This chapter focuses on routing and switching with ASR 1000 series routers. The chapter starts with a brief discussion of the capabilities of the router family, and then reviews how those strengths can be used to address relevant problems.
This approach, with detailed configuration examples where applicable, will allow you to understand the problems, the challenges they represent, and how you can use the ASR 1000 to address them.
Introduction to the Scalable and Modular Control Plane on the ASR 1000
The control plane is a logical concept that defines the part of the router architecture responsible for building and drawing the network topology map (also known as the routing table) and manifesting it to the forwarding plane (where actual packet forwarding takes place) in the form of the Forwarding Information Base (FIB).
While the forwarding capacity of the routers has continuously scaled throughout the years (the Cisco CRS-1, for example; the forwarding capacity for which boosted up to 92 terabits per second [Tbps]), the control-plane scale is given less attention. When routing products are compared, the focus is usually on the forwarding capacity (packets per second or bits per second).
Contrary to this popular notion, the control-plane scale is equally critical to ensure that the platform has the compute cycles in the form of Route Processor (RP) CPU to perform the following (among other things):
- CLI, and similar external management functions performed via Simple Network Management Protocol (SNMP) or Extensible Markup Language (XML)
- Routing protocols and their associated keepalives (including crypto functions in the control plane)
- Link-layer protocols and their associated keepalives
- Services such as RADIUS, TACACS+, DHCP, Session Border Controller (SBC), and Performance-based Routing (PfR) Master Controller function
- All other traffic that cannot be handled at the data plane (for example, legacy protocols such as IPX), including punt traffic
The Cisco ASR 1000 router series delivers complete separation of the control and data plane, which enables the infrastructure's control plane to scale independently of the data plane. The ASR 1000 has two RPs on the market today: ASR1000-RP1 (first generation) and ASR1000-RP2 (second generation). ASR1000-RP1 is based on a 1.5-GHz RP CPU, whereas ASR1000-RP2 hosts a dual-core Intel 2.66-GHz processor, literally increasing the scale many times over the ASR1000-RP1.
The central benefit of physically separating the forwarding and control planes is that if the traffic load becomes very heavy (the forwarding plane gets overwhelmed), it simply doesn't affect the control plane's capability to process new routing information.
Another way of looking at it is if the routing plane gets very busy because of any of the relevant tasks, causing the control plane to be busy, perhaps because of a flood of new route information (even worse, peer or prefix flaps), busy-ness doesn't adversely affect the capability of the forwarding plane to continue forwarding packets. This is a common problem that plagues all software-based routers (due to a single general-purpose CPU running both control and data planes).
Key applications that benefit the most, from a big picture perspective, are network virtualization, infrastructure consolidation, and rapid rollout of various network-based services.
Before delving further and discussing the actual use cases from real-world networks, a quick refresher is in order on some commonly used terms.