Meeting the requirements for video-grade Wi-Fi Access Points

There’s a proliferation of wireless video-consuming devices in modern home networks including both stationary devices such as set-top boxes (STBs) and smart TVs, and portable devices such as tablets and smart phones. As such, a new breed of Wi-Fi access points (APs) is required to support the robust delivery of multiple HD video streams to all these devices. Since the various video end points typically come from different vendors, each with different Wi-Fi capabilities (e.g. different number of antennas), a smart video-grade AP must allocate its resources according to each device’s needs and at the same time optimize the overall network performance.

Traditionally a Wi-Fi device was measured according to its receiver performance, using metrics such as receive sensitivity. However, a video-grade AP that streams multiple downlink streams is measured according to its transmitter performance. With video delivery, transmitter performance is ultimately measured according to user experience. Users are coming to expect a wireless video experience equivalent to video consumption over wires.

Raw throughput is no longer an adequate metric for measuring wireless network performance, as was the case with internet data traffic. Quality-of-service (QoS) related metrics, such as packet loss, delay and jitter, to which HD video is highly sensitive, must be optimized as well, sometimes at the expanse of raw throughput. If we allow for around 10-20 Mbps per HD stream, multiple HD stream delivery requires no more than 40-80 Mbps of actual throughput. The main challenge is to deliver this relatively modest throughput consistently and robustly across an entire home to multiple nodes. This must be accomplished over a constantly changing and hostile wireless fading, interference prone channel. The challenge becomes converting an unpredictable medium into a controlled and managed one.

This article shows how current-day Wi-Fi technology has matured and when properly implemented and tailored for the task, it can sustain and deliver multiple HD video streams across an entire home.

The Problem with Traditional Wi-Fi AP Devices
Over the past decade, Wi-Fi technology has evolved from the 802.11b CCK modulation in 2.4GHz, to 802.11a/g OFDM modulation, to 802.11n MIMO-OFDM. While the physical layer was significantly improved in each new Wi-Fi generation, increasing the theoretical maximum throughput significantly, the MAC mechanism has not significantly evolved from the first days of Wi-Fi.

No real QoS measures have been put in place to cope with the increasing demand for robust and jitter-free traffic. The more manageable and contention-free MAC modes of HCF and HCCA have not been embraced by the industry, and the dominant MAC policy has continued to be DCF, which is essentially an unmanaged, best-effort and collision prone mechanism.  802.11n has introduced two new MAC enhancements to increase MAC efficiency and reduce protocol overhead - namely frame aggregation and block acknowledgment. With frame aggregation, several IP frames can be aggregated into a single physical layer packet. With block acknowledgement, a receiver acknowledges a block of frames instead of each frame individually. While significantly reducing MAC protocol overhead, these two mechanisms can actually increase the delay and jitter of the underlying HD video streams.

Block acknowledgment introduces an inherent delay, and block aggregation is very sensitive to the proper selection of the over-the-air (OTA) rate, since the selection of a too-aggressive OTA rate can result in an immediate loss and a need to retransmit a group of packets, which significantly increases the underlying steam's jitter.

While many implementations boast theoretical OTA rates of 300 and even 450 Mbps for systems that support two and three spatial streams respectively, these rates can usually be sustained only in very high SNR (signal-to-noise ratio) scenarios, i.e. very close range, and cannot be guaranteed consistently across an entire home.

What is needed is a new breed of APs, tuned for HD video delivery that will transform an 802.11n core into a robust and consistent vehicle for multiple HD video stream delivery. These video-grade APs should employ a holistic cross-layer design approach, to form an expert system that can issue optimal transmit parameter decisions, based on analysis and knowledge of current channel condition and traffic demands.

Figure 1 illustrates the evolvement of 802.11 technologies from 802.11a/g to 802.11n. While maximum throughput was significantly increased, no particular measures were put in place to significantly increase coverage and robustness.



Figure 1 Evolution of 802.11 technologies

 A video-grade AP must solve the following three problems that have notoriously plagued traditional Wi-Fi home networks:

1. Increase link budget at a reasonable cost to increase coverage.
2. Make optimal and consistent use of the available link budget to increase throughput and reduce delay and jitter. This mainly involves optimal channel aware OTA rate selection.
3. Schedule and prioritize multiple HD streams to multiple end devices.

All of these tasks can be achieved with current day 802.11n technology, using a cross layer implementation approach, as will be detailed in the following sections.