802.11 AC Primer: Whats all the fuzz about?

802.11ac is the upcoming big standard with tremendous increase in the data rates and throughput if properly utilized. Lets take a Brief/Raw look at the all the new features and how they help to achieve greater data rates and spectrum utilization.

PHY Features

• 256 QAM

• Very high order modulation scheme which increases the spectral efficiency only when used with beamforming technology, as high order modulation schemes are susceptible to noise and interference.

• Compared to 11n 64QAM, spectrum efficiency improves by 33% 

• Require about 30dB increase in SNR and coverage area is reduced (beam forming can solve this)

• Sub-carriers
• The maximum subcarriers that can be used with OFDM in WLAN is 64/20MHz, as of now 11ac uses this limit most efficiently, the next standard will run out the this limit, its time to increase it to 128/20MHz #IEEE :-)
• 11a/11g ==> 52
• 11n        ==> 57
• 11ac      ==> ~59
• 80 MHz

• 160MHz

• 80+80 MHz

• Single continuous 160MHz and 2 discrete 80MHz can also be combined as 160MHz channel, increases the throuhgput but not the spctral efficiency.

• 8 Spatial Streams

• Sounds high, yeah for a single user it doesn't make sense, but with Multi User MIMO we can exploit this to increase the overall spectrum efficiency.

• MU-MIMO

• Instead of using all the Antennae for a single user (Even though some of them are not really used for some MCSes) we can use each antenna for a single user (max up to 4) and serve all of them in parallel.  This poses few issues like how do we identify each STA? What about group frames? How can a STA for which the data is not destined can ignore the frames? 

• The answer to these questions is the additional features introduced in MAC as explained below. 

• AES-256

• GMAC and GCMP

• Present in 11n in most of the enterprise AP's, now its official along with few other extra algorithms.

MAC Features

PHY ID's (Included in the VHT SIG field) 

Basic motivation is determine if the packet is not destined for you at the earliest possible stage (PHY instead of MAC) and go to micro sleep. (Most likely the case when MU-MIMO is in use)

GroupID

"An AP determines the possible combinations of STAs that can be addressed by an MU PPDU by assigning
STAs to groups and to specific user positions within those groups. (through a new GroupID management Frame).
So after decoding the TXVECTOR the STA can decided whether the frame is for itself (or) not."

Note: Group ID 0 is reserved for transmissions to AP and Group ID 63 is reserved for downlink SU transmissions

Partial AID: The partial AID is a non-unique identifier of a STA and is 9 bits conveyed in the TXVECTOR To identify whther the transmissions are destined to a STA/not, used in conjunction with GroupID.

PHY Power-saving with PHY ID's

TXOP Sharing 

In the TXOP won for a particular AC, we can also send frames destined for others AC's to other STA's as well.

"This mode only applies to an AP that supports DL-MU-MIMO. 

The AC associated with the EDCAF that gains an EDCA TXOP becomes the primary AC. TXOP sharing is allowed when primary AC traffic is transmitted in a VHT MU PPDU and resources permit traffic from secondary ACs to be included, targeting up to four STAs."

TXOP power save

Sounds weird but is a good feature. Basically in a TXOP for MU-MIMO, if the frame is not destined for the STA it can doze off for that TXOP duration.

"If the AP allows non-AP VHT STAs to enter Doze state during a TXOP, then a non-AP VHT STA that is in VHT TXOP power save mode may enter the Doze state till the end of that TXOP when one of the following

conditions is met:

— On receipt of a VHT MU PPDU, the STA determines that it is not a member of the group indicated by the RXVECTOR parameter GROUP_ID. 

— On receipt of an SU PPDU, the STA determines that the RXVECTOR parameter PARTIAL_AID is neither equal to 0 nor does it match the STA’s partial AID. 

— The STA finds that the PARTIAL_AID in the RXVECTOR matches its partial AID but the RA in the MAC header of the corresponding frame that is received correctly does not match the MAC address of the STA."

 References:

1. IEEE Discussions: Spectrum Efficiency 
2. IEEE Discussion: PHY Powersave
3. 802.11ac-draft 5.0

Roadmap of wifi: For coming years...

Is wifi saturated? whats coming up in wifi world?
Lately, i have been hearing this question a lot and decided to find an answer.

I believe that the field of wifi is very busy in the coming years, technology wise not so much change, the basic protocol remains same and so are the basic rules, but its more user-centric now and more real world scenario based. Lets take a look at the things coming up and soon in wifi

Lets divide the upcoming things in wifi in to technology based and user based.This time less text and more crisp no detailed technical analysis :-) , because most of these things have a roadmap of around 2015-2016, many of the TG's are in the nascent state.

Technology Centric: IEEE			User Centric: WFA
802.11 ac	PHY Changes	256 QAM, MU-MIMO, Beamforming, BW signalling,	Wifi Miracast	On top of P2P based Sharing ending Video and Audio over wifi, eg: project your video from laptop to TV
802.11aq	MAC Changes	Low Power Pre-association discovery	Neighbour aware networking	On top of P2P with Low power pre-association discovery and Information sharing, Get a pop up if the information matches then proceed through normal connection (P2P/AP)
802.11ah	PHY Changes	sub 1GHz < 1GHz (excluding TV) spectrum and 1,2,4,8,16 MHz bandwidths, over 1KM distances and 300 Kbps Low Power Large STA association Improvised legacy Powersave	Wifi Docking	P2P based Connecting Multiple peripheral using wifi
802.11ad	PHY Changes	Gigabit Wifi, 60 GHz Unlicensed	WNM Powersave	Max Idle Periods, Proxy ARP, WoW, Convert Multicast to Unicast (Direct/Flexible)
802.11af	PHY Changes	Wifi TV: SUB 1GHz TV band Optional AC for spectrum sense	Wifi Smart Grid	Smart Energy Profiles 2.0 compatible with non-wifi as well eg. Zigbee.
802.11ae	MAC Changes	Prioritization of Management frames
802.11ai	MAC Changes	Fast initial link setup: < 100ms data connection Lesser Management Frames

In short, we have got all the three tracks covered IEEE is working on huge

                     Major PHY changes : 

• Spectrum     : 3.7GHz, sub 1GHz, 60GHz
• Bandwidth   : 1, 2,4,8 and 16 MHz
• Low power   : Energey Efficient
• FSN OFDM : 64 subcarriers (Fixed Subcarrier Number)

                    Minor MAC changes 

• All advanced WFA programs are P2P based.
• Fast Discovery
• Fast Connection, 
• Efficient and Long powersave, 
• QoS for Management,

where as WFA is working on IOT and End-User scenarios and applications.

So for coming 5 years i would say wifi has lots of work to do. Especially the standard 802.11ah and ad are very popular among engineers and companies and are highest attended TG's.

Any more new ideas/use cases for wifi ? Please share your opinion.

MU-MIMO and 802.11n/ac:An User Level Overview

Multi User MIMO, the technology introduced first in 802.11ac (By first means in to the IEEE std's) and is known to increase the spectral efficiency of the wifi channel by using multiple antennas for multiple recipients in the Downlink.

When a 11n/11ac AP with 4 antennae transmitting to a smart phones with typically a single antenna, the AP is forced to use single antenna only, causing high spectral wastage. Now imagine, if we can make use of the other 3 antennas to transmit data to other connected clients while antenna1 is transmitting to STA1? 

Here's a quote from the 802.11-ac-draft-4.0:

The support for VHT transmit beamforming sounding and VHT MU PPDUs in a VHT AP and more than oneVHT STA within a VHT BSS enables the optional use of DL-MU-MIMO. 

With DL-MU-MIMO the AP can create up to four A-MPDUs each carrying MPDUs destined for an associated MU capable STA. The AP uses group identifiers (GIDs) to signal potential recipient STAs. 

The AP transmits the A-MPDUs simultaneously in separate space-time streams such that each recipient STA is able to demodulate the space-time streams carrying its A-MPDU. 

The simultaneous transmission of A-MPDUs in a single VHT MU PPDU provides a means to increase aggregate throughput over that which would be achieved by sending the A-MPDUs in separateSU PPDUs.

That's MU-MIMO causing us a higher DL throughput even when there are many STA's connected to AP. Now for a technical mind, we shall discuss the below queries in the upcoming 802.11ac article.

a) How the CSMA/CA works here?

b) How the ACK procedure works? 

c) Why TxBF is mandatory for DL-MU-MIMO

d) Can we explot this for Multi AP and Multi Channel COncurrency...so on

Here's a nice video demonstration by qualcomm R&D, Enjoy.

View More Qualcomm Videos

OFDM in WLAN: 802.11n: Similar to 802.11a and 802.11g with minor changes

In the previous article we have discussed about OFDM in 802.11 a/g, now let's take a look at how OFDM works in 11n. 

OFDM has little to do with the drastic 11n rate boost, MIMO is the key player there. Lets leave the MIMO part aside, its a candidate for next article. As far as OFDM is concerned it will work per antenna, but the basic concept per antenna is the same as 11g with few changes.

20 MHz:  This is same as 802.11 a/g but 

a) With an extra 4 sub carriers.

In 802.11a/g we use 48 sub-carriers for data and 4 sub-carriers for pilot. But in 11n we make use of an extra 4 sub-carriers from the reserved for data sub-carriers.So this results in increased throughput.

b) Short Guard Interval (Optional)

we have short guard interval introduced in 11nIn the previous article we have take the guard interval as 1/4 of symbol time, now in 11n they have still decreased the guard interval to 1/8 of the symbol time. So it comes down to,                          3.2us *1/8=0.4us 

Lets take an example to understand:

a) Full Guard Interval

20MHz/64=0.3125; 1/0.3125 =3.2us + 1/4 * 3.2 us=4 us

b) Short guard Interval

20MHz/64=0.3125; 1/0.3125 =3.2us + 1/8 * 3.2 us=3.6 us

Note: Important point to note here is that while doing the calculations for the FFT period we have considered the maximum carriers 64 not the used carriers, so the above calculation of 3.2us holds good in spite of increased sub carriers.

40 MHz: 

Its same as 20MHz but the process repeated in both primary and secondary channels.

So as derived above the total no of sub-carriers in  a 40MHz channel is 52*2 (primary and secondary) + 4 (extra).

The reason for these extra 4 data sub carriers is that we don't need that many (8) pilot sub-carriers combined for both the channels, so we converted some of them to data sub carriers.

Data Sub-Carriers = 52 *2 = 104

Pilot sub-carriers  = 4 + 4  = 8

======================

Total Sub-carriers = 112 with 8 pilot sub-carriers.

Now out of that if remove 2 sub-carriers from 8 and convert them to 2 and add 2 more (why?) 

======================

Data Sub-carriers = 104+2 (from pilot)+2 (extra)=108

Pilot sub-carriers = 4-1 + 4-1=6

======================

Total sub-carriers = 108+6 =114

I am not sure of the exact rationale behind adding some number of sub carriers  but its safe to attribute that to the evolution, as the IEEE std progresses they are trying to make use of all the reserved sub-carriers in that process.

With the same logic in the previous article we can similarly derive the rate calculations for the above 20MHz and 40Mhz as well. The above tables from the IEEE 802.11-2012 will help you.