WiMAX Worldwide Interoperability for Microwave Access Marina Yankelovich, Hadar Hen-zion, Michael Reznik

Why WiMAX? Over the past years we witness growing demand for Voice, Multimedia and Internet Access. The problem is that running wired communication infrastructure is expensive and complicated. The answer is broadband wireless. Erecting a big antenna on hill just outside of town is much easier and cheaper.

Why WiMAX? WiMAX is a MAN (Metropolitan Aria Network) protocol that will be a wireless alternative to DSL and TI level services for last mile broadband access. Internet/ Telephone Network Competing telecommunication companies have a great interest in providing a multimegabit wireless communication service for voice, internet movies on demand, etc.

WiFi vs. WiMAX 802.16 provides service to building, unlike 802.11 that was designed to be mobile Ethernet. Wi-Fi has a range of about 100m. WiMAX provide services over several Kms.

WiFi vs. WiMAX WiFi operates using CSMA/CA (Carrier Sense Multiple Access with Collision Avoidance) in its MAC layer while WiMAX uses a scheduler at the BSs (base stations) that allows better utilization of the bandwidth. Another issue is Qos. While 802.11 provides some support for real-time traffic , it was not really designed for telephony and heavy-duty multimedia usage.

WiFi vs. WiMAX In contrast, 802.16 is expected to support these applications completely because it is intended for residential as well as business use. Wi-Fi channels occupy a fixed width of the spectrum while WiMAX allow channel sizes to be decided based on requirements.

Scope of 802 Standard

WiMAX Phisycal Features The physical layers, as named in the standard, are: SC – A Single carrier physical layer used for 10-66 GHz and 2-11 GHz frequencies. OFDM – An OFDM physical layer used for 2-11GHz frequencies. OFDMA – An OFDMA physical layer used for 2-11GHz frequencies. All the above features, support both TDD and FDD . The standard defines all of this. However, all the WiMax networks used today (Seoul, Tokyo, Portland, Washington, Moscow,…) are OFDMA based (i.e. – there isn’t today, and will not be a SC or OFDM WiMax network).

OFDM OFDM is a very powerful transmission technology that subdivides the bandwidth into multiple frequency sub- carriers. Each sub-carrier is actually a narrow-band signal that is orthogonal to others signals like it. All the signals are transmitted simultaneously without being able to interrupt each other due to their orthogonality. Together they use the same bandwidth as one single-carrier signal.

OFDMA Originally the OFDM was designed for single signal transmission which was divided to sub-carriers. To realize the multiple user access a TDMA or FDMA scheme has to be added, allowing different users to use the same OFDM signal and turning it to OFDMA. The sub-carriers are divided into sets creating sub-channels.

OFDMA Now the system has to coordinate the time and the frequency allocations to transmit data that belongs to different users. Using sub-channels adds the frequency dimension to the scheduling problem-allocation of the channel resources.

Duplexing Scheme in WiMAX Duplexing refers to the way downlink and uplink data is arranged in a two-way wireless transmission. The downlink carries information from a Base Station (BS) to Subscriber Stations (SSs). The uplink carries information from a SS to a BS.

Duplexing Scheme in WiMAX There are two types of duplexing scheme: FDD and TDD. FDD (Frequency division duplexing) - divides the bandwidth into two channels and if one of them is temporarily empty of data, the other uses only half of the allocated resources. TDD (Time division duplexing) - achieves the full-duplex communication goal by dividing the time dimension while sharing the same frequency. TDD on the other hand, may easily change the downlink-uplink ratio to adapt itself to the traffic needs.

Duplexing Scheme in WiMAX An example of a TDD frame that is always divided to pair of sub-frames, the downlink followed by the uplink, having a guard interval between them.

TDD Frame Structure in WiMAX It should be noted that in the downlink direction one sub-channel may be intended for group of receivers, while in the uplink several sub-channel can be allocated to the same transmitter. FCH=frame control header

The MAC Sublayer The WiMAX base station schedules uplink (UL) and downlink (DL) packets wise. The scheduling is done with respect to priority of services. The scheduling information is broadcasted by the BS through the uplink map message (UL-MAP) at the beginning of each frame. After receiving the UL-MAP, each SS will transmit data in the predefined time slots.

The MAC sublayer The BS uplink scheduling module allocates bandwidth according to BW request sent from SSs to BS. In IEEE 802.16 standard, there are two modes of transmitting the BW-Request: contention mode and contention-free mode (polling). IEEE 802.16 defines four types of service flows, each with different QoS requirements and corresponding uplink scheduler policy. In contention mode, SSs send BW-Request during the contention period. Contention is resolved using back-off resolution. In contention-free mode, BS polls each SS and SSs reply by sending BW-request. Due to the predictable signaling delay of the polling scheme, contention-free mode is suitable for real time applications.

Types of service flow Unsolicited grant service (UGS)- this service supports constant bit-rate flows such as Voice over IP. These applications require constant bandwidth allocation. BW-Request: Not required. The four services are supported by the IEEE 802.16 standard.

Types of service flow Real-time polling service (rtPS)- this service is for real-time flows such as MPEG video. These applications have specific bandwidth requirements as well as a deadline. Late packets that miss the deadline will be useless. BW-Request: used only in the contention-free mode. The current queue size represents the current bandwidth demand and included in the BW-Request.

Types of service flow Non-real-time polling service (nrtPS)- this service is for non-real-time flows which require better than best effort service, e.g. bandwidth intensive file transfer. These applications are time-insensitive and require minimum bandwidth allocation. BW-request: uses either contention-free mode or contention mode. Current queue size is included in BW-request.

Types of service flow Best effort service (BE)- this service is for best effort traffic such as HTTP. There is no QoS guarantee. The applications in this service flow receive the available bandwidth after the bandwidth is allocated to the previous three service flows. BW-request: uses only contention mode. Current queue size is included in BW request.

The MAC sublayer scheduler IEEE 802.16 standard left the QoS based packet scheduling algorithms, that determine the uplink and downlink bandwidth allocation, undefined. In recent years, several schedulers were published. We would like to present a basic uplink scheduler.

Uplink Scheduler Article “Packet scheduling for QoS support in IEEE 802.16 broadband wireless access systems” INTERNATIONAL JOURNAL OF COMMUNICATION SYSTEMS By Kitti Wongthavarawatn and Aura Ganzz מעכשיו נתמקד באלגוריתמים אשר הוצעו במאמר.

Uplink Scheduler Proposed Module Uplink packet scheduling (UPS) resides in the BS to control all the uplink packet transmissions.

Uplink Scheduler Proposed Module The article suggest: At the BS A detailed description of the uplink packet scheduling module. Admission control module. At the SS The Traffic Policing module. The admission control and traffic policing are mandatory but not defined in the standard as well as the scheduler.

Traffic Policing-Token bucket A token bucket is a common algorithm used to control the amount of data that is injected into a network. The main idea of Token Bucket is to permit transmission of data burst, yet to control the maximum transmission rate. Packets arriving cannot be transmitted if a token is not available in the bucket. Tokens are filled in the bucket according to a specific rate defined by the algorithm. The tokens arrival rate equals the average transmission rate.

Token Bucket illustration Based on the presence of tokens in the bucket - an abstract container that holds aggregate network traffic to be transmitted. Each token represents a unit of bytes or a single packet of predetermined size. Tokens in the bucket are removed for the ability to send a packet.

Admission control module Admission control is the QoS mechanism that decides whether a new session (connection) can be established. This mechanism will ensure that existing sessions’ QoS will not be degraded and the new session will be provided QoS support. A connection is admitted if: (1) there is enough bandwidth to accommodate the new connection, (2) the newly admitted connection will receive QoS guarantees in terms of both bandwidth and delay and (3) QoS of existing connections is maintained.

Admission control for rtPS connections The admission control policy for rtPS connections is given by the following algorithm: Input: A new rtPS connection requests with parameters: bi - token bucket size of connection i. ri - token bucket rate (average data rate) of connection i. di - maximum delay requirement of connection i (ms). Current network parameters: average capacity (bps) allocated for all current connections. total capacity (bps) allocated for uplink transmission. f = duration of a time frame (ms) which includes uplink and downlink subframes.

Example for Admission control procedure Check for available bandwidth: If ri ≤ Cuplink-CUGS-CrtPS –CnrtPS, there is available bandwidth for the new rtPS connection. Otherwise reject the new connection. Check for delay guarantees: If we can provide delay guarantees for the new rtPS connection. Otherwise reject the new connection. Check for delay violations of existing rtPS connections: If for any connection this condition is not satisfied, reject the new connection. Update CrtPS: CrtPS←CrtPS+ ri. Pass token bucket parameters ri, bi to the traffic policing module.

The proposed scheduler One of the modes of uplink arbitration uses a TDMA MAC. The BS determines the number of time slots that each SS will be allowed to transmit in an uplink subframe. המאמר מציע אלגוריתם שמקצה slots בתוך frames אך ורק במישור הזמן. חשוב לציין שטכנולוגיית הwimax מאפשרת גם חלוקה במישור התדר. הדבר מצריך אלגוריתם מתאים המתזמן slots גם בזמן וגם בתדר. המאמר לא מדבר על שילוב שכזה.

The proposed scheduler (cont.) Overall bandwidth allocation: Bandwidth allocation per flow follows strict priority, from highest to lowest: UGS, rtPS, nrtPS and BE.

The proposed scheduler (cont.) Bandwidth allocation within UGS connections: The Scheduler allocates fixed bandwidth as defined by IEEE 802.16. Bandwidth allocation within rtPS connections: Apply earliest deadline first (EDF) service discipline to this service flow. Packets with earliest deadline will be scheduled first.

The proposed scheduler (cont.) Bandwidth allocation within nrtPS connections: We apply weight fair queue (WFQ) service discipline to this service flow. We schedule nrtPS packets based on the weight of the connection. Bandwidth allocation within BE connections: The remaining bandwidth is equally allocated to each BE connection.

Service Assignment The service assignment module determines the uplink subframe allocation in terms of the number of bits per connection. The number of bits will eventually be converted to the number of time slots which are the units used in the UL-MAP. The number of bits per time slot is determined by the physical layer of the wireless network.

Assume: current time = t Service Assignment UGS packet scheduling - schedule predetermined time slots for each UGS connection. After scheduling the packets, update Nuplink: Nuplink = total number of bits that SSs are allowed to transmit in an uplink subframe. Assume: current time = t The i means the i-th connection of UGS type.

Service Assignment rtPS packet scheduling - schedule the rtPS packets until either there are no rtPS packets left or there is no more bandwidth left. After scheduling the packets, update Nuplink :

Service Assignment In case the total number of bits in the column is greater than Nuplink, Nuplink will be distributed to each connection based on its weight.

Service Assignment r i = average data rate of connection i. For example, connection i will be scheduled with W i ∙ Nuplink bits. We can take the following two actions for the packets that missed their deadline: (1) drop the packets. (2) reduce the priority of the packets by moving them to the BE database, i.e. these packets will be scheduled with the same priority as BE.

Assume: current time = t Service Assignment nrtPS packet scheduling - schedule the packets based on connections’ weights Also update the Nuplink: Assume: current time = t

Assume: current time = t Service Assignment BE packet scheduling - schedule packets equally, i.e. provide the same amount of bandwidth to each BE connection. Assume: current time = t

Simulation Results In the article, they perform the experiment that shows the QoS support provided by the uplink packet scheduling. Assumptions: There are only two types of traffic (rtPS and BE). All traffic is already admitted to the network. BE traffic requires uplink bandwidth at all times. Ctotal =10 Mbps, Cuplink = 5 Mbps, Cdownlink = 5 Mbps. (Ctotal =total capacity (bps) of the wireless network, Ctotal = Cuplink ‏+Cdownlink) Frame size (f) = 10 ms Input traffic: Three rtPS sessions with average total bandwidth (CrtPS) of 3 Mbps:

The bandwidth allocation for rtPS and BE connections. Since our UPS is work conserving and there is always BE traffic available, rtPS and BE bandwidth allocation complement each other, i.e. in each frame the total rtPS and BE bandwidth equal to 5 Mbps: In this experiment there are no packets that miss their deadline.

Arrival and service curves The graphs clearly show that the service curve adapts and follows the arrival curve. We observe that the horizontal distance between these two curves (arrival curve and service curve) is bound by the maximum delay.

Downlink Scheduler On the downlink (from BS to SS), the transmission is relatively simple because the BS is the only one that transmits during the downlink subframe. The data packets are broadcasted to all SSs and an SS only picks up the packets destined to it. As it was said the IEEE 802.16 standard doesn’t define Downlink scheduler, many algorithms were proposed, but we will not present them here.

The End Thank You