802.11ax Technical Overview of Next Generation Intelligent Wifi

Introduction

When Steve Jobs launched iphone in 2007, during the event he had shown 3 products on screen- a phone, an ipod, an internet communicator, few mins later suddenly he surprised everyone when he had combined it into one product and named it iphone, later story everyone knows how it has changed internet usage landscape and helped in exponential growth of data consumption.  

Before smart phones come to market wifi was majorly dealing with laptop or zero mobility devices, only idea was to get more and more data speed with every release of 802.11 amendments b>a>g>n>ac, which we fully accomplished and enjoyed the exhilarating speed of wifi at home, airports, hospitals, stadiums etc.

However, in last one decade due to unprecedented growth of smart phones, simplicity of implementation and economic viability of designing free wifi networks, has attracted more user towards wifi, because of this we have seen a tremendous pressure on wlan network and evolving highly dense mobility use cases, which was never imagined before for wifi. So to overcome all these challenges and make it future ready, wifi fraternity started looking into it and joined hands to bring a new amendment’s and make wifi more intelligent, which will open the path for next generation WIFI.  So here we discuss different technical aspect of the upcoming major amendment 802.11ax, which will play a big role in making wifi smarter.

Recent statistic of network-king  cisco, it is expected that wifi will have a lion share in data consumption by 2019-2020 and global data traffic will be 10-12 times higher than the level measured in 2014 and wifi will have staggering 53 % data share.

Excluding carrier garde wifi, other major part of it like home and enterprise wifi being used in indoor environment, which are designed haphazardly without any proper network planning, which give rise to OBSS(overlapping BSS), which in turn have a major impact on QOS and QOE of individual users. So recently IEEE 802.11 standardization committee approved the development of a new WLAN standard IEEE 802.11ax which will deal with dense deployment scenario and help in achieving average through put per device at mac layer.

Current Wifi Challenge and bottlenecks:

Here we see what are the inherent technical challenges ahead to solve the problem of wifi networks.

Clear channel access method in existing standard is overly protective, leading to reduced spatial reuse.

• Always on Carrier sensing and Energy detection consuming more resources in wifi

More no of messages to complete the initial connection and authentication, which leads to more latency and delay

• Around 27 message exchange between STA and AP to complete the initial authentication.

It will be standardized as part of 802.11ai.

  

Retransmissions are inefficient and use a lot of airtime

• Wi-Fi network have a lot of retransmissions consuming airtime and causing overhead in transmission

RF spreads evenly everywhere

• RF is sent to all directions and receiver tries to receive it from all directions.
• Benefits of antenna directivity(beamforming) and beam steering has minimal use.

No dynamic transmit power control

• Nearby devices transmit static high power levels

Control and management traffic takes a lot of airtime from user data traffic

• In dense areas, majority of packets are control and management frames and rare piggyback concept in wifi like LTE, where Attach accept can carry a PDN connectivity request.

Legacy device protection reduces network capacity significantly

• Legacy devices are over protected, benefits of new technologies are reduced

Channel access gets congested with large amount of devices

• Channel access is contention based(DCF) and efficiency could be better

Wi-Fi signal processing does not work well with large delay spread

• Large delay spread causes receivers problems decoding the data

One size fits all -- Home Wi-Fi = Stadium Wi-Fi = Medical Wi-Fi = School Wifi

• No major differentiation in operation, capability, implementation and optimization.

Radio traffic flows not properly prioritized for system level capacity

• Protocols are inefficient with high/burst load, clients and APs are equal

Wi-Fi lacks performance management capability

• No visibility to user experience and capability optimize network

Wi-Fi is half duplex technology – cannot receive when transmits

• This cuts efficiency by 50%

New use cases have not been considered with the 802.11 standard

• Wi-Fi is used in ways which were not imagined during earlier standardization

Mobile/cellular networks interfere(LAA/eLAA/LWA/LTE-U/Multifire)-Wi-Fi

• Interference of LTE TDD band(2.3 GHz) 40 with Wifi 2.4GHz in dense deployment

Evolving Applications

Every new amendment designed to overcome the demand supply gap in requirement and new use cases, and of course market forces, competition are also behind it. In last one decade, there is a tectonic shift in new application development and usage below we can see few examples.

• Cellular offload due to carrier grade wifi mass adoption
• Wifi densification due to smartphone and IOT
• Smart & Connected City Wifi
• Connected Vehicles (802.11p)
• Industrial IOT
• New and Enhanced Applications (VR/AR)
• 1024p/HD video traffic from Netflix/Youtube etc.
• Wireless networking
• Wireless backhaul

New Improvements

When current wifi standards IEEE 802.11(n/ac) were developed, main aim was to get aggregative multi station throughput at PHY in the network, where no consideration given to individual user level guaranteed throughput at MAC. But 802.11ax has following primary goals to achieve to be called as next generation or intelligent smart wifi.

• Increase in guaranteed throughput for individual STA in a dense environment. (UL/DL OFDMA)
• Increase in power efficiency
• Efficient use of Spectral Resources-Frequency Reuse
• Indoor and Outdoor Operations over 1 GHz to 6 GHz Frequency Bands
• Better mechanism to handle OBSS  
• Increasing the robustness outdoors as well as over the uplink transmissions
• Enabling Backward Compatibility
• IOT requirement below 1GHz band
• Mission Critical communication
• Better device Power Management
• LTE-wifi co-existence requirement
• Fast moving support for cellular offload
• Usage of wifi in Vehicular environment
• MU-MIMO improvement
• Enhanced Mobility

Overview of the Amendment: IEEE 802.11ax

IEEE 802.11 as a protocol basically deals with PHY and MAC layer, so 802.11ax will bring some fundamental changes in the lower layers. So here main goal is to improve the user experience by reducing interference and provide improved throughput for individual STA in a dense environment. So, same approach of LTE adopted here like MIMO and OFDMA to transmit more bits per transmission opportunity (TxOP).

Following important features or new design changes proposed for IEEE 802.11ax amendments and it is divided into 4 categories PHY, MAC, Spatial Reuse, Simulation & calibration and others.

Physical Layer Enhancement

Physical Coding Decision

L1 deals with different types of coding mechanism to work efficiently, legacy wifi majorly worked on the principle of BCC and LDPC rarely used in some standards. However, in 802.11ax focus is on LDPC with higher MCS.

Physical Coding Decision based on two coding mechanism LDPC and BCC.

LDPC- Low density parity check- High cost in computation may be used for higher order modulation like 1024.

BCC- Binary convolutional coding

Default forward error correction (FEC) scheme proposed for IEEE 802.11n and IEEE 802.11ac is based on binary convolutional coding (BCC) with frequency interleaving per orthogonal frequency division multiplexing (OFDM) symbol.

Larger FFT size- Enhancement for Outdoor Communication

Larger FFT size is proposed to increase robustness outdoors as well as to improve the average indoor throughput. Large path loss and channel delay suffered in outdoor large hotspots, TGax defines a new high efficiency PPDU (HE-PPDU) format, called Extended Range Single User (SU) PPDU, in which the fields that contain the information required to interpret packets are repeated.

1024 QAM- 4x improvement in throughput with more (10) bits per symbol, 66% improve over 64 QAM in 802.11n and 33% improvement w.r.t 802.11ac (256 QAM).

Frequency Selective Scheduling

OFDMA systems benefit from frequency selectivity in terms of frequency diversity and Frequency Selective Scheduling(FSS). In TGax, FSS is being actively pursued to provide throughput gains to far away stations (with respect to AP) by allocating physical resource blocks with least amount of fading for their transmissions.

Beam Forming

It was introduced in 802.11n but not standardized, so it never took off due to difference in proprietary implementation by AP and client side vendors. So finally it was standardized in 802.11ac wave 2 with transmit beamforming which is based on explicit CSI (Channel State Information) feedback matrix.

Now in 802.11ax it is further enhanced to support cell edge performance improvement in dense deployment scenario where overlapping BSS and Interference issues are causing major bottlenecks.

MAC Layer Enhancement

To support the requirement of Multiuser and OFDMA, major following changes are proposed to improve spatial reuse.  

PHYCCA Modifications and DSC

The legacy IEEE 802.11 utilizes physical clear channel assessment (PHYCCA) modules to sense state of the channel (i.e., either busy or idle) by measuring the received energy or best fixed threshold. To make a better approach to handle spatial reuse and multiple concurrent transmission, 802.11ax amendment proposed dynamic CCA threshold. To support this dynamic CCA , one innovative algorithm designed known as DSC-Dynamic Sensitive Control which can improve overall throughput.

What DSC exactly doing?

The basic idea of the DSC scheme is to optimize the existing deployments by appropriately tuning the CCA threshold for each node in a distributed manner. DSC tries to confine the increase and decrease of CCA threshold for a station in a bounded area to avoid both extremely aggressive and conservative behaviour.

The throughput gains achieved by DSC are more than 20 percent on average when combined with optimal channel selection (gain increases beyond 40 percent when stations use slow bit rates and send long frames.

It should be a configurable parameter at AP level in beacon and probe response.

How CCA threshold calculated?

Multiple factors included to decide the value of CCA threshold, like Tx Power, MCS of transmission, path loss and CCA threshold should be different at each STA and it should vary dynamically with Tx Power, MCS and path loss/mobility etc.

Transmit Power Control

What improvement done for TPC?

Like LTE and CDMA transmit power control mechanism improved to minimize the interference as well as increasing the spatial reuse. Here goal is the dynamically adjust the lowest possible power for station with highest path loss with an intention to reach a target signal to interference noise ratio(SNIR), which will be enough to decode the received frames.

BSS Colouring

It is an interesting new scheme to enhance throughput in dense WLAN network, where every BSS/AP is assigned with a color (set in L-Sig bit of physical header). When STA receive any frames from neighbour BSS it will stop listening from same neighbour BSS and assume that medium is idle and provide more TxOP.

How it will work?

BSS coloring is depending on dynamic CCA threshold to work effectively.  BSS color in HE PHY SIG is used for a STA/AP to decide whether the detected 802.11 transmission is from the self BSS or neighbor BSS.

• When the transmission is from the self BSS and the CCA level is lower than the defined lower CCA level, a STA/AP can decide the medium is busy.
• When the transmission is from the neighbor BSS and the CCA level is lower than the defined higher CCA level, a STA/AP can decide the medium is idle.
• The decoding of PHY header before HE SIG is based on normal CCA level.

Multiple NAVs

Legacy 802.11, used virtual carrier sensing to solve the Collison problem of hidden nodes. This technique worked by reserving a channel based on RTS/CTS frames which precede data frames, neighbouring over hearing stations upon receiving the RTS/CTS frames set a timer called NAV (Network Allocation Vector) that blocks them from transmitting for specific time.

Why two NAVs required?

However, The IEEE 802.11ax amendment proposes to utilize two NAV timers at each station (one identified as an intra-BSS NAV and the second called inter-BSS NAV) where the intra-BSS NAV is reset or increased only by the frames from that BSS. Thus, spatial reuse can be increased by allowing the station to ignore RTS/CTS frames transmitted from the inter-BSS or OBSS. This will helps in improving the spatial resuse, as stations ignores RTS/CTS from IBSS or OBSS.

How it will work?

Each STA maintains multiple NAVs.

Each NAV corresponds to a particular BSS heard by the STA.

• E.g., in case of 3 BSSs operating in the same area, each STA maintains 3 NAVs: one NAV for its own BSS and the other 2 NAVs for the alien 2 BSSs.

NAV corresponding to a particular BSS can be reset or increased only by reception of frames from that BSS.

Since some frames may not contain BSSID (e.g. CTS, etc.), BSS color can be used to distinguish overlapped BSSs.

If at least one NAV indicates that the medium is busy, then the medium is considered to be busy.

Interference Management Aspects

Due to possible dense deployment scenarios, interference management is major aspect for wifi, so following aspect taken into consideration and further study is ongoing.

1. TD-LTE(2370-2390MHz) and Wifi(2400-2483.5MHz) band interference around 2.4 GHz band.

2. LTE band-40 (23 GHz) Out of Band(OOB) Interference to Wifi 2.4 GHz Band

3. Beam forming under Overlapping BSS due to co-channel Interference.

4. Adjacent Channel Interference mitigation as it is affecting BSS throughput

Downlink and Uplink OFDMA

In 802.11ax major new addition is OFDMA, which will enhance channel efficiency significantly and bring a whole new user experience in wifi, like how it is being used in LTE.

What are the new changes proposed to support OFDMA?

• RTS/CTS for UL/DL OFDMA control for resource allocation and better channel utilization. Allocate all the channels during downlink operation and for UL resource allocation based on STA’s transmit demand.
• New channel (APCH) to transmit 802.11ax control frames and legacy control frames via legacy channels.
• Better Scheduling and CCA feedback mechanism to handle massive UL OFDMA control operation.
• Below 20 MHz operation like 1MHz, 2.5MHz or 5 MHz to support IOT use cases.
• Trigger frame based random access mechanism to handle UL OFDMA resource allocation for STA
• RAPS(Random access parameter set) element in beacon and probe response to handle Random access procedure.
• 9 or 10 STA MU MIMO OFDMA operation in a 20MHz channel with 26 resource unit (consist of no of subcarrier, in LTE it is known as Resource Block) proposed to increase the physical efficiency.

OFDMA operates on top of OFDM where a control tower or base station(AP) allocates sub carrier in a tight fashion manner to each user to accommodate multiple simultaneous transmission in synchronous way which will reduce contention leads to fewer collision.

Currently in MU-MIMO supported AP only 2-4 users supported if it is 4x4 or 8x8 spatial streams is used. So now as pointed above min 26 resource units in 20 MHz are allocated to end users (9-10 STA can transmit simultaneously) by AP using a specific format called HE-PPDU, HE trigger-based PPDU, which allows the announcement of scheduling decisions. This feature helps to reduce synchronization complexity.

Downlink and Uplink Multi-User (MU) MIMO

MU MIMO support introduced in legacy wifi 802.11ac wave 2 only in DL with 4 spatial streams. Now with 802.11ax it is extended to UL as well. Now in both UL and DL multiple user can enjoy more data throughput.

How it will work?

In UL MU-MIMO, multiple stations can transmit simultaneously over the same frequency resources to the receiver.

• AP transmits an Uplink MU-MIMO Poll frame to initiate the uplink MU-MIMO PPDUs transmission
• HE STAs transmit uplink MU-MIMO PPDUs SIFS after receiving the Uplink MU-MIO Poll frame
• AP replies with an Acknowledgement of the received uplink MU-MIMO PPDUs

Multi-User Aggregation

Frame aggregation was introduced in IEEE 802.11n to reduce overhead by allowing the transmission of multiple data frames in a single channel access (provided that they have the same destination). IEEE TGax aims to further extend the aggregation procedure by defining a multi-user aggregation scheme that will allow a single access to send frames to multiple recipients. This scheme operates to reduce transmission overheads.

Legacy 11n/ac have added/extended AMPDU for SU aggregation so MPDUs can be aggregated if they arrive within the same burst, or within some time limit and DL MU-MIMO has MU aggregation method, but is optional and limited and from beginning Wifi efficiency is not good to handle short packets/burst. But with advent of new application requirement of short packets transmission is huge, which wifi has to handle to reduce the overhead.

Where transmission over head normally generate without aggregation?

Clients ACK in individual frames. So even with a DL MU aggregation, sending ACKs individually adds overhead, specially for shorter packets

In WiFi, AP sends ACK/BA in individual frames. It’s possible to aggregate the ACK/BA in a single frame in response to an UL aggregation method.

Why we need MU aggregation?

• MU aggregation would help in high-density environments, where bursts from multiple STAs could be aggregated in one PPDU
• Considering application with high likelihood of short packets/bursts, MU aggregation methods help to enhance MAC efficiency and reduce medium access overhead
• Lately, WiFi industry has shown interest to add UL/DL OFDMA and/or UL MU to 11ax
• This is good news, but for a complete story there should be mechanisms to aggregate ACKs/BAs in the opposite direction

How to implement?

• Enhancement in polled Ack/Block Ack aggregation mechanism implemented in DL MU- MIMO can be reused for MU OFDMA.

OR

• Use of CDMA-based signaling is also well known in OFDMA-based cellular technologies, In LTE/WiMax, there are dedicated common channel where STAs can put their request for UL airtime, where such resource is available to all STAs.

Enhancement in Power Saving Technique

We have many legacy power saving mechanism like Power Saving Mode, Power Saving Polling and U-APSD. To extend it further TAGx actively working on it to reduce the awake time or extend the sleep time, which is required in high density network or for low power mode operation to support LPWAN use cases.

In addition, the TGax is also exploring the possibility to reuse different energy efficiency techniques proposed for the upcoming IEEE 802.11ah standard such as target wake time (TWT), where a routine and schedule for sleep is permitted by the AP to the associated stations.

How TWT works?

Target Wake Time (TWT) is a power saving mechanism, negotiated between a STA and its AP, which allows the STA to sleep for periods of time, and wake up in pre-scheduled (target) times to exchange information with its AP. We can discuss about it more in a separate article.

Challenges

Due to explosive traffic volume growth and fierce competition for unlicensed spectrum has make it a battle ground. Till 802.11ac unlicensed spectrum bands like 2.4GHz and 5GHz only used by IEEE group and also it was used for (particularly LBT) scientific purpose. Now many other forum like LTE-U and 3GPP started experimenting on unlicensed band. So now lot of technical changes required for wifi to co-exist in the same bands.

Wifi IEEE 802.11 PHYCCA primarily worked on LBT-Listen before Talk where LTE is completely opposite it is an always ON rude technology which will make the medium busy all the time. So from both side some amount of changes is required to share the band equally between different technologies.

Also unlike WiFi, in which devices use a distributed mechanism to contend for access to the wireless medium, LTE relies on base stations as central schedulers for medium access of all associated nodes in a cell. Since operation in unlicensed bands is non-exclusive, medium access inherently needs to employ means for fair spectrum sharing.

Even though LBT is the fundamental requirement for Unlicensed spectrum, LTE-U implemented it in a different way using a technique called dynamic on/off scheme or carrier-sensing adaptive transmission (CSAT) and 3GPP based LAA implemented almost similar to Wifi enhanced distributed channel access (EDCA) scheme. Both of these technology needs further evaluation, can be considered a current challenge for the IEEE 802.11ax standard.

So lot of new techniques proposed under 802.11ax to simplify and enable co-existence between LTE unlicensed and Wifi.

In addition to above challenges 802.11ax has to full fill the requirement and use cases of Internet of Things. Even though separate amendment 802.11ah is there, which will work below 1GHz band and covered major requirements, still during development of 802.11ax has to give some consideration for IOT use cases.

A comparative study on 802.11ac,802.11ah and 802.11ax

Time Lines

Like 802.11ac, this new amendment will be released in 2 waves/phases, first draft version is already released in November 2016 and second draft version we may see by the end of 2017 and final amendment by 2018 end.

Conclusion

Any technology is obsolete if it is not evolving, same holds true for wifi as well. Currently wifi has to compete with many new and evolving technologies like LAA, LTU, LWA, Multipath TCP, LTE-Multifire, LORA, Sigfox, CBRS(3.5GHz), Lifi and many more. So development of 802.11ax is vital to provide affordable connectivity and connecting the un-connected.

References

802.11ax forum

IEEE paper on 802.11ax

3GPP specs and Multifire forum