High Availability and LAN Redundancy – Chapter Summary

The following section is a summary of the major points you should be aware of in this chapter.

Understanding High Availability

• The term ‘High Availability’ is considered a very broad and general term
• Availability is calculated mathematically and is often expressed in percentages
• These percentages are calculated from failure rates and repair times
• Reliability and serviceability affect availability

Understanding Reliability

• Reliability measures the ability of a system or a solution to function without interruptions
• The most common measurement of reliability is MTBF, or Mean Time Before Failure
• Factors that influence reliability include the following:

1. Redundancy
2. Technology

Understanding Serviceability

• Serviceability refers to the time it takes to restore a system to service following a failure
• A commonly used serviceability metric is the Mean Time To Repair (MTTR)
• Serviceability is facilitated by several factors, which include the following:

1. Network support and engineering staff
2. Event logging and network monitoring tools
3. Documentation
4. Policies, Processes, and Controls

The Different Types of High Availability

• The different types of High Availability include the following:

1. Service-Level High Availability
2. System-Level High Availability
3. Network-Level High Availability

Hot Standby Router Protocol

• HSRP is a Cisco-proprietary First Hop Redundancy Protocol (FHRP)
• Two versions are HSRP are supported in Cisco IOS software: versions 1 and 2
• HRSP version 1 is the default HSRP version
• HSRP version 1 restricts the number of configurable HSRP groups to 255
• HSRP version 1 sends updates to Multicast group address 224.0.0.2 using UDP port 1985
• HSRP version 2 uses the new Multicast address 224.0.0.102
• The version 2 packet format uses a Type/Length/Value (TLV) format
• HSRP version 2 packets are ignored by gateways running version 1
• HSRP version 1 is not capable of advertising or learning millisecond timers; version 2 is
• HSRP version 2 numbers have been extended from 0 to 4095
• HSRP version 2 includes a 6-byte Identifier field that contains the router MAC address
• HSRP version 1 uses the MAC range 0000.0C07.ACxx
• HSRP version 2 uses the MAC range 0000.0C9F.F000 to 0000.0C9F.FFFF
• The default HSRP gateway priority is 100; the range is 1 – 255
• HSRP routers exchange three types of messages:

1. Hello Messages
2. Coup Messages
3. Resign Messages

By default, preemption is disabled for HSRP
HSRP interfaces transition through several states, which are:

1. Disabled
2. Init
3. Listen
4. Speak
5. Standby
6. Active

• HSRP uses a default plain-text authentication password of ‘cisco’
• HSRP supports plain text and MD5 authentication
• MD5 authentication can be configured with or without key chains
• HSRP supports interface tracking configuration
• Multiple HSRP groups can be configured on the gateway for load balancing

Virtual Router Redundancy Protocol

• VRRP is an open standard First Hop Redundancy Protocol, similar to HSRP
• VRRP is defined in RFC 2338, which was made obsolete by RFC 3768
• VRRP sends advertisements to the Multicast destination address 224.0.0.18
• VRRP uses IP protocol number 112
• VRRP uses MAC addresses in the range 00-00-5e-00-01xx
• VRRP elects a virtual router master and virtual router backup
• You configure up to 255 virtual routers on an interface
• The number of supported virtual routers that can be configured depends on:

1. Router processing capability
2. Router memory capability
3. Router interface support of multiple MAC addresses

• The default VRRP priority value is 100; the valid range is 1 – 254
• By default, preemption is enabled for VRRP
• The default VRRP version is version 2; there is no version 1
• VRRP version 3 is still in the draft stage
• The virtual router master sends advertisements to other routers in the same group
• Like HSRP, VRRP supports both plain text and MD5 authentication

Gateway Load Balancing Protocol

• GLBP allows multiple gateways in the same GLBP group to actively forward traffic
• GLBP gateways communicate via Hello messages that are sent every 3 seconds
• GLBP Hello messages are sent to the Multicast address 224.0.0.102, using UDP port 3222
• GLBP group members elect one gateway to be the AVG for that group
• The other gateways in the GLBP group provide backup for the AVG in case it fails
• The AVG answers all ARP requests for the virtual router address
• In addition, the AVG assigns a virtual MAC address to each member of the GLBP group
• Each gateway is an AVF for the virtual MAC address it has been assigned
• A GLBP group allows up to four virtual MAC addresses to be used per group
• A primary virtual forwarder is assigned a virtual MAC address by the AVG
• A secondary virtual forwarder is one that has learned the virtual MAC address
• GLBP uses two timers to migrate away from an old forwarder address:

1. The redirect timer
2. The timeout timer

• By default, GLBP preemption is disabled; however, this feature can be manually enabled
• GLBP uses a weighting scheme to determine the forwarding capacity of each gateway
• By default, each gateway is assigned a default weight of 100
• GLBP supports three different load sharing methods:

1. Host-dependent
2. Round Robin
3. Weighted

• The client cache contains information about hosts using a GLBP group as default gateway
• The maximum number of cache entries that may be stored can be up to 2000
• In production environments, is recommended that this number never exceed 1000
• GLBP supports plain-text and MD5 authentication

ICMP Router Discovery Protocol

• IRDP uses ICMP router advertisements and ICMP router solicitation messages
• IRDP is an alternative gateway discovery method
• IDRP eliminates the need for manual configuration of gateway addresses on network
• IRDP is independent of any specific routing protocol
• By default, ICMP router advertisements are sent out as Broadcast packets
• ICMP router advertisements can also be sent as Multicasts
• Cisco IOS software sends out IRDP advertisements between every 450 and 600 seconds

Supervisor Engine Redundancy

• Cisco Catalyst 4500 and 6500 series switches support redundant Supervisor modules
• The first Supervisor that boots up is referred to as the Primary or Active Supervisor Engine
• The second Supervisor is referred to as the Standby or Redundant Supervisor Engine
• A failover or switchover to the Standby or Redundant Supervisor Engines happens when:

1. The Primary Supervisor Engine fails or crashes
2. The Primary Supervisor Engine is rebooted
3. The administrator forces a manual failover
4. The Primary Supervisor Engine is physically removed

• Cisco IOS software supports three redundancy modes for redundant Supervisor Engines:

1. Route Processor Redundancy (RPR)
2. Route Processor Redundancy Plus (RPR+)
3. Stateful Switchover (SSO)

• With RPR, the Standby Supervisor Engine is only partially booted and initialized
• With RPR, not all switch subsystems on The Redundant Supervisor become operational
• With RPR, clock synchronization occurs between Primary and Backup every 60 seconds
• With RPR, when the Standby Supervisor becomes operational, the following occurs

1. All switching modules are reloaded and powered up again
2. Remaining subsystems on the MSFC are brought up
3. ACLs are reprogrammed into Supervisor Engine hardware

• The RPR failover or switchover process takes generally takes between 2 to 4 minutes
• RPR+ improves on RPR and provides failover generally within 30 to 60 seconds
• With RPR+, the Redundant Supervisor is fully initialized and configured
• With RPR+, although initialized, the Redundant Supervisor is not fully operational
• RPR+ synchronizes user-entered CLI commands incrementally line-by-line
• When failover or switchover occurs with RPR+, the following events occur on the switch:

1. Traffic is disrupted until the Redundant Supervisor Engine completes the takes over
2. The switch maintains any static routes across the switchover
3. The switch does not maintain any dynamic routing protocol information
4. The switch clears the FIB Tables on switchover
5. The switch clears the CAM Tables on switchover
6. State information, such as active TCP sessions, is not maintained on switchover

• SSO is the preferred redundancy mode for Supervisor Engines
• With SSO, the Redundant or Standby Supervisor Engine is fully booted and initialized
• With SSO, Configuration information and data structures are synchronized
• SSO maintains state information between the redundant Supervisor Engines
• Failover or switchover with SSO redundancy generally happens within 0 to 3 seconds
• Administrators can initiate a manual failover to the Standby Supervisor Engine

StackWise Technology

• Cisco Catalyst 3750 series switches support Cisco StackWise technology
• This allows up to nine (9) switches to be combined into a single logical unit
• The switch stack is managed and configured from the master switch
• The stack master is elected upon initialization based on the following criteria:

1. The switch with the highest stack member priority value is elected
2. The switch with the highest hardware and software priority will be elected
3. The switch with non-default configuration is elected
4. The switch with the longest system uptime is elected
5. The switch with the lowest MAC will be elected

• The stack master election is held when one of the following events occurs:

1. When the whole switch stack is reset or rebooted
2. When the stack master is reset or powered off
3. When the stack master is removed from the stack
4. When the stack master switch has failed
5. When switches are added to the existing stack

• Cisco StackWise Technology supports the following High Availability mechanisms:

1. CrossStack Etherchannel technology
2. Equal Cost Paths
3. 1:N Master Redundancy
4. Stacking Cable Resiliency
5. Online Insertion and Removal (OIR)
6. Distributed Layer 2 Forwarding
7. RPR+ for Layer 3 Resiliency

Catalyst Switch Power Redundancy

• Cisco Catalyst 4500 and 6500 series switches support redundant power supplies
• Two power redundancy modes are supported:

1. Combined
2. Redundant

• In combined mode, both switch power supplies are used at the same time by the switch
• In combined mode, the total power load cannot exceed the sum of both supplies
• Combined mode is typically used when the switch has a large amount of PoE modules
• In redundant mode, the switch draws power from both power supplies
• In redundant mode, the switch uses no more power than the capacity of a single supply
• Catalyst 3750 switches do not support internal redundant power supplies
• The RPS 2300 is used, with a UPS, to provide the following for Catalyst 3750 series switches:

1. Internal power supply failures in network devices
2. Failure of an AC circuit (a circuit breaker tripping, for example)
3. Interruption of utility power

Non-Stop Forwarding

• Cisco Non-Stop Forwarding (NSF) works with in conjunction with SSO
• NSF minimizes the amount of time a network is unavailable following a switchover
• NSF is used to ensure the continued forwarding IP packets after switchover
• NSF is supported by BGP, OSPF, EIGRP, IS-IS, and CEF
• NSF allows routing protocols to detect a switchover
• NSF allows routing protocols to recover route information from the NSF-capable peers
• With NSF, routing protocols depend on CEF to continue forwarding packets
• NSF is configured on a per-routing protocol basis

Simple Network Management Protocol

• Simple Network Management Protocol, or SNMP, is a widely used management protocol
• SNMP provides a means to monitor and control network devices
• SNMP is an Application Layer (Layer 7) protocol
• SNMP uses UDP as the Transport protocol, using UDP ports 161 and 162
• An SNMP-managed network consists of three core elements, which are:

1. A Management System
2. SNMP Agents
3. Managed Devices

• The SNMP agent gathers data from the Management Information Base, or MIB
• The MIB is a virtual information storage area for network management information
• MIBs are comprised of two components which are:

1. Managed Objects
2. Object Identifiers

• There are three versions of SNMP, which are SNMP versions 1, 2, and 3
• SNMPv1 is the initial implementation of the SNMP protocol
• SNMPv1 is widely used and is the default version when SNMP is enabled in Cisco IOS
• SNMPv1 supports Get, GetNext, Set, and Trap operations
• SNMPv2 revises the SNMPv1 standard and includes improvements
• SNMPv2 also defines two new operations: GetBulk and Inform
• SNMPv3 provides three additional security services not available in previous versions
• The additional security features provided in SNMPv3 are:

1. Message Integrity
2. Authentication
3. Message Encryption

• Both SNMPv1 and SNMPv2 use a community-based form of security
• SNMPv3 is a security model that is an authentication strategy for a user and the group
• Each  SNMPv3 agent has an engine ID that uniquely identifies the SNMP agent
• SNMP passwords are localized using the SNMP ngine ID of the authoritative engine
• SNMP notifications can be sent by SNMP agents as either Traps or Inform requests

IP Service Level Agreement

• Cisco IOS IP SLA allows you to monitor, analyze, and verify IP service levels
• IP SLA uses active traffic monitoring for measuring network performance
• IP SLA can measure and monitor performance metrics such as jitter, latency, and packet loss
• IP Service Level Agreement is comprised of two components:

1. Source
2. Target

• IP SLA operations can be broadly categorized into the following five functional areas:

1. Availability monitoring
2. Network monitoring
3. Application monitoring
4. Voice monitoring
5. Video monitoring

• IP SLA supports two MIBs, which are CISCO-SYSLOG-MIB and CISCO-RTTMON-MIB
• The CISCO-SYSLOG-MIB is needed only for packet loss, average jitter, or MOS violations

System Logging

• Syslog is a protocol that simply allows a host to send out event notification messages
• A Syslog Daemon or Server is an entity that listens to the Syslog messages
• Syslog uses User Datagram Protocol (UDP) as the underlying transport mechanism
• Syslog messages have a destination port of 514
• Syslog messages cannot exceed 1,024 bytes in size; there is no minimum length
• All Syslog messages have a header, a priority, and a message

Additional HA Monitoring Tools

• Additional HA monitoring tools include:

1. Cisco IOS Embedded Event Manager
2. Cisco Network Analysis Module
3. CiscoWorks LAN Management Solution

• Cisco IOS EEM provides real-time network event detection as well as onboard automation
• EEM scripts can be configured using the CLI or Tcl
• NAM improves network performance visibility to help manage application delivery
• The NAM is supported in Catalyst 6500 series switches
• The CiscoWorks LAN LMS is comprised of software applications
• CiscoWorks LAN LMS applications that can be used for availability monitoring include:

1. Resource Manager Essentials (RME)
2. CiscoWorks Health and Utilization Monitor
3. Device Fault Manager (DFM)
4. Internetwork Performance Monitor (IPM)