Brief Analysis on whether Wi-Fi can carry VoIP?

For wireless data networks, voice is a "killer application ". The high-performance Wi-Fi mesh network system is a killer IP wireless network. However, not all mesh networks are identical. As wireless mesh networks become increasingly popular-almost every day, people announce the newly deployed public and private networks. To add voice applications to business needs, the network needs to improve overall performance so as to process real-time applications.

Once Multiple Relay segments (hops) appear in the mesh, some problems may occur, such as bandwidth decline, network latency, and application competition for priority. If the network covers a wide range of regions, these problems will become more serious. The lack of network performance seriously affects some real-time services that Wi-Fi service providers most want to provide, such as VoIP.

Mesh networks have four important requirements for the performance of real-time applications. However, three mesh network architectures can be used to deploy wireless IP voice transmission (wVoIP) on a Wi-Fi mesh network; is there a positive impact on the capital cost required for the deployment of multiple wireless devices and multi-RF architectures? What are the costs for O & M of these networks?

Wi-Fi mesh network has many advantages of Wi-Fi, and greatly saves the cost for planning, deploying and running such networks, especially in environments that are hard to connect and cannot be connected, and/or a man environment that cannot be connected to hundreds of nodes within dozens of square miles. A converged Wi-Fi network that supports data and voice can even benefit from the mesh network architecture, because multiple wireless devices deployed increase the coverage distance and bandwidth density. High-capacity mesh network nodes usually provide the lowest bandwidth cost for wireless infrastructure devices and installation. In addition, it is easy to reduce network operation costs.

Four requirements of the mesh network

Network Infrastructure performance must be able to provide high throughput, low latency, and end-to-end service quality, not just between wireless phones and access points, it also extends the network link across the mesh to the wired endpoint (usually an IP Switch ). Because of this, the mesh backbone network must provide:

High throughput across multiple relay segments. Regardless of the number of relay segments (usually 3 to 10), the mesh backbone network must support traffic load. The ability to provide high throughput is directly related to the number of speech and data users supported. If the bandwidth is insufficient across multiple relay segments, the user density will be unsatisfactory. Additional devices and more wired endpoints are also needed on the network.

Low latency across multiple relay segments. High throughput is not enough. To avoid signal jitter, each relay segment must minimize the packet delay. In the mesh network, the waiting time of data packets in any node must be minimized (preferably reduced to only 5 milliseconds for each relay segment ). Because of this, a data packet should be sent out before receiving all data packets from a data stream sent from the previous node. Data transmitted on the mesh network must be asynchronous rather than synchronous. In the case of synchronization, a highly synchronous inter-node packet routing protocol is required.

End-to-End Service Quality-prioritize voice packets. If the network load is large, high throughput and low latency alone are not enough. In order to deal with contention and spontaneous load demands, voice streams on the backbone network of the network must be processed first, and end with an end-to-end traffic priority division mechanism. It is far from enough to provide a service level between AP wireless devices (just like wired AP) that provide services for wireless phones and the devices. The mesh network must have a quality of service for the entire backbone network to avoid contention for each relay segment in the mesh network. This service level needs to be automated (depending on the infrastructure). It is best to process it through different virtual LAN/Service Set Identifiers (VLAN/SSID) dedicated to voice. There is still a long distance between 802.11e and actual deployment. Do not expect 802.11e to be widely used in wireless infrastructure and all client devices in the near future.

L2 switching network. The L2 network minimizes the roaming problems in the L3 network. Layer-3 networks also need to be carefully planned for different types of advanced protocols. These two factors cause performance problems and Protocol configuration problems. The second-layer wireless network can act as an advanced "line ".

The above four factors have a direct impact on scalability (in terms of the number of users and network coverage) and speech quality. If the topology of a specific multi-relay segment does not meet these requirements, the functions are limited and the speech quality function is missing.

Three deployment methods

The wireless mesh network solution varies, but most of the technology comes from the initial concept of the wireless distribution system (WDS. WDS is a Wireless AP model that uses wireless bridging and wireless relay. The AP only communicates with each other and does not allow wireless clients to access itself. The latter means that AP can not only communicate with each other, but also communicate with wireless clients. Before leaving the network, user traffic must pass through several nodes (such as through a wired LAN). This is an inherent feature of various mesh networks. The number of relay segments for user traffic to pass to the destination depends on the network design, link length, technology used, and other uncertainties.

Single-wireless device solution-one-channel transmission of various signal single-wireless device methods is the weakest wireless mesh network solution. The access point only uses one radio channel and is shared by the wireless client and the backbone network (traffic is forwarded between the two APs ). If more APs are added to the network, a larger proportion of the wireless bandwidth will be used to forward the backbone network traffic, so that the wireless client does not have much capacity to use, because the wireless network is a shared medium. In addition, the AP cannot send and receive signals at the same time. When another AP within a valid distance transmits signals, the AP cannot send signals at the same time. In this way, after only three relay segments are passed, the latency will be unacceptable.

A simple calculation proves that in this single-wireless device solution, each wireless client can only use a limited throughput. For example, if you have five APs, each of which is connected to 20 wireless clients, because all APs and clients share the same 802.11b channel (5 Mbps ), therefore, each user only has 50 kb to kbps-the same throughput as the dial-up connection. Because all wireless clients and APS must work on the same channel, network contention and RF interference will lead to unpredictable latency.

Dual-wireless device solution-shared return transmission if dual-wireless device solution is used, a wireless device can be specially used to support wireless clients, another wireless device is dedicated to supporting wireless return transmission-the return transmission channel is shared by inbound traffic and outbound traffic. Because the dual-wireless device solution provides a separate wireless device for client access and return transmission, this reduces the congestion (Low Throughput and low latency) on the client to a certain extent ), however, the inbound traffic and outbound traffic share the network channel of the returned transmission network, because the wireless device of the return transmission must still switch between the inbound network of the returned transmission network and the outbound network of the returned transmission network. Therefore, compared with the single-wireless device architecture, this solution has little effect on solving the bottleneck problem of back-trip transmission, and it is only a little help to improve the latency of the entire mesh network.

Multi-wireless device solution-structured Wireless Mesh Network has several dedicated link interfaces in the Multi-wireless device or "structured mesh network" solution, and each network node uses at least three wireless devices, includes a wireless device for wireless client traffic, a second wireless device for 802.11a wireless return transfer traffic inbound, and a third wireless device for 802.11a return transfer traffic outbound. Compared with single-wireless devices or dual-wireless devices, this wireless mesh network solution greatly improves the performance. It allows a dedicated mesh network to send and receive requests at the same time, because each link is on a different channel.

A dedicated wireless device is responsible for the following functions:

(1) ensure high throughput of 10 relay segments;

(2) The latency of each relay segment is also limited to 4 to 5 milliseconds, and the total latency of 10 relay segments is 50 milliseconds-far lower than the 120 milliseconds required by voice.

(3) If each wireless device supports service quality and multiple SSID/VLANs, voice traffic is handled with a reasonable priority throughout the process from a wireless phone to a wired endpoint through a mesh network.

How can we achieve the best operation?

The investment in the purchase and installation of wireless infrastructure depends on the capital fee per region and the number of users who can provide services in a specific region. Daily operating fees include not only network management and maintenance fees, but also the fees typically paid to infrastructure broadband providers by DSL, T1, T3, and/or OC3 broadband endpoints.

For the service provider's capital fee, the most important thing is the cost of each wireless device (one wireless device can provide 54 Gbit/s of bandwidth, equivalent to supporting a certain number of users based on the amount of bandwidth allocated to each user), and the cost of deploying a wireless device to meet user needs according to the required density. Multi-wireless device, multi-RF and multi-channel wireless mesh network architecture can deploy the most density wireless device with the most cost-effective solution. Compared with the three dual-wireless device nodes, the cost per region for purchasing and deploying six single-wireless devices should be lower. The number of concurrent voice calls that can be processed by multiple wireless device nodes in a partition should be at least three times that of a typical single wireless device or dual wireless device node.

As an important operation expense, the Broadband Access Point (PoP) provider usually charges fees based on the bandwidth of each PoP in each channel. That is to say, although the bandwidth of one T3 line is equivalent to 30 T1 lines, the cost of the T3 line is roughly three times that of the T1 line, so the bandwidth cost of the T3 line is only one tenth of that of the latter. In the final analysis, using 3 T3 lines instead of 90 T1 lines can significantly reduce costs. For example, in terms of deployment architecture, the T3 link allows each PoP to use a mesh network node ten times more, so that the provider can make full use of all the T3 bandwidth. The only way to achieve this is to use a multi-wireless device, multi-RF, multi-channel wireless mesh network architecture to make the most effective use of Wired PoP. For example, if a single wireless device node is used, 10 endpoints are required every 5 square miles. Instead, a wired end point is enough to use multiple wireless device nodes.

For each wireless device, the cost of Multi-wireless devices and multi-RF nodes is usually low, and the installation cost is also low. In addition, you can use a bandwidth with a lower cost but higher capacity per region, take advantage of a much more economical broadband endpoint. To meet the needs of real-time communication applications such as voice, Wi-Fi mesh networks require multi-wireless devices, multi-RF, and multi-channel architectures. To deploy a cost-effective network with the necessary transmission capacity and coverage to achieve high throughput, low latency, and high-priority voice traffic, A high-capacity node is required to support the client inbound, the mesh return transmission inbound, And the mesh return transmission output.

1. Detailed description of Wi-Fi mesh networks suitable for VoIP services
2. Integration of VoIP and Wi-Fi: IP Phones move towards Wireless Networks