Cisco IE3000: Resilient Redundancy Protocol

IE3000 is a din-rail mount switch with eight fast ethernet copper ports and two gigabit ethernet combo ports. People who worked for datacentre, enterprise, and ISP might not have seen or used this kind of switch before.

The IE3000 is designed for used in Industrial network, most of the control systems like PLCs are mounted on rails within a cabinet the switches that use to interconnect the PLCs are hence designed this way as shown above.

As shown you need an AC-DC power converter to power up these switches, on the far left is the AC-DC power supply, the input jacks are very different from those rack-mount switches which most of the network engineers used, that’s right you need to use a screw driver and insert the power cables.

Industrial switches are built for harsh environment such as higher shock and vibration tolerance, higher operating temperature which sometime can be up to 60 deg Celcius or 75 deg Celcius. For some highly hazardous environment like oil and gas industry Cisco IE3000 also has zone 2 certified chassis.

The above showed the power supply fully clipped to the side of the IE3000, at the rear of the chassis are mechanically movable plastic clips for attaching to the rail.

Resilient Ethernet Protocol aka REP is Cisco’s proprietary protocol designed solely for ring topology. This protocol is included for industrial network switches and metro ethernet switches that Cisco is selling. Almost every industrial network is a ring or multiple ring architecture, conventional design methodology from Cisco certification curriculum focus mainly on mesh architecture and ring architecture is not mentioned at all.

Traditionally industrial network was not using TCP/IP, the traditional industrial network used several kinds of serial protocol which traversed over a ring, then came token ring and FDDI for dual ring redundancy.

REP has a recovery time of 50ms on fibre which my friend and I had tested, on copper using Cat5E the recovery time is not deterministic the range is between 390ms and 790ms. A note is REP is only for ring network and not for mesh, for mesh you should use RSTP or MSTP.

Each ring network running REP is a segment represented by a segment ID, REP is compatible with RSTP which RSTP bpdu can also traverse in REP enabled ring network. Each segment has one switch that has an edge primary port which is forwarding data and edge secondary port which is discarding data.

Switch#sh run int g1/1
Building configuration…

Current configuration : 109 bytes
!
interface GigabitEthernet1/1
switchport mode trunk
rep segment 1 edge primary
rep preempt delay 30
end
Switch#sh run int g1/2
Building configuration…

Current configuration : 101 bytes
!
interface GigabitEthernet1/2
switchport mode trunk
rep segment 1 edge
rep preempt delay 30
end

Switch A’s configuration as above. REP preempt delay is not necessary, by default there is no preemption delay, if a link fails, the secondary edge port will be enabled and forward data.

Switch#sh run int g1/1
Building configuration…

Current configuration : 96 bytes
!
interface GigabitEthernet1/1
switchport mode trunk
rep segment 1
rep preempt delay 30
end

Switch#sh run int g1/2
Building configuration…

Current configuration : 96 bytes
!
interface GigabitEthernet1/2
switchport mode trunk
rep segment 1
rep preempt delay 30
end

Switch B’s Configuration as above. The rest of the switches within segment id 1 will not need edge ports.