Control Plane Overview
The Cisco CRS-1 Series Carrier Routing System 16-Slot Line Card Chassis control plane provides a communication path between cards, modules, and components in the chassis. The control plane is a logical entity that ties physical chassis components and software functions into a unified entity. The control plane connects the system controller functionality on the route processor (RP) to the service processor (SP) module used to control each card and module in the chassis.
The control plane is used for:
- System discovery and inventory
- Configuration management, system boot, and upgrades
- Inventory control and asset tracking
- Fault detection and recovery, and performance monitoring
The data plane is the path that packets take through the routing system from the physical layer interface module (PLIM) to the modular services card (MSC) to the switch fabric to another MSC and out a PLIM. The control plane and data plane may share some physical components. For instance, the control plane uses the switch fabric for some types of intrasystem communication, just as the data plane uses it to switch packets.
The control plane hardware provides for system discovery and inventory. This process includes mechanisms to determine system topology of the control plane and switch fabric before the system has been configured. In addition to topology discovery, the control plane hardware must also provide mechanisms for card- or module-presence detection and tracking information, such as the card type, revision, and serial number. These mechanisms allow system management software to build a database that represents the routing system configuration, including individual board identification and location information. The control plane hardware provides online insertion and removal (OIR) detection.
The Cisco CRS routing system hardware detects, isolates, and recovers from a broad range of faults, and provides failover mechanisms to redundant hardware. The control plane is a central element in achieving high availability, as it must isolate failures and direct failover events, both in the data plane and in the control plane. To ease serviceability, chassis identification displays and critical, major, and minor alarm indicators are clearly visible. Each MSC, RP, fan controller card, and switch fabric card has an alphanumeric display and green OK LED to show current board status. Environmental conditions, including temperature and voltage levels, are monitored by several internal measurement points and reported to the routing system operator.
The RPs function as the system controller in a Cisco CRS 16-slot line card chassis. Note that the PLIMs are connected to the control plane through their respective MSCs. The control plane includes a switched point-to-point Fast Ethernet (FE), driven by these FE switches, for control plane network messages, and some other paths for system communication. The dual RPs and midplane FE traces provide redundant connections between all cards in the Cisco CRS 16-slot line card chassis. Most cards or modules contain a service processor (SP) module that provides the communication for that device within the control plane.
Some of the important functions and implementations of the control plane are:
- Online insertion and remove (OIR) detection—Every MSC, RP, switch fabric card, power module, and so on provides a presence-detection signal to the system controller function on the RP cards. This dedicated hardware signal indicates the physical presence of a card in every slot. The presence-detection signal allows the Cisco IOS XR configuration software to quickly detect OIR events, and identify cards that have been inserted but cannot communicate over the control plane.
- PLIM inventory—Every PLIM slot is probed by the master RP to get the board ID and type and other inventory information. The RP can read an identification chip on each PLIM, even if the PLIM is not powered on. The PLIM inventory chip can be accessed by the RP, whether or not an MSC is plugged into the MSC slot associated with the PLIM.
- RP active/standby arbitration—Both RP slots are directly connected by dedicated midplane signals to special hardware arbitration logic. During the boot process, this logic selects one of the RPs to be the master (active) device; the other RP functions in standby mode. See Route Processor Active and Standby Arbitration for more information.
After hardware arbitration, software should verify single RP mastership via control plane FE messaging. The arbitration hardware could elect two masters due to an unusual hardware fault. The control plane FE provides a redundant path so mastership can be verified absolutely.
- Node reset—Each RP has a dedicated reset line to every node in the chassis. Nodes include MSCs, RPs, and fabric cards. The reset lines fan out from each RP and are connected to the SP on the node cards. Only the master RP can assert these reset lines; the standby RP reset lines are isolated by the RP arbitration logic. The reset lines allow the RP to force a board reset from hardware, and is used only if a board does not respond to control network messages. When this mechanism is used to reset an SP, power to all other chips on that node are turned off until the reset SP has rebooted and enabled power to the local board. To prevent glitches on the reset lines from causing inadvertent resets, as might occur during an RP OIR event, a reset from this signal can be triggered only from an encoded string of high to low transitions.