manager
 
RIB
Routing Information Base
 
SA
Security Association
 
VPN
Virtual Private Network
 
VSA
vendor-specific attribute

Reference List

[1] Installing Release 6.4.1
[2] Technical Product Description, SmartEdge OS for SmartEdge Routers, Release 6.4.1
[3] Configuring Queuing and Scheduling