Introduction to ten core router performance indicators and Routers

Speaking of the performance of the vro, I would like to give you a brief introduction to what is a vro and what a vro is used!

A router is a network device that connects multiple networks or CIDR blocks. It can translate data between different networks or CIDR blocks ", so that they can "read" each other's data to form a larger network.

Vro has two typical functions: Data tunnel and control. The data channel functions include forwarding decision, backplane forwarding, and output link scheduling, which are generally completed by specific hardware. The control functions are generally implemented by software, it includes information exchange with neighboring routers, system configuration, and system management.

In the past few years, the development of vrouters has continued. In the middle of 1990s, traditional routers became the bottleneck restricting the development of the Internet. Instead, the ATM switch becomes the core of the IP backbone network, and the vro becomes a secondary role. In the end of 1990s, the Internet scale was further expanded, and traffic doubled every six months. The ATM network became a bottleneck, and the vro started again. After the Gbps route switch was launched in 1997, people began to replace the ATM switch with a Gbps route switch, and the architecture of the backbone network with a vro as the core.

I believe that you have a basic understanding of the concept of vro. Let's talk about ten performance indicators of the core vro:

The system switching capability and processing capability of a high-speed router are different from those of a general router. At present, the high-speed Router backplane switching capacity should reach 40 Gbps or above, and even if the system does not provide OC-192/STM-64 interface, this interface must also be supported without upgrading the existing interface card and general-purpose parts in the future. In terms of device processing capability, when the system is running at full capacity, all interfaces should be able to process short packets at line speed, such as 40 bytes and 64 bytes. At the same time, the switching matrix of the high-speed router should be able to process the switching of all interfaces at the wire speed without affecting the traffic type.

One metric: Throughput

Throughput is the packet forwarding capability of the router. Throughput is related to the number of vro ports, port rate, packet length, packet type, route computing mode distribution or concentration), and test methods. throughput generally refers to the ability of the processor to process packets. The packet forwarding capability of a high-speed router must be at least 20 mpps. Throughput mainly includes two aspects:

1. Machine Throughput

The package forwarding capability of a device is an important indicator of the device's performance. The router selects routes based on the IP header or MPLS label. Therefore, the performance indicator refers to the number of packets forwarded per second. The total throughput is usually less than the sum of the throughput of all ports on the router.

2. port throughput

Port throughput refers to the packet forwarding capability of a vro on a port. Usually two test interfaces with the same rate are used. Generally, the test interface may be related to the interface location and relationship. For example, the throughput tested between ports on the same plug-in card may be different from the throughput value between ports on different plug-in cards.

Indicator 2: route table capability

A Router usually depends on the route table created and maintained to determine packet forwarding. The route table capability refers to the maximum number of route table entries in a route table. Because vrouters that execute the BGP protocol on the Internet usually have hundreds of thousands of Route entries, this project is also an important embodiment of vro capabilities. Generally, a high-speed router should support at least 0.25 million routes, and each destination address should provide at least 2 routes. The system must support at least 25 BGP peers and at least 50 IGP peers.

Indicator 3: backplane capability

Backplane refers to the physical path between the input and output ports. The backplane capability is the internal implementation of routers. Traditional routers use shared backplanes. However, as a high-performance router, congestion is inevitable. Secondly, it is difficult to design a high-speed shared bus, therefore, the existing high-speed routers generally adopt the design of a switchable backplane. The backplane capability can be reflected in the router throughput. The backplane capability is generally greater than the value calculated based on the throughput and test packet length. However, the backplane capability can only be reflected in the design and cannot be tested.

Metric 4: Packet Loss Rate

Packet loss rate refers to the proportion of data packets that cannot be forwarded due to lack of resources under a stable and sustained load of the router in the data packets to be forwarded. Packet Loss Rate is usually used to measure the performance of a router when it is overloaded. Packet loss rate is related to the packet length and packet sending frequency. in some environments, you can perform test simulation by adding route jitter or a large number of routes.

Metric 5: latency

Latency refers to the time interval between the first bit of a data packet entering the router and the last bit output from the router. This interval is the processing time of the router that stores the forwarding method. Latency is related to the packet length and link rate. It is usually tested within the throughput range of the router port. Latency has a great impact on network performance. As a high-speed router, in the worst case, the latency of 1518 bytes or less of the IP packet must be less than 1 ms.

Metric 6: Number of back-to-back frames

The number of back-to-back frames refers to the number of data packets when the maximum number of data packets sent at the minimum frame interval does not cause packet loss. This indicator is used to test the router cache capability. Vro with full-duplex line rate forwarding capability, this indicator value is infinite.

Metric 7: latency Jitter

Latency jitter refers to latency changes. Data Services are not sensitive to latency jitter. Therefore, this indicator is generally not an important indicator for measuring high-speed routers. This indicator is necessary for testing services other than data on IP addresses, such as voice and video services.

Indicator 8: Service Quality capability

1. Queue Management Mechanism

The queue management control mechanism usually refers to the router Congestion Management Mechanism and Its Queue scheduling algorithm. Common methods include RED, WRED, WRR, DRR, WFQ, and WF2Q.

Queuing policy:

● Supports fair queuing algorithms.

● Weighted Fair Queuing Algorithm is supported. This algorithm gives each queue a weight), which determines the link bandwidth that the queue can enjoy. In this way, real-time businesses can achieve the required performance, while non-flexible business flows can be isolated from common Best-effort business flows.

● The management of input/output queues should adopt the virtual output queue method.

Congestion Control:

● Congestion control mechanisms such as WFQ and RED must be supported.

● A mechanism must be supported to mark a High Discard priority for traffic that does not meet the CIR/Burst contract of the business level, this priority should be higher than the discard priority of the traffic that meets the contract and the traffic that best serves.

● In an exchange environment where an output queue may compete, an effective method must be provided to eliminate header congestion.

2. Number of port hardware queues

Generally, the priority supported by the router is guaranteed by the port hardware queue. The priority of each queue is controlled by the queue scheduling algorithm.

Metric 9: Network Management

Network administrators manage network resources in a centralized manner through network management programs, including configuration management, billing management, performance management, error management, and security management. The degree of network management supported by devices reflects the manageability and maintainability of devices. Generally, SNMPv2 is used for management. The granularity of network management indicates the granularity of vro management, such as ports, CIDR blocks, IP addresses, and MAC addresses. The management granularity may affect the router forwarding capability.

Metric 10: reliability and availability

1. device Redundancy

Redundancy can include interface redundancy, plug-in Card Redundancy, power supply redundancy, system board redundancy, clock board redundancy, and equipment redundancy. Redundancy is used to ensure the reliability and availability of equipment. The design of redundancy quantity should compromise between equipment reliability requirements and investment. Vrouters can use VRRP and other protocols to ensure router redundancy.

2. Hot swapping Components

Because a vro usually requires 24 hours of work, changing parts should not affect the work of the vro. Hot swapping of components ensures the 24-hour operation of routers.

3. No-fault working time

This indicator indicates the time when the device has no fault according to the statistical method. Generally, it cannot be tested. It can be calculated based on the fault-free working time of the main device or the working conditions of a large number of identical devices.

4. Internal clock precision

Vro interconnection with ATM ports for circuit simulation or POS ports usually needs to be synchronized. When an internal clock is used, its precision will affect the bit error rate.

In the high-speed Router technical specifications, the high-speed Router reliability and reliability requirements should meet the following requirements:

① The system shall reach or exceed 99.999% availability.

② Continuous working time without failure: MTBF> 0.1 million hours.

③ Fault recovery time: the system fault recovery time is less than 30 mins.

④ The system should have the automatic protection switching function. The master-slave switchover time should be less than 50 ms.

⑤ SDH and ATM interfaces should have automatic protection switching function, and the switching time should be less than 50 ms.

⑥ The equipment is required to have high reliability and high stability. The main components of the system, such as the primary processor, primary storage, switching matrix, power supply, bus arbitration, and management interface, should have Hot Backup Redundancy. The line card requires m + n backup and supports remote test and diagnosis. Power supply failure can maintain the connection validity.

7. The system must not have a single fault point.

(T113)