QoS policies for lan ip Phones

Source: Internet
Author: User

QoS policies for lan ip Phones this article discusses the main factors that affect the quality of voice service over IP networks, and proposes solutions based on LAN VoIP. This paper mainly analyzes the mechanism of ensuring the Quality of Service in the LAN, and puts forward the QoS Assurance policy in the campus network and backbone access. VoIP IP Phone QoS lan I. Introduction With the development of the Internet, more and more applications are added to the IP network due to ease of management and economic savings. It has been widely believed that today's IP networks are "best-effort" Quality assurance methods and cannot meet the existing service and market requirements. In terms of speech, people are used to high-quality voice of PSTN. Therefore, to transmit voice over IP networks, we must improve the quality of Voice over IP networks. At the same time, it has become a hot topic in the industry. Ethernet is currently the most widely used LAN technology. Both 100 Mbit/s fast Ethernet and 1 Gbit/s Gigabit Ethernet are based on the IEEE 802.3 CDMA/CD specification. Some may argue that QoS is not a concern in the LAN. However, considering the critical points in the LAN, such as the upper line and local area trunk, when an organization is configured with more and more applications requiring bandwidth, there may be situations that exceed expectations. This will eventually lead to the need to use QoS tools in the LAN. Ii. Factors Affecting the Quality of IP speech existing networks (the Internet) have roughly the following impacts on the quality of service for IP phones and faxes: real-time bandwidth: the original design of the IP network does not guarantee the expected bandwidth for real-time voice transmission. Although the demand for real-time voice transmission bandwidth is very low, it requires a constant or direct available bandwidth. Latency: end-to-end (less than 250-350 ms) latency includes latency caused by encoding and decoding, package and unpackage latency, and network transmission latency. Packet loss: the integrity of IP network data is achieved through the re-transmission mechanism of TCP. The transmission of IP network voice packets adopts the UDP mode, and the packet integrity cannot be guaranteed. To ensure the call quality, the packet loss rate is generally less than 5%. Non-linear Speech Encoding: the nonlinear speech compression encoding scheme does not retain all spectrum information. 3. mechanism for ensuring voice QoS in the LAN. 1. classification in a centralized LAN environment, when it is necessary to protect voice communication from data communication, the classification of LAN communication becomes very important. QoS in a LAN mainly adds priority fields to the Ethernet frame header on the second layer to distinguish different priorities. Strictly speaking, a LAN can only be divided into business priorities by region. It cannot be measured by precise definitions and detailed parameter indicators like ATM QoS. Classification technology provides network-first functions on Ethernet, virtual LAN (VLAN), and other networks. This solution differentiates businesses of different priorities based on the processing of IEEE 802.1 p/Q protocol fields. IEEE 802.1 p/Q belongs to the same subset. It adds four bytes to the Traditional Ethernet frame header, of which 802.1p occupies three places. 802.1p extends the 802.1d Protocol. A maximum of eight priority levels can be provided with a three-bit priority. 802.1Q uses VI (VLAN Identifier, virtual network Identifier) bits to identify which virtual network the transmitted frame belongs. A total of 12 VI bits are supported. The maximum number of virtual networks supported is no more than 4096. 2. qoS in the queue management LAN is used to check when a communication enters a vswitch, classify frames according to certain rules, and then queue important data, in order to give priority to the transmission of these important data. The actual handling method of queuing can be changed according to the specific switch used. Cisco Catalyst 2900 and 3500 series L2 switches support two separate queues for each physical interface. A vswitch can classify communications based on the 802.1Q/p marking method described earlier, or, as defined by the network administrator for unlabeled frames, classification can be completed on each port. Once the communication has been classified, the high-priority communication (CoS = 4-7) can be sent to the acceleration (high-priority) queue, while the low-priority communication (CoS = 0-3) is sent to the common queue. This scheduling is essentially a priority queue algorithm. Generally, a layer-3 LAN switch is used for LAN core. For example, the Catalyst 6500 series provides additional QoS features for priority voice communication and supports IP priority classification, you can also provide multiple queues for each interface and support various priority scheduling technologies, such as Weighted Fair Queuing (WFQ, Weighted Fair Queuing), Custom Queue (CQ, Custom Quening), prior Queue (PQ, Precedence Queuing), and Weighted Random Early Detection (WRED, Weighted Random Early Detection ). (1) IP priority: three bits in the Type of Service (ToS) field of the IP packet header are used, which means that by using the priority bits, you can obtain up to 8 values. Because the two values are usually reserved for the routing protocol, ICMP message, and other network overhead, this note allows a maximum of six service types to be defined. Once the IP priority value has been set, the QoS feature on the network device can be enabled. The router can use this information to determine the service type required by a specific group. For example, WFO and WRED can use IP priority to control the QoS status. (2) Weighted Fair Queuing: the goal of WFQ is to provide a more equitable available bandwidth allocation for small-traffic communication, or to interact with the communication type to improve on the congested link, without increasing bandwidth. The WFQ algorithm dynamically distributes communications to information flows based on Packet characteristics, such as source/destination addresses, protocols, and port/socket numbers, layer 2 features can also be used for communication classification, such as the source/destination MAC address or frame relay DLCI. WFQ can identify IP priority values and schedule these packages to deliver them faster, reducing latency and response time. In WFQ, the formula below shows the percentage of the total interface bandwidth given to each IP communication stream based on the priority level and the number of information flows. When the IP priority level is high, WFQ allocates more bandwidth for this communication when congestion occurs. WFQ can avoid a large-traffic application like FTP occupying all available bandwidth on one link, while other small-traffic communication such as voice is starved to death due to insufficient resources ". WFQ supports RSVP. One disadvantage of WFQ is that if there is a lot of information flow in the network within a given period of time, it may become too fair to guarantee the bandwidth of a specific communication type. (3) class-based weighted fair queuing (CBWFQ): It is an extended version of the standard WFQ function and adds support for user-defined communication classes. With CBWFQ, the network administrator can separate communications and put various communications into the queue according to the protocol, access control list (ACL) or input interface. A maximum of 64 types of communication can be defined in CBWFQ. Once the communication has been specified as a type, you can specify the bandwidth for this type and ensure the bandwidth during the congestion period. CBWFQ ensures bandwidth for a specific communication type, making it a better WFQ Method for speech. (4) priority queue (PQ): the network administrator determines which communication must be first queued and served. Various Network streams can be classified based on the following items: Protocol or protocol type, inbound interface, group size, segmentation, and access list. The PQ queue priority can be set to high, medium, normal, or low. The router provides services for the queue in the order of priority from highest to lowest. The service order is as follows: if there is communication in the high-priority queue, normal queues cannot forward any groups until all groups in the high-priority queue are transferred. This is a good mechanism for ensuring key applications, such as interactive speech. The service is a high-priority queue and never serves a low-priority queue. CQ allows communication of a certain percentage of other queues. The capability of this percentage ensures that each queue will be regularly served and bandwidth at some levels is guaranteed. (6) weighted Random Early Detection (WRED) WRED attempts to overcome the discard problem by randomly carrying out packet loss before the buffer congestion. WRED determines when packet loss starts based on the average queue length. Once the number of groups in the queue exceeds the defined queue limit, WRED begins to packet loss within the queue limit. Packet Loss is completely different for network streams. Because the classification is randomly discarded in the queue, only a few sessions will be restarted. This provides the network with an opportunity to empty the queue. As the remaining sessions continue to flow, the buffer can be cleared and other TCP sessions can be restored. WFQ, CBWFQ, CQ, PQ, and WRED are mutually exclusive on an interface. Iv. QoS Assurance Policies for IP phones in LAN basic technologies currently are based on Ethernet (traditional 10 Mb/s Ethernet, fast Ethernet, or Gigabit Ethernet. For the IP network telephone service, the impact of delay will not become a problem. Because of the distribution of network traffic from 100 M to M, the Buffer of the switching port is easy to fill, and the packet is lost. To ensure the quality of voice/video, the following QoS mechanisms must be used in the campus network: 1. the policy for classifying different applications is: (Precedence is the priority) (a) Speech: Precedence = 5 (B) Speech Control: Precedence = 3 (c) video conferencing: precedence = 4 (d) Video Stream: Precedence = 1 (e) data: Precedence = 0-2. select a switch that supports multiple queues on the input/output ports. The following switches have multiple queues: Listen st6000, listen st4000, listen st3500, and Catalyst 2900. 3. use different queue output policies for voice: PQ Video conferencing Video: PQ Video Stream: CBWFQ voice control: CBWFQ (8kb/s) 5. Summary This paper designs the implementation strategy and method of this scheme in the speech QoS of LAN Ethernet. This paper explores how to use the packet priority field and the routing queue mechanism to implement high-quality calls under LAN conditions. It is of great research and practical value to further explore QoS in LAN. As LAN information construction and application systems gradually increase, the network is a multi-media communication that integrates voice, data, video, and other media forms. How to Ensure speech QoS is still a complicated issue and deserves further research.

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.