Congestion management refers to the strategies and techniques used by a network to handle and control traffic when congestion occurs. One of the simplest methods is the FIFO (First In First Out) queue, which does not classify packets based on their type or importance. Instead, it queues packets in the order they arrive at the interface and forwards them in the same sequence. This approach follows a best-effort forwarding model, which may not be suitable for time-sensitive applications like VoIP.
To better manage traffic, networks often use traffic classification and marking. This involves identifying different types of traffic based on specific fields within the packet headers, such as the CoS field in 802.1Q Ethernet frames, the ToS field in IP packets, or the EXP field in MPLS packets. Once marked, these packets can be processed according to predefined rules by other devices in the network.
For example, at the network edge, voice data packets (VoIP) might be classified into an EF (Expedited Forwarding) service class and marked with a DSCP value of EF or IP precedence of 5. Voice control packets could be assigned to AF (Assured Forwarding) service class with a DSCP value of AF31 or IP precedence of 4. These markings allow intermediate nodes to prioritize traffic appropriately. For instance, EF traffic ensures low latency and minimal jitter, while AF traffic guarantees a certain level of bandwidth during congestion.
Congestion typically arises when the data transmission rate exceeds the capacity of the network link. For example, if a branch office sends data to a headquarters over a 100Mbps connection, but the WAN link only supports 10Mbps, congestion will occur at the router's serial interface. This situation highlights the need for effective congestion management techniques.
Common congestion management technologies include FIFO, PQ (Priority Queuing), CQ (Custom Queuing), WFQ (Weighted Fair Queuing), and CBWFQ (Class-Based Weighted Fair Queuing). Each of these uses different queueing algorithms to manage traffic based on priority, fairness, or specific service requirements.
The FIFO queue operates on a simple principle: packets are queued in the order they arrive and dequeued in the same order. While this method is easy to implement and has low overhead, it lacks the ability to differentiate between packet types, making it unsuitable for real-time applications that require guaranteed performance.
In contrast, PQ prioritizes traffic based on predefined categories. It divides packets into four priority queues—high, medium, normal, and low—and processes them in that order. This ensures that critical traffic, such as VoIP, gets transmitted first, while lower-priority traffic waits until higher-priority packets are sent. However, a major drawback of PQ is that it can starve lower-priority queues if higher ones remain constantly busy.
UL RHW-2,Lines for Construction,UL Certified Cable, Industrial Cables Wire
Jiangyin City Weicheng Special Cable Co.,Ltd , https://www.weichengcable.com