Campus Networks

Campus Network Infrastructure-Deployment Guide: HyperEdge Mixed Stack (FastIron Release 8.0.00)

by nshaw on ‎07-11-2013 02:18 PM - edited on ‎04-09-2014 02:02 PM by Community Manager (728 Views)

Contents

 

INTRODUCTION

HyperEdge mixed stacking combines the ICX6610 and ICX6450 into a single logical stack. This utilizes the full L3 functionality of the ICX6610 and the low-cost port density of the ICX6450. With mixed stacking, the ICX6450 can leverage the ICX6610 platform to perform L3 functionality.

Mixed stacking currently supports up to 8 units in various topologies – with a maximum of 2 ICX6610s. It provides better resilience and shorter paths between units. The ICX6610 is a high end device with full L3/L2 unicast/multicast capabilities and 160GbE(two 40GbE and two 4*10GbE) stacking port bandwidth and full routing capability. One 40GbE and one 4*10GbE ports are trunked to form a stacking trunk by default. It has eight 10GbE ports and either 24 or 48 1GbE ports.

The ICX6450 is the high end model of the ICX64XX family. It supports full L2 switching and basic L3 routing capabilities. It has 24 or 48 1GbE data ports and four 10GbE ports for stacking and uplink, two of which require a license to be 10GbE. They are by default Gig ports. The ICX6610 and ICX6450 are available in 24 and 48 ports, with or without POE capability. Currently, the ICX6430 cannot join a mixed stack.

Each HyperEdge mixed stack can be divided into two groups: the backbone (core), and the peripheral (sub-stack). The backbone group consists of ICX6610 devices linked by their 40GbE stacking ports. In each HyperEdge stack, the master and standby are part of the backbone group. They form an ICX stack that can be a linear or ring topology.

The peripheral devices do not have 40GbE stacking ports. Each peripheral unit can link to one or two backbone units because a peripheral unit has one or two stacking port/trunk. Peripheral units can form a sub-stack, and a sub-stack can connect to one or two backbone units. A peripheral unit can never become a master or a standby. Thus, a sub-stack cannot operate if it does not connect to an ICX6610.

 

Figure 1: ICX 6610s form the backbone. ICX6450s can connect to one or two ICX6610 units.

 

02_Introduction.png

 

ICX6450 can never be a master or standby in a mixed stack even if there is not an  ICX6610.

 

Unicast example that uses the ICX6610s L3 capability

Combined with the ICX6610 in a HyperEdge stack, the ICX6450 offers low cost port density solution. L3 packets arriving at the ICX6450 will be forwarded to an ICX6610 by the forwarding database (FDB). Then the ICX6610 router engine routes the packet.

 

03_UnicastExample.png

Figure 2: L3 packets arriving at ICX6450 are L2 forwarded to ICX6610 routing engine.

 

Example: Pinging PC2 from PC1.

The following steps will be performed:

Step 1: PC1 sends a packet destined to 1.1.20.10.

Step 2: The packet’s destination MAC is the stack MAC address because PC1’s gateway is 1.1.10.1.

The packet destination MAC matches the stack FDB at ICX6450-2, and is forward to

ICX6610- 1.

Step 3: The routing on a cascade port is enabled on ICX6610.  There should be a host route of 1.1.20.10 in TCAM. So ICX6610-1 routes the packet to 1.1.20.10.

Step 4: If there is no such host route TCAM in step 3, the packet goes to the CPU. The system sends an ARP request and installs the host route when it receives an ARP reply.

Step 5: The packet is sent back to ICX6450-1 and 2, and to PC2.

 

Stack Formation

There are three methods to form a stack:

1.       Secure Setup: An interactive utility to construct a stack. User can determine who can join and assign IDs.

2.       Configure & Join: Each unit configures “stack enable”, does an election to choose a master, and then physically joined to form a stack. A new command called “stack suggested-id” lets user control the ID assignment. Otherwise, the ID could vary due to timing.

3.       Unit replacement: The master has a reserved (provisional) configuration for every unit. Other units are clean units (meaning that it has no startup configuration in flash). If clean units match the master configuration, the master grabs them into stack.

 

The configure & join method is the most important because the other two methods depend on it. Configure & join operate as follows:

1.       Every unit must have “stack enable” configured on it.

2.       Every unit sends mail containing information about itself and its configuration to its neighbors. When a unit learns about other units, it adds the other unit’s information to its database and to its probe mail.

3.       When a unit finds that mail from both sides match its software database, it means that this unit has received all of topology information. Because every unit has the same database, they should elect the same master. Every unit also runs the same ID assignment algorithm on the same database. The master reloads units that have an assigned ID different from its boot up ID. It reloads other units that are boot up standalone or masters.

4.       When other units reboot, they become members with assigned IDs. Everyone participates in the election, and the same master wins. The stack then detects that no unit needs to be rebooted. The master goes through different learning states to get member information and to build register caches. Finally, the stack is operational with the master controlling most of the member’s activities.

 

System specifications:

  • Backwards compatible. Features supported in the previous releases are still supported.
  • Support up to 8 units (Two ICX6610s and up to six ICX6450s.)
  • Only an ICX6610 can be the Active or Standby controller.
  • Trunk number reduced from 124 to 120.
  • If there is no Active or Standby, ICX6450 fall into an orphan state forever.
  • The ICX6610 always has two trunk stacking ports. ICX6450 may have one or two stacking ports or trunks. (Same as in previous releases)

Core vs. peripheral:

  • All traffic between core units never go through peripheral units except when the core connections are broken.
  • Broadcast/multicast packets flood all core units. Only necessary broadcast/multicast go through peri-ports.
  • Active units does not copy its startup config to peri-units.
  • ICX6450 does not need licenses even if the core is running premium protocols.
  • Peripheral units must be running router code.

Best Practice Recommendations

  • Limit the number of the VLANs of peripheral units to reduce broadcast/multicast packets flooding to peri-units.
  • A ring topology is better than a linear topology.
  • Link a sub-stack to master and standby.
  • Avoid a large sub-stack due to limited 10GbE bandwidth.

 

How to configure a HyperEdge Mixed Stack topology

 

The following is an example on how to configure a mixed stack.

 

04_HowToConfigure.png

 

Step 1:  The following hardware is required to bring up the above topology.

  • Two ICX6610s and six ICX6450 units.
  • Two stacking cables to connect between ICX6610 units.
  • Eight stacking cables to connect ICX6450 units to ICX 6610 units.
  • Power cords for eight units

Step 2:  Connect the power cables to the units and power up all the above units.

Step 3:  Load the 08.0.00 software image on all the units.

Step 4:  Licenses:

Please note that you do not need premium licenses on any switch for mixed stacking to work, but one of the key benefits is theability for advanced features to be distributed from the ICX6610s down to the ICX6450s.

Also, you do not need the premium licenses on the ICX6450 even if the ICX6610 units have the license. If licensing is required, then all ICX6610 units will require the premium/advanced license.

If ICX6450 has premium license and ICX6610 with premium is introduced into a mixed stack,  the ICX6610 licenses will supersede anything on the ICX6450.

If you do plan on utilizing all four of the 10GbE ports on the ICX6450 and all eight ports on the ICX6610, you will need additional licensing. The ICX6450 comes with two 10GbE ports enabled. You will need an additional POD license to activate the remaining two. On the ICX6610, since all ports are 1GbE by default, you can purchase additional POD licenses in multiples of 4 to activate the 10GbE ports.

 

Step 5:  Connect the stack cables between the units (port information included for simplicity)

 

05_Step5.png

 

Step 6: In this step, we will show you how to configure the stack using ‘secure setup’ or ‘manual config’.

 

Secure setup:

Note: Please refer to Appendix C on how to configure mixed stacking with peri-trunks.

 

1.       Configure speed as 10GbE on peri-ports of ICX6610.

In the above example,

1/3/5 and 1/3/6 are peri-ports for first ICX6610.

ICX6610-24F Router(config-if-e10000-1/3/5)#speed-duplex 10g-full

ICX6610-24F Router(config-if-e10000-1/3/6)#speed-duplex 10g-full

1/3/7 and 1/3/8 are peri-ports for second ICX6610.

ICX6610-24F Router(config-if-e10000-1/3/7)#speed-duplex 10g-full

ICX6610-24F Router(config-if-e10000-1/3/8)#speed-duplex 10g-full

 

2.       Perform stack secure setup on active unit.

ICX6450-48P Router(config)#stack enable

Enable stacking. This unit actively participates in stackingstacking is enable. optical monitoring for stacking ports 1/2/1, 1/2/3 is not available.

ICX6610-24F Router#stack secure-setup

ICX6610-24F Router#Discovering the stack topology...

 

Manual config:

1.       Configure the following commands on all the units.

The following individual stacks will be formed initially:

  • ICX6610 two unit stack(core)
  • ICX6450 three unit stack(peripheral1)
  • ICX6450 three unit stack(peripheral2)

Configuration: (Please note that the output will be different due to differing MAC addresses.)

 

ICX6450-48P Router#conf t

ICX6450-48P Router(config)#stack enable

Enable stacking. This unit actively participates in stacking

stacking is enable. optical monitoring for stacking ports 1/2/1, 1/2/3 is not available.

ICX6610-24F Router#stack secure-setup

ICX6610-24F Router#Discovering the stack topology...

Current Discovered Topology - RING

Available UPSTREAM units

Hop(s)  Id      Type          Mac Address

1       2    ICX6610-24     748e.f834.3318

Available DOWNSTREAM units

Hop(s)  Id      Type          Mac Address

1       2    ICX6610-24     748e.f834.3318

No new units found...

Selected Topology:

Active  Id      Type          Mac Address

        1    ICX6610-24F    748e.f834.1ea0

Selected UPSTREAM units

Hop(s)  Id      Type        Mac Address

1       2    ICX6610-24     748e.f834.3318

Selected DOWNSTREAM units

Hop(s)  Id      Type        Mac Address

1       2    ICX6610-24     748e.f834.3318

Do you accept the unit id's (y/n)?: y

ICX6610-24F Router#

ICX6610-24F Router(config)#stack mac 748e.f834.1ea0

ICX6610-24F Router(config)#hitless-failover enable

 

Configure ‘peri-port’ on  ICX6610 units.

 

ICX6610-24F Router(config)#stack uni 1

ICX6610-24F Router(config-unit-1)# peri-port 1/3/5

ICX6610-24F Router(config-unit-1)#  peri-port 1/3/6

ICX6610-24F Router(config-unit-1)#stack uni 2

ICX6610-24F Router(config-unit-2)#  peri-port 2/3/7

ICX6610-24F Router(config-unit-2)#  peri-port 2/3/8

 

2.       Perform stack secure setup on the active units , write memory and then reload the router.

 

ICX6610-24F Router#stack secure-setup

ICX6610-24F Router#wr mem

 

Write startup-config done.

Automatic copy to member units:  2

ICX6610-24F Router#Flash Memory Write (8192 bytes per dot) .

Flash to Flash Done.

 

ICX6610-24F Router#reload

 

Are you sure? (enter 'y' or 'n'): y

 

Step 7:HyperEdge mixed stack topology will be successfully formed after reload.

 

ICX6610-24F Router(config)#sh stack

alone: standalone, D: dynamic config, S: static config

ID   Type         Role    Mac Address    Pri State   Comment                  

1  S ICX6610-24F  active  748e.f834.1ea0 128 local   Ready

2  S ICX6610-24   standby 748e.f834.3318   0 remote  Ready

3  S ICX6450-48P  member  748e.f882.eac0   0 remote  Ready

4  D ICX6450-48P  member  748e.f8d2.8540   0 remote  Ready

5  D ICX6450-48P  member  748e.f8d2.c200   0 remote  Ready

6  D ICX6450-48P  member  748e.f8d2.d5c0   0 remote  Ready

7  D ICX6450-24   member  748e.f882.e4a0   0 remote  Ready

8  D ICX6450-48P  member  748e.f882.ee00   0 remote  Ready

 

    active       standby                                                      

     +---+        +---+                                                       

=2/6| 1 |2/1==2/6| 2 |2/1=                                                   

|   +---+        +---+   |                                                   

|                        |                                                   

|------------------------|                                                  

    active                                              standby               

      ---         +---+        +---+        +---+         ---                 

     ( 1 )3/6--2/1| 4 |2/3--2/1| 5 |2/3--2/3| 6 |2/1--3/8( 2 )                

      ---         +---+        +---+        +---+         ---                

    standby                                             active                

      ---         +---+        +---+        +---+         ---                 

     ( 2 )3/7--2/1| 3 |2/3--2/3| 8 |2/1--2/3| 7 |2/1--3/5( 1 )                

      ---         +---+        +---+        +---+         ---                

Will assign standby in 12 sec due to all ready

Standby u2 - protocols ready, can failover

Current stack management MAC is 748e.f834.1ea0

 

Recommended Use cases

General recommendations:

 

Only ICX6610s form the core of the stack, with current support of 1-2 ICX6610 units. Two ICX6610s can form a full mesh ring or linear topology as part of the core. Connections through the QSFP stacking ports can be in a set of 1, 2 or 4 connections. One connection creates a linear stack, two creates a linear trunk or a simple ring, and four allows for a fully redundant ring. Four connections are recommended; but in the case of cost constraints, two connections can be considered as it allows for some redundancy. More connections between the ICX6610 units are recommended since it allows for greater resiliency in the core, in case of cable or connection failure. Two ICX6610s generally recommended since only core units can be the master and/or standby of the mixed stack.

 

ICX6450s are considered peripheral units and, when combined, can form a sub-stack. Sub-stacks must link to the core. Ideally, peripheral sub-stacks connect to the core units in a ring topology for full redundancy. In a mixed stack, up to 6 ICX6450s are supported. No ICX6450 can be a part of the core since the 10GbE connections will slow down the QSFP connections that should exist between the ICX6610s. By keeping the ICX6610s as the core units and the main brains of the stack, you maintain the processing power and scalability capacity of the ICX6610s.

 

More details on the advantages and disadvantages for each use case will be provided below.

 

Use case 1 (Brocade’s Recommended Topology):

                The following topology is highly recommended since it allows for the greatest network resiliency.

Two ICX6610s form a ring topology as part of the core. Peripheral units(ICX6450) form two sub-stacks. Two peripheral sub-stacks are connected to two ICX6610 units to form two ring topologies. The ICX6610s can have fully redundant connections to each other, but less connections are supported as well. The more connections you have between the two core units, the greater protection your network will have in the case of cable or connection failure. The connections from the ICX6610s to the ICX6450s can be peripheral ports or peripheral trunks (with trunking up to 2 ports.) This use case is the best and maximum use of resources since it fully utilizes all link connections between the ICX6610s and between the ICX6610 and the ICX6450s.

The one disadvantage is that it may become costly as network engineers distribute this topology across the network. Depending on the importance of resiliency, other networks can be considered.

 

06_UseCase1.png

 

Use case 2 (Alternate Recommendation):

In this use case, one ring is formed between the ICX6610s and all ICX6450s. This topology allows for some cost savings on SFP+ cables as cross-connections between the sub-stacks and the other ICX6610s is not needed. This does allow for some failover in case any cable/connection between any unit fails. However, it only allows for 1 possible link failure. Additional link failures can result in the loss of part of the sub-stack.

You may consider this scenario if the peripheral units need to be placed over a longer distance. The ICX6450 units can be horizontally stacked over a greater distance (up to 150 meters) and may not reach back to the core units.

The use cases can have 2 6610 in a ring or line deployment. All the use cases presented can have it etiher way.

 

07_UseCase2.png

Use case 3:

Two ICX6610s form a ring topology as part of the core. Peripheral units(ICX6450) form two sub-stacks. One peripheral sub-stack (‘Sub-stack B’ in the diagram) is connected to one ICX6610 to form a linear topology. The other peripheral sub-stack (‘Sub-stack A’ in the diagram) is connected to two ICX6610 units to form a ring topology.  The ICX6610s can have fully redundant connections to each other, but less connections are supported as well. The more connections you have between the two core units, the greater protection your network will have in the case of cable or connection failure. The connections from the ICX6610s to the ICX6450s can be peripheral ports or peripheral trunks (with trunking up to 2 ports.)This use case, while supported, is not recommended over Use Case 2. It lacks the redundancy and resiliency of the previous case.  If the ICX6610 on the right fails, then the connections and layer 3 capabilities for Sub-stack B fail. Sub-stack B will become and orphan and may function until its tables time out .

 

 

 

08_UseCase3.png

 

 

Use case 4:

Two ICX6610s form a ring topology as part of the core. Peripheral units(ICX6450) form two sub-stacks. One peripheral sub-stack is connected to one ICX6610 to form a ring topology. The other peripheral sub-stack is connected to the other ICX6610 unit to form another ring topology. The ICX6610s can have fully redundant connections to each other, but less connections are supported as well. The more connections you have between the two core units, the greater protection your network will have in the case of cable or connection failure. The connections from the ICX6610s to the ICX6450s can be peripheral ports or peripheral trunks (with trunking up to 2 ports.)This use case is also not recommended over use case 2, if possible. However, if the ICX6610s are stacked horizontally over a longer distance, and the respective sub-stacks are closer to a specific ICX6610, then it might be a good idea to follow this stacking topology. The risk that runs with this topology is that the failure of one ICX6610 can cause the loss of an entire sub-stack.


09_UseCase4.png

Use case 5 (Generally Not Recommended):

Two ICX6610s form a ring topology as part of the core. Peripheral units(ICX6450) form two sub-stacks. One peripheral sub-stack is connected to one ICX6610 to form a linear topology. The other peripheral sub-stack is connected to the other ICX6610 unit to form another linear topology. The ICX6610s can have fully redundant connections to each other, but less connections are supported as well. The more connections you have between the two core units, the greater protection your network will have in the case of cable or connection failure. The connections from the ICX6610s to the ICX6450s can be peripheral ports or peripheral trunks (with trunking up to 2 ports.)If cost savings is a significantly higher priority than resiliency, then this use case can be considered. However, keep in mind that if either ICX6610 fails, then you have the potential to lose an entire sub-stack. If possible, consider use case 4.

 

10_UseCase5.png

 

Use case 6:

Two ICX6610s form a ring topology as part of the core. Peripheral units(ICX6450) form four sub-stacks.The ICX6610s can have fully redundant connections to each other, but less connections are supported as well. The more connections you have between the two core units, the greater protection your network will have in the case of cable or connection failure. The connections from the ICX6610s to the ICX6450s can be peripheral ports or peripheral trunks (with trunking up to 2 ports.)Hanging off of each ICX6610 are 2 sub-stacks. Each sub-stack must be connected to the ICX6610 in a linear fashion, since a ring would result in using up the allotted 2 peri-ports/trunks per ICX6610.If possible, it would be recommended to join the sub-stacks that are under each ICX6610 to form 2 rings if possible. This would allow for redundancy. Currently, in the following image, if the connection between the ICX6610 and any sub-stack fails, the sub-stack becomes an orphan. If the sub-stacks were connected, then a link-failure would not be catastrophic anymore.This topology will most likely become a scenario that occurs when connections between the ICX6450s fails – thereby creating 2 substacks.

 

11_UseCase6.png

 

For the following use cases, only one ICX6610 is used in the core. This is supported, but not recommended due to the possibility of link or unit failure. In the case of a core unit failure, the sub-stacks will become orphan units and will eventually die out. This is very risky in scenarios that require high availability and little tolerance for failure.

 

Use case 7 (Single Core Unit, Limited Recommendation):

One ICX6610 unit will be placed as part of the core. Peripheral units(ICX6450) form a sub-stack. Peripheral sub-stack is connected to one ICX6610 to form a ring topology.While this topology is supported, it is highly not recommended because of the sheer lack of redundancy with respect to the ICX6610s. If the ICX6610 fails, both sub-stacks become orphaned and will lose all premium capabilities that came with having the ICX6610s.

 

12_UseCase7.png

 

Use case 8 (Generally not recommended. If possible, consider Use Case 7):

One ICX6610 unit will be placed as part of the core. Peripheral units(ICX6450) form a sub-stack. One end of Peripheral sub-stack is connected to one ICX6610 to form a linear topology.

Any topology with just one ICX6610 is not recommended without fully considering the loss of redundancy and resiliency. In the case of an ICX6610, the sub-stack(s) will become orphaned and will lose the functionality gained from mixed stacking. If possible, consider using scenario 7 since a link failure will at least prevent sub-stacks from turning into orphan units.

 

13_UseCase8.png

 

 

Use case 9:

One ICX6610 unit will be placed as part of the core. Peripheral units(ICX6450) form two sub-stacks. Two peripheral sub-stacks are connected to one ICX6610 to form a linear topology.

Any topology with just one ICX6610 is not recommended without fully considering the loss of redundancy and resiliency. In the case of an ICX6610, the sub-stack(s) will become orphaned and will lose the functionality gained from mixed stacking.

Please consider using use case 7 so that a link failure will not cause sub-stacks and/or units to become orphans.

 

14_UseCase9.png

 

HyperEdge Mixed Stack Limitations

 

The following are the limitations to form a mixed stack:

  • Supports up to two ICX6610 units
  • Supports up to six ICX6450 units
  • Supports up to two logical peripheral ports per ICX6610 (including trunks)
  • If there is a peripheral trunk from an ICX6610 to an ICX6450, then it should always start from odd number port and the ports should be consecutive.
  • Renumbering between core and peripheral is not supported. Renumbering can be done for units of same type within ICX6610 or within ICX 6450.
  • Switch image is not supported. All units must be running router code. (License not required.)
  • Cross trunking between ICX6610 and ICX6450 is not supported
  • LAGs cannot be formed between core and peripheral units.

Following warning message will be printed on the console for the invalid topologies:

*** Warning! The current Mixed-stack topology is not a validated topology

Please refer to the configuration guide for more details.

There will also be a reason printed along with this warning to help the user to correct the topology as needed.

  • If there are more than two ICX6610 units in the stack:
    Reason: There should not be more than "two" ICX6610 units in the core-stack.
  • If there are more than six ICX6450 units in the stack:
    Reason: There should not be more than "six" ICX6450 units in the peripheral sub-stacks.
  • If there are more than two peripheral links (trunk or port) on any of the ICX6610 units:
    Reason: There should not be more than "two" peripheral-links (Trunk-or-port) on each ICX6610 unit.

Upgrade/Downgrade images

  • Backwards compatible in upgrading to 8.0
  • The ICX6450 has no issues in downgrading the image.
  • The ICX6610 loses ICX6450 config and peri-port/trunk when downgrade is performed.
  • All units are required to run images of the same version and same image type.
  • Stack cannot be formed if mixing 8.0 and previous releases.
  • Unit will be non-operational if the running image has a minor version difference.
  • Both ICX6610s and ICX6450s should run with Router image to form a mixed stack.
  • ‘Manifest file upgrade’ method can be used to upgrade the images in a mixed stack. Please refer to ‘Appendix B’ for the detailed procedure. This method can be used only after forming the a valid topology.

Layer-3 features support

VRF lite has been added in this release on Brocade's SX/FCX/ICX 6610 platforms, and is configurable/supported in the HyperEdge Mixed Stack Environment also. VRF allows multiple instances of a routing table to co-exist within the same router at the same time. Because the routing instances are independent, the same or overlapping IP addresses can be used without conflicting with each other.

Additionally BGP4+ with IPv6 Unicast Address family support is provided in a mixed stack, starting from this release.

 

L3 Considerations/Caveats for HyperEdge Mixed Stacking:

1.     Physical ports can not be configured as part of non-default (user) vrfs. Only VEs can be configured to be part of non-default(user) vrfs.

2.       Maximum number of user vrfs (i.e. non-default vrfs) configurable is 16.

3.       The following L3 routing protocols have been made VRF aware (configurable on a per VRF basis) :

  • OSPFv2
  • RIP
  • IPv4 Static Routes
  • OSPFv3
  • IPv6 static Routes

4.     The following L3 protocols are not VRF aware and consequently are configurable only on the default-vrf

  • BGPv4
  • BGPv4+ with IPv6 Address Family support
  • RIPng

5.     The number of IPv4, IPv6 routes for default-vrf and non-default vrfs are configurable through the system-max commands.

 

Deployment models

HyperEdge mixed stacks can be positioned in different network segments. The following scenarios are verified in four different profiles: data center, secure enterprise, VOIP and Three tier.

1.       ICX6610 (Distribution) and ICX6450 (Edge)

2.       HyperEdge mixed stacks at Distribution level with ICX6610 for L3/L2 and ICX6450 for hosts devices

3.       HyperEdge mixed stacks at Edge level with ICX6610 for L2 uplinks and ICX6450 to host devices

 

Datacenter profile:

The Datacenter profile represents a typical campus environment with a highly-redundant infrastructure for end-to-end Voice/Video/Data traffic.  End-users can communicate and access shared server resources across Autonomous Systems.  Key features include Spanning Tree, Trunking, OSPF, PIM, GRE Tunneling, MSDP, and QoS to selectively prioritize traffic from end to end.   Objective of this profile is to provide real-time network monitoring, possible network solutions, induce real-world network triggers and create realistic stateful End-to-End traffic

The following features are incorporated in this profile:

  • PIM-SM, SSM, IGMP snooping
  • Spanning Tree ( STP ,RSTP, PVST )
  • Security Authentication ( Local , TACACs , Radius , 802.1x )
  • VRRP and VRRP-E, Trunking, Tunneling ( GRE , IPv6-v4), Qos
  • L3 protocols ( IPv4 , IPv6 , OSPF,  RIP and BGP)
  • Rate Limiting, Security ACL

 

15_DC.png

Secure Enterprise profile:

Enterprise campus LAN networks consist of an access, distribution and core layer architecture.  Key features includes a large 10GbE trunk core ring running Brocade proprietary Metro Ring Protocol (MRP) ,  Virtual Switch Redundancy Protocol (VSRP) and RFC standard Virtual Routing Redundancy Protocol (VRRP)  to provide high-bandwidth, sub-second fail over and fast convergence time for Layer 2, Layer 3 and multicast data traffic.   Secure enterprise networks incorporate many security features as well as the latest L3 functionalities for all FastIron product line.

The following features are incorporated in this profile:

  • Port MAC Security,802.1X,MAC Filters,MAC Authentication,MAC Based VLANs
  • DHCP Snooping,Dynamic Arp Inspection,DOS ATTACK,IP Source Guard
  • xSTP/MRP/VSRP
  • ACL v4/v6, VLANs, LAGs (Static/Dynamic)
  • VRRP/VRRP-E
  • OSPF,BGP
  • IGMP/MLD/PIM

16.png

VoIP profile:

Large Enterprise VoIP profile is emulating one campus network which transmits concurrent voice, data and video traffic. Video and voice applications (such as VoIP, IPTV and streaming video) are time sensitive and require high bandwidth.  Impairments found in IP networks, such as delay and packet reordering, can have a significant effect on voice and video traffic.

Networks engineered to deliver triple-play content must prioritize traffic to guarantee adequate bandwidth and latency metrics for high-bandwidth video applications and medium-bandwidth voice applications while accommodating the variable-bandwidth needs of data traffic. Bursts of data traffic, such as file transfers, must not affect voice or video applications.

Evaluate the performance of emerging technologies such as Video on Demand (VOD) and VoIP. Assess the sensitivity of voice, video and data applications to variations in traditional QoS parameters such as bandwidth, packet loss, delay, and jitter. Determine how resource sharing can lead to degradation of signals due to network congestion. Identify implementation (buffering, QoS, bandwidth shaping) to assure quality of experience for the end user. Ensure an acceptable MOS rating through various VoIP testing scenarios.

The following features are incorporated in this profile:

  • Spanning-Tree, Trunking, VLANS
  • Routing, Multicast
  • VSRP, VRRP-E, MRP
  • POE/POE+, Topology groups, GRE tunnel
  • IPV6OSPFv3(IPSEC), IPV6 tunnel, MSDP, VRRP-E IPv6

17_ThreeTier.png

 

Three Tier profile:

Three tier profile represents a typical campus environment with a highly-redundant infrastructure for end-to-end Video/Data traffic.  End-users can communicate and access shared server resources across Autonomous Systems.  Key features include Rapid Spanning Tree, Trunking, MRP, IGMP, OSPF, BGP, PIM, GRE Tunneling, and QoS.

The following features are incorporated in this profile:

  • IPv4/IPv6 ROUTES
  • IPv4 ECMP
  • IPv4/IPv6  ROUTES
  • Routing protocols: OSPF, BGP, RIP
  • PBR, IPv6 OVER IPv4 TUNNEL
  • GRE TUNNEL, VRFs

18_ThreeTier.png

 

SCALABILITY NUMBERS

Following scale numbers are verified with HyperEdge mixed stack.

 

  1. 1. Layer-3 Features:

Layer-3 Feature

Scale Number

VRRP/VRRP-E

254

VRRP-e IPv6

128

IPv4  ROUTES

15000

IPv6  ROUTES

2884

IPv4 STATIC ROUTES

2k

IPv6 STATIC ROUTES

2k

IPv4 ECMP

8

IPv4 OSPF ROUTES

15000

IPv4 OSPF NEIGHBORS

20

IPv4 OSPF AREAS

4

IPv6 OSPFv3 Neighbors with IPSEC

20

IPv4 BGP ROUTES

15000

IPv4 BGP NEIGHBORS

20

BGP RIB-in

15000

BGP RIB-out

15000

IPv4 RIP ROUTES

2800

IPv6 OSPF NEIGHBORS

20

IPv6 OSPF ROUTES

2884

IPv6 RIP ROUTES

2800

IPv6 OVER IPv4 TUNNEL

8

IPv6 Tunnel

8

GRE TUNNELS

16

VRFs

16

 

  1. 2. Layer-1&2 Features:

Layer-1&2 Feature

Scale Number

VLANs

4095

LAGs (Static/Dynamic)

120/64

MSTP (802.1s)

16

RSTP (802.1w)

254

VSRP

254

MRP

128

POE/POE+

48/24

Topology Groups

128

MAC entries

16K

 

  1. 3. Security Features:

Security Feature

Scale Number

Number of MAC Filters per port

256

Number of MAC Filters per system

512

ACL rules per system

8000

IPv6 ACLs per device(TCAM rules)

1500

IPV4 ACLs per device(TCAM rules)

3000

IPv4 Egress ACLs per device

3000

Dot1x sessions per port

250

MAC based VLANs(clients per port)

32

MAC port security per port

64

ACLs

sys max ip-filter-port

3069

sys max ip-filter-sys

8192

  1. No. of rules possible to config in 1 ACL -ID before binding to interface

3068

  1. No. of rules permitted to config in 1 ACL-ID after  binding to interface

3068

Max HW TCAM entries possible in 1 device, including default deny rule

3069

Max Rules Configuration, in overall system

8192

Max rules possible to bind in overall system

8192

Max Hw TCAM entries in overall system (include default deny rules)

8195

 

  1. 4.Multicast Features:

MULTICAST Scale Number

Layer2 Multicast

Layer3 Multicast

IGMP

IGMP

MLD

MLD

IGMP

PIM (S,G) mcache

MLD

PIM6 (S,G) mcache

Groups

mcache

Groups

mcache

Groups

Single VRF

Max VRF

Mcache in Max VRF scenario

Groups

Single VRF

Max VRF

Mcache in Max VRF scenario

8K

8K

8K

8K

8k

6k

16

384

8k

1k

16

64

 

  1. 5. Power-Over-Ethernet Scaling:

Power Over Ethernet Scaling for ICX6610 platform

Platform

PoE Capacity of Power Supply

Goal/Result

ICX6610-24P

748W

24 PoE ports

24 PoE+ ports

ICX6610-48P

748W

48 PoE ports

24 PoE+ ports

Power Over Ethernet Scaling for ICX6450 platform

SKU

Power Supply Combination1

Supported Class of PD

Redundancy

Internal PS

EPS 1

EPS 2

Class 4 PD

Class 3 PD

ICX6450-24P

370W

-

N/A

12

24

No

-

370W

N/A

12

24

No

370W

370W

N/A

12

24

Up to12 Class 4 PD24 Class 3 PD

ICX6450-48P

740W

-

-

24

48

No

-

370W

-

12

24

No

740W

370W

-

37

48

Up to12 Class 4 PD24 Class 3 PD

740W

370W

370W

48

48

Up to24 Class 4 PD48 Class 3 PD

1.       Cumulative watt based on Power Supplies present, if a power supply fails ref cumulative watt in the rows above. Ex: 740W EPS1 and EPS2 present in ICX6450-48P. If a single EPS fails the cumulative watt support is Internal PS + EPS1 row.

 

How to configure a HyperEdge mixed stack topology

In a mixed stack, the ICX6450 units might be part of independent homogenous stacks before joining a ICX6610 stack as peripheral sub-stacks to form a HyperEdge mixed stack.

Form a mixed stack with 3 individual substacks by using manual configuration (securesetup won’t detect the substack).

 

Substack1: 2 Unit ICX6610 homogeneous stack.

 

ICX6610-24 Router#sh stack

alone: standalone, D: dynamic config, S: static config

ID   Type         Role    Mac Address    Pri State   Comment                  

1  S ICX6610-24   active  748e.f894.094a 128 local   Ready

2  D ICX6610-24P  standby 748e.f834.5080   0 remote  Ready

    active       standby                                                      

     +---+        +---+                                                       

=2/6| 1 |2/1==2/1| 2 |2/6=                                                   

|   +---+        +---+   |                                                   

|                        |                                                   

|------------------------|                                                  

Standby u2 - No hitless failover. Reason: hitless-failover not configured

Current stack management MAC is 748e.f894.094a

 

Substack2: 2 Unit ICX6450 homogeneous stack.

 

ICX6450-48P Router#sh stack

alone: standalone, D: dynamic config, S: static config

ID   Type         Role    Mac Address    Pri State   Comment                  

1  S ICX6450-48P  active  748e.f882.eb80 128 local   Ready

2  D ICX6450-48P  standby 748e.f882.f140   0 remote  Ready

    standby      active                                                       

     +---+        +---+                                                       

  2/1| 2 |2/3--2/1| 1 |2/3                                                    

     +---+        +---+                                                      

Standby u2 - No hitless failover. Reason: hitless-failover not configured

Current stack management MAC is 748e.f882.eb80

 

Substack3: 3 Unit ICX6450 homogeneous stack.

 

ICX6450-48P Router#sh stack

alone: standalone, D: dynamic config, S: static config

ID   Type         Role    Mac Address    Pri State   Comment                  

1  S ICX6450-48P  active  748e.f883.1920 128 local   Ready

2  D ICX6450-24   standby 748e.f8b1.2660   0 remote  Ready

3  D ICX6450-24P  member  748e.f8b1.ca80   0 remote  Ready

    active       standby                                                      

     +---+        +---+        +---+                                          

  2/1| 1 |2/3--2/1| 2 |2/3--2/3| 3 |2/1                                       

     +---+        +---+        +---+                                         

Standby u2 - No hitless failover. Reason: hitless-failover not configured

Current stack management MAC is 748e.f883.1920

 

HyperEdge Mixed Stack:

 

A HyperEdge mixed stack is constructed from Two ICX6450 stacks and another two-unit ICX6610 stack. The two ICX6450 stacks joins the ICX6610 stack as peripheral sub-stacks.

ICX6610-24 Router(config-unit-2)#sh stack

alone: standalone, D: dynamic config, S: static config

ID   Type         Role    Mac Address    Pri State   Comment                  

1  S ICX6610-24   active  748e.f894.094a 128 local   Ready

2  S ICX6610-24P  standby 748e.f834.5080   0 remote  Ready

3  D ICX6450-24P  member  748e.f8b1.ca80   0 remote  Ready

4  D ICX6450-24   member  748e.f8b1.2660   0 remote  Ready

5  D ICX6450-48P  member  748e.f883.1920   0 remote  Ready

6  D ICX6450-48P  member  748e.f882.f140   0 remote  Ready

7  D ICX6450-48P  member  748e.f882.eb80   0 remote  Ready

    active       standby                                                      

     +---+        +---+                                                       

=2/6| 1 |2/1==2/1| 2 |2/6=                                                   

|   +---+        +---+   |                                                   

|                        |                                                   

|------------------------|                                                  

    standby                                             active                

      ---         +---+        +---+        +---+         ---                 

     ( 2 )3/7==2/1| 5 |2/3--2/1| 4 |2/3--2/3| 3 |2/1==3/1( 1 )                

      ---         +---+        +---+        +---+         ---                

    standby                                                                   

      ---         +---+        +---+                                          

     ( 2 )3/5==2/1| 6 |2/3--2/1| 7 |2/3                                       

      ---         +---+        +---+                                         

Will assign standby in 19 sec due to all ready

Standby u2 - No hitless failover. Reason: hitless-failover not configured

Current stack management MAC is 748e.f894.094a

 

How to unconfigure a HyperEdge mixed stack topology

 

Step 1:  Unconfigure HyperEdge mixed stacking with the following command

After executing command “stack unconfigure mixed-stack” on the Active controller, the following prompt is shown on the console. The user must respond with ‘y’ (for yes) or ‘n’ (for no).

if the user responds with ‘y’, then:

  • The peripheral ICX6450 units recover their pre-mixed stack configuration and go for a reload.
  • All the peri-port and peri-trunk configuration is removed from the ICX6610 stack.

This command recovers the startup-config (startup configuration) on the ICX6450 units. So, the user is advised to save the running configuration on them before making a HyperEdge, using the command ‘write memory’.

 

ICX6610-24 Router#stack unconfigure mixed-stack [

All the peri-ports/trunks will be removed and all the ICX6450 units will recover

pre-mixed-stacking configuration. Are you sure? (enter 'y' or 'n'): y

Removed peri-ports from configuration: 

Removed peri-trunks from configuration: 1/3/1-to-1/3/2 2/3/5-to-2/3/6 2/3/7-to-2/3/8

ICX6610-24 Router#

ICX6610-24 Router#

ICX6610-24 Router#T=15m43.0: Active unit 1 deletes u3 but keeps its static config.

T=15m43.1: Active unit 1 deletes u4 but keeps its static config.

T=15m43.2: Active unit 1 deletes u5 but keeps its static config.

T=15m43.2: Active unit 1 deletes u6 but keeps its static config.

T=15m43.3: Active unit 1 deletes u7 but keeps its static config.

T=15m43.0: Election, was active, no change, ID=1, pri=128, 2U(1-2), A=u1, nbr#=1 1, reason: u2: stk-po-chg, ,

Step 2:  verify the substacks after unconfigure HyperEdge.

 

 

GLOSSARY

Term

Meaning

HyperEdge

Mixing ICX6610 and ICX6450 in the same stack

core unit

ICX6610 in HyperEdge

peripheral unit or peri-unit

ICX6450

Stacking port/trunk

Stacking port/trunk is used to link to similar type unit. The ICX6610 always has two stacking trunks. The ICX6450 could have one or two stacking ports or trunks.

Peripheral port/trunk

Peri-port/trunk is defined only on the ICX6610 and is used to link to a stacking port/trunk of a peri-unit.

(peri-port/trunk)

The ICX6610 may have up to 8 peri-port/trunk. The ICX6450 has no peri-port/trunk.

Core

Topology formed by ICX6610 in a HyperEdge Domain.

Sub-stack

Topology formed by adjacent ICX6450 in a HyperEdge Domain. It can include one or more peri-units. If two ICX6450s are separated by core units, they belong to different sub-stacks.

Active Controller

The formal name for master.

Clean unit

A unit has no startup config flash. Issuing “erase startup” and reload without “write memory” make a clean unit.

 

APPENDIX A

 

Equal-cost paths load balance

This system uses the shortest path approach to define unicast and multicast paths. All traffic between core units never go through ICX6450 units.

There could be two equal paths to reach a device in a ring topology. The system tries use different path to reach different destinations. The following two diagrams shows these two situations. The granularity is based on device, not on packet header hashing. The path from one device to a specific device remains unchanged once it is defined. If the path consists of trunks, load balance among ports of a trunk is based on packet header hashing.

Unicast packets travel only through necessary units and ports. However, multicast/broadcast packets flood all stacking ports of core units. Peri-port/trunk only allows multicast/broadcast packets of the VLANs associated with its sub-stack ICX6450 ports.

Multicast/broadcast flooding also use the shortest path. In the case of multi-equal-cost-paths, the multicast/broadcast chooses different path from unicast.

 

Broken core (Invalid topology)

If the core units are partitioned into more than one segment, but they are still linked by peri-port/trunk, the stack still works. However, the peri-port/trunk may be saturated by the traffic between cores. The system prints a warning message every 10 minutes to remind the user to correct this situation. You can configure “stack suppress-warning” to suppress the message if the broken core is what users want.

Once the problem is corrected, any traffic between core units does not go through ICX6450 units. The system stops printing warning messages.

19_Invalid.png

In the above example, the system prints the following message. It prints only one message even there are multiple fragmentations.

*** Warning! U2 <--> U1 must go through peri-port/trunk 2/3/4 and 1/3/3.

Stack still works. But please correct this problem.

Traffic always uses shortest path, so traffic between U1 and U2 goes through U3, not U4-U5 because U3 cost is lower than U4 and U5 combined. If two sub-stacks have the same cost, the system may distribute traffic through equal-cost paths. For example, suppose U5 does not exist, then U1 to U2 have two equal cost paths through U3 or U4.

APPENDIX B

Manifest File Upgrade:

The manifest file upgrade feature will simplify the upgrade process into a single cli command. It will use the official release manifest file as an input, perform a multiple download or a copy of images specified in the manifest file and upgrade the active and member units of the system.

This feature will also improve the robustness of the system upgrade process and reduce the likelihood of a having a non-operational unit in the stack due to image-mismatch.

The following command will only accept manifest file in .txt format extension. This command will first install the ICX6610 image to all the ICX6610 units in the stack and then it will install the ICX6450 image to all the ICX6450 units in the stack. After all the relevant images have been installed on to the system, the feature will notify the user that the upgrade is done and prompt to reload the system for the new images to take into effect.

Note: Manifest file upgade will not be supported when using secure copy(SCP).

ICX6610-24P Router#copy tftp system-manifest 10.20.65.49 \08000q058\FI08000q058_Manifest.txt all-images-primary

ICX6610-24P Router#Flash Memory Write (8192 bytes per dot) .

DOWNLOADING MANIFEST FILE   Done.

Load to buffer (8192 bytes per dot)

Automatic copy to member units:  1

................................................................Write to boot flash..

DOWNLOAD OF FCX BOOT  Done.

Flash Memory Write (8192 bytes per dot)

Automatic copy to member units:  1

COPY FCX SIGNATURE TFTP to Flash Done

Flash Memory Write (8192 bytes per dot)

Automatic copy to member units:  1

...................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................

Copy FCX from TFTP to Flash Done.

Flash Memory Write (8192 bytes per dot)

Automatic copy to member units:  3  4  5  6  7  8

.

COPY KX SIGNATURE TFTP to Flash Done

Flash Memory Write (8192 bytes per dot)

Automatic copy to member units:  3  4  5  6  7  8

...............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................PLEASE WAIT. MEMBERS SYNCING IMAGE TO FLASH. DO NOT SWITCH OVER OR POWER DOWN THE UNIT....

Copy ICX from TFTP to Flash Done.

KX BOOT IMAGE COPY IS DONE

  Manifest image download is complete, please reload the system

 

HyperEdge configuration:

ICX6610-24F Router#sh run

Current configuration:

ver 08.0.00q045T7f3

stack unit 1

  module 1 icx6610-24f-sf-port-management-module

  module 2 icx6610-qsfp-10-port-160g-module

  module 3 icx6610-8-port-10g-dual-mode-module

  priority 128

  stack-trunk 1/2/1 to 1/2/2

  stack-trunk 1/2/6 to 1/2/7

  stack-port 1/2/1 1/2/6

  peri-port 1/3/5

  peri-port 1/3/6

stack unit 2

  module 1 icx6610-24-port-management-module

  module 2 icx6610-qsfp-10-port-160g-module

  module 3 icx6610-8-port-10g-dual-mode-module

  stack-trunk 2/2/1 to 2/2/2

  stack-trunk 2/2/6 to 2/2/7

  stack-port 2/2/1 2/2/6

  peri-port 2/3/7

  peri-port 2/3/8

stack unit 3                                                     

  module 1 icx6450-48p-poe-port-management-module

  module 2 icx6450-sfp-plus-4port-40g-module

  stack-port 3/2/1 3/2/3

  connect 1/3/6

  connect 5/2/3

stack unit 4

  module 1 icx6450-24-port-management-module

  module 2 icx6450-sfp-plus-4port-40g-module

  stack-port 4/2/1 4/2/3

  connect 1/3/5

  connect 5/2/1

stack unit 5

  module 1 icx6450-48p-poe-port-management-module

  module 2 icx6450-sfp-plus-4port-40g-module

  stack-port 5/2/1 5/2/3

  connect 3/2/3

  connect 4/2/3

stack unit 6

  module 1 icx6450-48p-poe-port-management-module

  module 2 icx6450-sfp-plus-4port-40g-module

  stack-port 6/2/1 6/2/3

  connect 2/3/7

  connect 7/2/3                                                  

stack unit 7

  module 1 icx6450-48p-poe-port-management-module

  module 2 icx6450-sfp-plus-4port-40g-module

  stack-port 7/2/1 7/2/3

  connect 6/2/3

  connect 8/2/3

stack unit 8

  module 1 icx6450-48p-poe-port-management-module

  module 2 icx6450-sfp-plus-4port-40g-module

  no legacy-inline-power

  stack-port 8/2/1 8/2/3

  connect 2/3/8

  connect 7/2/1

stack enable

stack rconsole-off

stack mac 748e.f834.1ea0

vlan 1 name DEFAULT-VLAN by port

vlan 2 by port

tagged ethe 3/1/48                                              

router-interface ve 2

vlan 3 by port

tagged ethe 3/1/48

router-interface ve 3

vlan 52 by port

tagged ethe 3/1/20

router-interface ve 52

vlan 53 by port

tagged ethe 3/1/20

router-interface ve 53

ip route 10.20.0.0/16 10.20.79.1

ip route 131.1.1.0/24 132.1.1.1

ip route 131.1.2.0/24 132.1.2.1

ip route 131.1.50.0/24 132.1.50.1

hitless-failover enable

interface management 1

no ip dhcp-client enable

ip address 10.20.79.35 255.255.255.128

interface ve 2

ip address 133.1.1.1 255.255.255.0

interface ve 3

ip address 133.1.2.1 255.255.255.0

interface ve 52

ip address 132.1.1.2 255.255.255.0

!                                                                

interface ve 53

ip address 132.1.2.2 255.255.255.0

end

ICX6610-24F Router(config)#sh stack

alone: standalone, D: dynamic config, S: static config

ID   Type         Role    Mac Address    Pri State   Comment                  

1  S ICX6610-24F  active  748e.f834.1ea0 128 local   Ready

2  S ICX6610-24   standby 748e.f834.3318   0 remote  Ready

3  S ICX6450-48P  member  748e.f882.eac0   0 remote  Ready

4  D ICX6450-48P  member  748e.f8d2.8540   0 remote  Ready

5  D ICX6450-48P  member  748e.f8d2.c200   0 remote  Ready

6  D ICX6450-48P  member  748e.f8d2.d5c0   0 remote  Ready

7  D ICX6450-24   member  748e.f882.e4a0   0 remote  Ready

8  D ICX6450-48P  member  748e.f882.ee00   0 remote  Ready

    active       standby                                                      

     +---+        +---+                                                       

=2/6| 1 |2/1==2/6| 2 |2/1=                                                   

|   +---+        +---+   |                                                   

|                        |                                                   

|------------------------|                                                  

    active                                              standby               

      ---         +---+        +---+        +---+         ---                 

     ( 1 )3/6--2/1| 4 |2/3--2/1| 5 |2/3--2/3| 6 |2/1--3/8( 2 )                

      ---         +---+        +---+        +---+         ---                

    standby                                             active                

      ---         +---+        +---+        +---+         ---                 

     ( 2 )3/7--2/1| 3 |2/3--2/3| 8 |2/1--2/3| 7 |2/1--3/5( 1 )                

      ---         +---+        +---+        +---+         ---                

Will assign standby in 12 sec due to all ready

Standby u2 - protocols ready, can failover

Current stack management MAC is 748e.f834.1ea0

 

APPENDIX C

How to Configure HyperEdge with peri-trunks:

The following is an example on how to configure HyperEdge with peri-trunks with secure setup method.

 

20_AppendixC.png

 

Step 1:  The following hardware is required to bring up the above topology.

·        Two ICX6610 and six ICX6450 units.

·        Two stacking cables to connect between ICX6610 units.

·        Twelve stacking cables to connect ICX6450 units to ICX 6610.

·        Power cards for eight units

 

Step 2:  Connect the power cables to the units and power up all the above units

 

Step 3:  Load the 08.0.0.0 software image on all the units.

 

Step 4:  License:

 

Please note that you do not need premium licenses on any switch for HyperEdge to work, but one of the key benefits of HyperEdge is that ability for advanced features to be distributed from the ICX6610 down to the ICX6450.

Also, you do not need the premium licenses on the ICX6450 if the ICX6610 units have the license. If licensing is required, then all ICX6610 units will require the premium/advanced license.

If ICX6450 has premium license and ICX6610 with premium is introduced into HyperEdge,  ICX6610 will supersede ICX6450.

Step5:  Connect the stack cables between the units.

Step6:

Secure setup:

In the above example, four peri-trunks will be formed between ICX6610 and ICX6450.

 

1.       Configure speed as 10GbE on all the peri-ports of ICX6610.

1/3/5&1/3/6, 1/3/7&1/3/8 will be part of peri trunks on first ICX6610.

ICX6610-24F Router(config-if-e10000-1/3/5)#speed-duplex 10g-full

ICX6610-24F Router(config-if-e10000-1/3/6)#speed-duplex 10g-full

ICX6610-24F Router(config-if-e10000-1/3/7)#speed-duplex 10g-full

ICX6610-24F Router(config-if-e10000-1/3/8)#speed-duplex 10g-full

1/3/5&1/3/6, 1/3/7&1/3/8 will be part of peri trunks on second ICX6610.

ICX6610-24F Router(config-if-e10000-1/3/5)#speed-duplex 10g-full

ICX6610-24F Router(config-if-e10000-1/3/6)#speed-duplex 10g-full

ICX6610-24F Router(config-if-e10000-1/3/7)#speed-duplex 10g-full

ICX6610-24F Router(config-if-e10000-1/3/8)#speed-duplex 10g-full

 

2.       Perform stack secure setup on active unit.

ICX6450-48P Router(config)#stack enable

Enable stacking. This unit actively participates in stacking

stacking is enable. optical monitoring for stacking ports 1/2/1, 1/2/3 is not available.

ICX6610-24F Router#stack secure-setup

ICX6610-24F Router#Discovering the stack topology...

 

Step7:HyperEdge mixed stack topology will be successfully formed after reload.