For creating a stack of NS9600 Sensors, you need to take the following factors into consideration:
Ensure that you have the required QSFP28 Direct Attach Copper (DAC) cables to create the stack.
For standalone stack, ports G0/3 and G0/4 can be used as monitoring ports.
In a 3-node, 4-node, or 5-node stack, ports G0/1, G0/2, G0/3, and G0/4 are dedicated stacking fabric ports.
For HA the monitoring ports start at G1/3, in standalone stack it starts at G1/1.
For NS9600 stack licensing, you can combine multiple licenses of different capacity to achieve the throughput required and assign them to the stack. For example, the throughput requirement of NS9600 stack (2-node) is 120 Gbps. While assigning licenses, you can import and assign any combination of licenses (20 + 40 + 60 Gbps) or, (60 + 60 Gbps) to achieve the throughput requirement of 120 Gbps. refer to the table below NS9600 Sensor stack and licensing requirements:
To reach the required total stack throughput, combine multiple capacity licenses (20 Gbps, 40 Gbps, or 60 Gbps). Assign the combined licenses to the stack.
A 2-node stack requires a total throughput capacity of 120 Gbps (for example, 60 + 60 Gbps).
A 3-node stack requires a total throughput capacity of 180 Gbps (for example, 60 + 60 + 60 Gbps).
A 4-node stack requires a total throughput capacity of 240 Gbps (for example, 60 + 60 + 60 + 60 Gbps).
A 5-node stack requires a total throughput capacity of 300 Gbps (for example, 60 + 60 + 60 + 60 + 60 Gbps).
License SKU
Throughput
No of Sensors
Throughput required
NS96X20ECE-AT
MSP-NS96-20-OT
20 Gbps
2, 3, 4, or 5 NS9600 Sensors
120 Gbps (2-node)
180 Gbps (3-node)
240 Gbps (4-node)
300 Gbps (5-node)
NS96X40ECE-AT
MSP-NS96-40-OT
40 Gbps
NS96X60ECE-AT
MSP-NS96-60-OT
60 Gbps
Unsupported features list:
Suricata Snort engine
SSL resumption for stack
Configure packet capture settings in span port
Import and export Sensor configuration
Allocating interfaces at child domain
Important
Changing the stack topology between a 2-node stack and a 3-node, 4-node, or 5-node stack requires a system reboot.
Considerations for failover in stacked Sensors
While configuring failover for a stack, you need to take the following factors into consideration:
Both the stacks must be identical to each other with identical connections, network modules, and capacity.
In a 2-node stack, port G0/3 handles failover, and port G0/4 is unusable.
In a 3-node/4-node/5-node stack, G1/1 is failover interconnect, G1/2 is unused, G1/3-G1/4 is a monitoring port.
G1 is a pluggable IO module, and it needs a 100G Pluggable IO module inserted for HA creation, without it HA cannot be created for 3/4/5 node stack, anyone of the following IO Modules has be plugged-in to G1 slot for HA:
The 4-port SR (Short Reach) MTP (Multi-fiber Termination Push-on)/MPO (Multi-Fiber Push-on) network interface module provides 100 Gbps /40 Gbps ethernet performance on each port.
The 4-port SR (Short Reach) MTP (Multi-fiber Termination Push-on)/MPO (Multi-Fiber Push-on) with internal fail-open network interface module 100 Gbps /40 Gbps ethernet performance on each port.
The 4-port 100 Gbps LR4 (Long Reach) with internal fail-open network interface modules provides 100 Gbps ethernet performance on each port.
The 4-port BiDi (bidirectional) with internal fail-open network interface module provides 100 Gbps /40 Gbps ethernet performance on each port.
The supported DAC cable length for the failover is 1 meter. It is recommended that the stack configuration is installed within one rack. If the stack is configured in multiple racks, you will need to use the QSFP28 fiber transceiver modules.
Ensure you have the correct license to configure failover for a stack. The licenses required are as follows:
License SKU
Throughput
No of Sensors
Throughput required
FO96X20ECE-AT
MSP-NS96FO20-OT
20 Gbps
2, 3, 4, or 5 NS9600 Sensors
120 Gbps (2-node)
180 Gbps (3-node)
240 Gbps (4-node)
300 Gbps (5-node)
FO96X40ECE-AT
MSP-NS96FO40-OT
40 Gbps
FO96X60ECE-AT
MSP-NS96FO60-OT
60 Gbps
For more information on NS9600 licensing, see Management of NS9600 Sensor licenses.
Note
In the event of node failure in a stack, the remaining nodes will continue to function and a fault is generated in Manager → Troubleshooting → Logs → System Faults.