Multicast Considerations over IEEE 802 Wireless Media

Performance issues have been observed when multicast packet transmissions of IETF protocols are used over IEEE 802 wireless media. Even though enhamcements for multicast transmissions have been designed at both IETF and IEEE 802, incompatibilities still exist between specifications, implementations and configuration choices. Many IETF protocols depend on multicast/broadcast for delivery of control messages to multiple receivers. Multicast is used for various purposes such as neighborhood discovery, network flooding, address resolution, as well minimizing media occupancy for the transmission of data that is intended for multiple receivers. In addition to protocol use of broadcast/multicast for control messages, more applications, such as push to talk in hospitals, video in enterprises and lectures in Universities, are streaming over wifi. Many types of end devices are increasingly using wifi for their connectivity. IETF protocols typically rely on network protocol layering in order to reduce or eliminate any dependence of higher level protocols on the specific nature of the MAC layer protocols or the physical media. In the case of multicast transmissions, higher level protocols have traditionally been designed as if transmitting a packet to an IP address had the same cost in interference and network media access, regardless of whether the destination IP address is a unicast address or a multicast or broadcast address. This model was reasonable for networks where the physical medium was wired, like Ethernet. Unfortunately, for many wireless media, the costs to access the medium can be quite different. Multicast over wifi has often been plagued by such poor performance that it is disallowed. Some enhancements have been designed in IETF protocols that are assumed to work primarily over wireless media. However, these enhancements are usually implemented in limited deployments and not widespread on most wireless networks. IEEE 802 wireless protocols have been designed with certain features to support multicast traffic. For instance, lower modulations are used to transmit multicast frames, so that these can be received by all stations in the cell, regardless of the distance or path attenuation from the base station or access point. However, these lower modulation transmissions occupy the medium longer; they hamper efficient transmission of traffic using higher order modulations to nearby stations. For these and other reasons, IEEE 802 working groups such as 802.11 have designed features to improve the performance of multicast transmissions at Layer 2 . In addition to protocol design features, certain operational and configuration enhancements can ameliorate the network performance issues created by multicast traffic. as described in . In discussing these issues over email, and in a side meeting at IETF 99, it has been generally agreed that these problems will not be fixed anytime soon primarily because it's expensive to do so and multicast is unreliable. A big problem is that multicast is somewhat a second class citizen, to unicast, over wifi. There are many protocols using multicast and there needs to be something provided in order to make them more reliable. The problem of IPv6 neighbor discovery saturating the wifi link is only part of the problem. Wifi traffic classes may help. We need to determine what problem should be solved by the IETF and what problem should be solved by the IEEE (see ). A "multicast over wifi" IETF mailing list has been formed (mcast-wifi@ietf.org) for further discussion. This draft will be updated according to the current state of discussion. This document details various problems caused by multicast transmission over wireless networks, including high packet error rates, no acknowledgements, and low data rate. It also explains some enhancements that have been designed at IETF and IEEE 802 to ameliorate the effects of multicast traffic. Recommendations are also provided to implementors about how to use and combine these enhancements. Some advice about the operational choices that can be taken is also included. It is likely that this document will also be considered relevant to designers of future IEEE wireless specifications.

This document uses the following definitions: IEEE 802.11 Access Point. The "lowest common denominator" data rate at which multicast and broadcast traffic is generally transmitted. Delivery Traffic Indication Map (DTIM): An information element that advertises whether or not any associated stations have buffered multicast or broadcast frames. Modulation and Coding Scheme. 802.11 station (e.g. handheld device). Traffic Indication Map (TIM): An information element that advertises whether or not any associated stations have buffered unicast frames.

In this section we describe some of the issues related to the use of multicast transmissions over IEEE 802 wireless technologies.

Multicast traffic is typically much less reliable than unicast traffic. Since multicast makes point-to-multipoint communications, multiple acknowledgements would be needed to guarantee reception at all recipients. Since typically there are no ACKs for multicast packets, it is not possible for the Access Point (AP) to know whether or not a retransmission is needed. Even in the wired Internet, this characteristic often causes undesirably high error rates. This has contributed to the relatively slow uptake of multicast applications even though the protocols have long been available. The situation for wireless links is much worse, and is quite sensitive to the presence of background traffic. Consequently, there can be a high packet error rate (PER) due to lack of retransmission, and because the sender never backs off. It is not uncommon for there to be a packet loss rate of 5% or more, which is particularly troublesome for video and other environments where high data rates and high reliability are required.

One big difference between multicast over wired versus multicast over wired is that transmission over wired links often occurs at a fixed rate. Wifi, on the other hand, has a transmission rate which varies depending upon the client's proximity to the AP. The throughput of video flows, and the capacity of the broader wifi network, will change and will impact the ability for QoS solutions to effectively reserve bandwidth and provide admission control. For wireless stations associated with an Access Points, the power necessary for good reception can vary from station to station. For unicast, the goal is to minimize power requirements while maximizing the data rate to the destination. For multicast, the goal is simply to maximize the number of receivers that will correctly receive the multicast packet; generally the Access Point has to use a much lower data rate at a power level high enough for even the farthest station to receive the packet. Consequently, the data rate of a video stream, for instance, would be constrained by the environmental considerations of the least reliable receiver associated with the Access Point. Because more robust modulation and coding schemes (MCSs) have longer range but also lower data rate, multicast / broadcast traffic is generally transmitted at the lowest common denominator rate, also known as the basic rate. The amount of additional interference depends on the specific wireless technology. In fact backward compatibility and multi-stream implementations mean that the maximum unicast rates are currently up to a few Gb/s, so there can be a more than 3 orders of magnitude difference in the transmission rate between the basic rates to optimal unicast forwarding. Some techinues employed to increase spectral efficiency, such as spatial multiplexing in mimo systems, are not available with more than one intended reciever; it is not the case that backwards compatibility is the only factor responsible for lower multicast transmission rates. Wired multicast also affects wireless LANs when the AP extends the wired segment; in that case, multicast / broadcast frames on the wired LAN side are copied to WLAN. Since broadcast messages are transmitted at the most robust MCS, many large frames are sent at a slow rate over the air.

Transmissions at a lower rate require longer occupancy of the wireless medium and thus take away from the airtime of other communications and degrade the overall capacity. Furthermore, transmission at higher power, as is required to reach all multicast clients associated to the AP, proportionately increases the area of interference.

One of the characteristics of multicast transmission is that every station has to be configured to wake up to receive the multicast, even though the received packet may ultimately be discarded. This process can have a large effect on the power consumption by the multicast receiver station. Multicast can work poorly with the power-save mechanisms defined in IEEE 802.11e, for the following reasons. Clients may be unable to stay in sleep mode due to multicast control packets frequently waking them up. Both unicast and multicast traffic can be delayed by power-saving mechanisms. A unicast packet is delayed until a STA wakes up and requests it. Unicast traffic may also be delayed to improve power save, efficiency and increase probability of aggregation. Multicast traffic is delayed in a wireless network if any of the STAs in that network are power savers. All STAs associated to the AP have to be awake at a known time to receive multicast traffic. Packets can also be discarded due to buffer limitations in the AP and non-AP STA.

This section identifies some representative IETF protocols, and describes possible negative effects due to performance degradation when using multicast transmissions for control messages. Common uses of multicast include: Control plane signaling Neighbor Discovery Address Resolution Service discovery Applications (video delivery, stock data, etc.) On-demand routing Backbone construction Other L3 protocols (non-IP) User Datagram Protocol (UDP) is the most common transport layer protocol for multicast applications. By itself, UDP is not reliable -- messages may be lost or delivered out of order.

The following list contains a few representative IPv4 protocols using multicast. ARP DHCP mDNS After initial configuration, ARP and DHCP occur much less commonly, but service discovery can occur at any time. Apple's Bonjour protocol, for instance, provides service discovery (for printing) that utilizes multicast. It's often the first service that operators drop. Even if multicast snooping is utilized, many devices can register at once using Bonjour, causing serious network degradation.

IPv6 makes extensive use of multicast, including the following: DHCPv6 IPv6 Neighbor Discovery Protocol (NDP) Duplicate Address Detection (DAD) Address Resolution Service Discovery Route Discovery Decentralized Address Assignment Geographic routing IPv6 NDP Neighbor Solicitation (NS) messages used in DAD and Address Lookup make use of Link-Scope multicast. In contrast to IPv4, an IPv6 Node will typically use multiple addresses, and may change them often for privacy reasons. This multiplies the impact of multicast messages that are associated to the mobility of a Node. Router advertisement (RA) messages are also periodically multicasted over the Link. IPv6 NDP Neighbor Solicitation (NS) messages used in DAD and Address Lookup make use of Link-Scope multicast. In contrast to IPv4, an IPv6 Node will typically use multiple addresses, and may change them often for privacy reasons. This multiplies the impact of multicast messages that are associated to the mobility of a Node. Router advertisement (RA) messages are also periodically multicasted over the Link. Neighbors may be considered lost if several consecutive Neighbor Discovery packets fail.

Multicast Listener Discovery(MLD) is often used to identify members of a multicast group that are connected to the ports of a switch. Forwarding multicast frames into a WiFi-enabled area can use such switch support for hardware forwarding state information. However, since IPv6 makes heavy use of multicast, each STA with an IPv6 address will require state on the switch for several and possibly many multicast solicited-node addresses. Multicast addresses that do not have forwarding state installed (perhaps due to hardware memory limitations on the switch) cause frames to be flooded on all ports of the switch.

On the Internet there is a "background radiation" of scanning traffic (people scanning for vulnerable machines) and backscatter (responses from spoofed traffic, etc). This means that routers very often receive packets destined for machines whose IP addresses may or may not be in use. In the cases where the IP is assigned to a host, the router broadcasts an ARP request, gets back an ARP reply, and caches it; then traffic can be delivered to the host. When the IP address is not in use, the router broadcasts one (or more) ARP requests, and never gets a reply. This means that it does not populate the ARP cache, and the next time there is traffic for that IP address the router will rebroadcast the ARP requests. The rate of these ARP requests is proportional to the size of the subnets, the rate of scanning and backscatter, and how long the router keeps state on non-responding ARPs. As it turns out, this rate is inversely proportional to how occupied the subnet is (valid ARPs end up in a cache, stopping the broadcasting; unused IPs never respond, and so cause more broadcasts). Depending on the address space in use, the time of day, how occupied the subnet is, and other unknown factors, on the order of 2000 broadcasts per second have been observed at the IETF NOCs. On a wired network, there is not a huge difference between unicast, multicast and broadcast traffic. Due to hardware filtering (see, e.g., ), inadvertently flooded traffic (or high amounts of ethernet multicast) on wired networks can be quite a bit less costly, compared to wireless cases where sleeping devices have to wake up to process packets. Wired Ethernets tend to be switched networks, further reducing interference from multicast. There is effectively no collision / scheduling problem except at extremely high port utilizations. This is not true in the wireless realm; wireless equipment is often unable to send high volumes of broadcast and multicast traffic. Consequently, on the wireless networks, we observe a significant amount of dropped broadcast and multicast packets. This, in turn, means that when a host connects it is often not able to complete DHCP, and IPv6 RAs get dropped, leading to users being unable to use the network.

This section lists some optimizations that have been specified in IEEE 802 and IETF that are aimed at reducing or eliminating the issues discussed in .

The AP knows the MAC address and IP address for all associated STAs. In this way, the AP acts as the central "manager" for all the 802.11 STAs in its BSS. Proxy ARP is easy to implement at the AP, and offers the following advantages: Reduced broadcast traffic (transmitted at low MCS) on the wireless medium STA benefits from extended power save in sleep mode, as ARP requests for STA's IP address are handled instead by the AP. ARP frames are kept off the wireless medium. No changes are needed to STA implementation. Here is the specification language as described in clause 10.23.13 of : When the AP supports Proxy ARP "[...] the AP shall maintain a Hardware Address to Internet Address mapping for each associated station, and shall update the mapping when the Internet Address of the associated station changes. When the IPv4 address being resolved in the ARP request packet is used by a non-AP STA currently associated to the BSS, the proxy ARP service shall respond on behalf of the non-AP STA"

As used in this section, a Low-Power Wireless Personal Area Network (6LoWPAN) denotes a low power lossy network (LLN) that supports 6LoWPAN Header Compression (HC). A 6TiSCH network is an example of a 6LowPAN. In order to control the use of IPv6 multicast over 6LoWPANs, the 6LoWPAN Neighbor Discovery (6LoWPAN ND) standard defines an address registration mechanism that relies on a central registry to assess address uniqueness, as a substitute to the inefficient Duplicate Address Detection (DAD) mechanism found in the mainstream IPv6 Neighbor Discovery Protocol (NDP) . The 6lo Working Group has specified an update to RFC6775. Wireless devices can register their address to a Backbone Router, which proxies for the registered addresses with the IPv6 NDP running on a high speed aggregating backbone. The update also enables a proxy registration mechanism on behalf of the registered node, e.g. by a 6LoWPAN router to which the mobile node is attached. The general idea behind the backbone router concept is that broadcast and multicast messaging should be tightly controlled in a variety of Wireless Local Area Networks (WLANs) and Wireless Personal Area Networks (WPANs). Connectivity to a particular link that provides the subnet should be left to Layer-3. The model for the Backbone Router operation is represented in .

LLN nodes can move freely from an LLN anchored at one IPv6 Backbone Router to an LLN anchored at another Backbone Router on the same backbone, keeping any of the IPv6 addresses they have configured. The Backbone Routers maintain a Binding Table of their Registered Nodes, which serves as a distributed database of all the LLN Nodes. An extension to the Neighbor Discovery Protocol is introduced to exchange Binding Table information across the Backbone Link as needed for the operation of IPv6 Neighbor Discovery. RFC6775 and follow-on work (e.g., , do address the needs of LLNs, and similar techniques are likely to be valuable on any type of link where sleeping devices are attached, or where the use of broadcast and multicast operations should be limited.

Methods have been developed to help save battery life; for example, a device might not wake up when the AP receives a multicast packet. The AP acts on behalf of STAs in various ways. To enable use of the power-saving feature for STAs in its BSS, the AP buffers frames for delivery to the STA at the time when the STA is scheduled for reception. If an AP, for instance, expresses a DTIM (Delivery Traffic Indication Message) of 3 then the AP will send a multicast packet every 3 packets. In fact, when any single wireless client associated with an access point has 802.11 power-save mode enabled, the access point buffers all multicast frames and sends them only after the next DTIM beacon. But in practice, most AP's will send a multicast every 30 packets. For unicast there's a TIM (Traffic Indication Message); but since multicast is going to everyone, the AP sends a broadcast to everyone. DTIM does power management but clients can choose whether or not to wake up or not and whether or not to drop the packet. Unfortunately, without proper administrative control, such clients may no longer be able to determine why their multicast operations do not work.

IPv6 uses Neighbor Discovery Protocol (NDP) instead of ARP. Every IPv6 node subscribes to a special multicast address for this purpose. Here is the specification language from clause 10.23.13 of : "When an IPv6 address is being resolved, the Proxy Neighbor Discovery service shall respond with a Neighbor Advertisement message [...] on behalf of an associated STA to an [ICMPv6] Neighbor Solicitation message [...]. When MAC address mappings change, the AP may send unsolicited Neighbor Advertisement Messages on behalf of a STA." NDP may be used to request additional information Maximum Transmission Unit Router Solicitation Router Advertisement, etc. NDP messages are sent as group addressed (broadcast) frames in 802.11. Using the proxy operation helps to keep NDP messages off the wireless medium.

It is often possible to transmit multicast control and data messages by using unicast transmissions to each station individually.

There are situations where more is needed than simply converting multicast to unicast. For these purposes, DMS enables a client to request that the AP transmit multicast group addressed frames destined to the requesting clients as individually addressed frames [i.e., convert multicast to unicast]. Here are some characteristics of DMS: Requires 802.11n A-MSDUs Individually addressed frames are acknowledged and are buffered for power save clients The requesting STA may specify traffic characteristics for DMS traffic DMS was defined in IEEE Std 802.11v-2011 DMS requires changes to both AP and STA implementation. DMS is not currently implemented in products. See and for more information.

GCR (defined in ) provides greater reliability by using either unsolicited retries or a block acknowledgement mechanism. GCR increases probability of broadcast frame reception success, but still does not guarantee success. For the block acknowledgement mechanism, the AP transmits each group addressed frame as conventional group addressed transmission. Retransmissions are group addressed, but hidden from non-11aa clients. A directed block acknowledgement scheme is used to harvest reception status from receivers; retransmissions are based upon these responses. GCR is suitable for all group sizes including medium to large groups. As the number of devices in the group increases, GCR can send block acknowledgement requests to only a small subset of the group. GCR does require changes to both AP and STA implementation. GCR may introduce unacceptable latency. After sending a group of data frames to the group, the AP has do the following: unicast a Block Ack Request (BAR) to a subset of members. wait for the corresponding Block Ack (BA). retransmit any missed frames. resume other operations which may have been delayed. This latency may not be acceptable for some traffic. There are ongoing extensions in 802.11 to improve GCR performance. BAR is sent using downlink MU-MIMO (note that downlink MU-MIMO is already specified in 802.11-REVmc 4.3). BA is sent using uplink MU-MIMO (which is a .11ax feature). Additional 802.11ax extensions are under consideration; see Latency may also be reduced by simultaneously receiving BA information from multiple clients.

This section lists some operational optimizations that can be implemented when deploying wireless IEEE 802 networks to mitigate the issues discussed in .

An ARP Sponge sits on a network and learn which IPs addresses are actually in use. It also listen for ARP requests, and, if it sees an ARP for an IP address which it believes is not used, it will reply with its own MAC address. This means that the router now has an IP to MAC mapping, which it caches. If that IP is later assigned to an machine (e.g using DHCP), the ARP sponge will see this, and will stop replying for that address. Gratuitous ARPs (or the machine ARPing for its gateway) will replace the sponged address in the router ARP table. This technique is quite effective; but, unfortunately, the ARP sponge daemons were not really designed for this use (the standard one , was designed to deal with the disappearance of participants from an IXP) and so are not optimized for this purpose. We have to run one daemon per subnet, the tuning is tricky (the scanning rate versus the population rate versus retires, etc.) and sometimes the daemons just seem to stop, requiring a restart of the daemon and causing disruption. Some routers (often those based on Linux) implement a "negative ARP cache" daemon. Simply put, if the router does not see a reply to an ARP it can be configured to cache this information for some interval. Unfortunately, the core routers which we are using do not support this. When a host connects to network and gets an IP address, it will ARP for its default gateway (the router). The router will update its cache with the IP to host MAC mapping learnt from the request (passive ARP learning). The distribution of users on wireless networks / subnets changes from meeting to meeting (e.g the "IETF-secure" SSID was renamed to "IETF", fewer users use "IETF-legacy", etc). This utilization is difficult to predict ahead of time, but we can monitor the usage as attendees use the different networks. By configuring multiple DHCP pools per subnet, and enabling them sequentially, we can have a large subnet, but only assign addresses from the lower portions of it. This means that we can apply input IP access lists, which deny traffic to the upper, unused portions. This means that the router does not attempt to forward packets to the unused portions of the subnets, and so does not ARP for it. This method has proven to be very effective, but is somewhat of a blunt axe, is fairly labor intensive, and requires coordination. In general, the router does not need to ARP for hosts; when a host connects, the router can learn the IP to MAC mapping from the ARP request sent by that host. This means that we should be able to disable and / or filter ARP requests from the router. Unfortunately, ARP is a very low level / fundamental part of the IP stack, and is often offloaded from the normal control plane. While many routers can filter layer-2 traffic, this is usually implemented as an input filter and / or has limited ability to filter output broadcast traffic. This means that the simple "just disable ARP or filter it outbound" seems like a really simple (and obvious) solution, but implementations / architectural issues make this difficult or awkward in practice. The broadcasts are overwhelmingly being caused by outside scanning / backscatter traffic. This means that, if we were to NAT the entire (or a large portion) of the attendee networks, there would be no NAT translation entries for unused addresses, and so the router would never ARP for them. The IETF NOC has discussed NATing the entire (or large portions) attendee address space, but a: elegance and b: flaming torches and pitchfork concerns means we have not attempted this yet. Another obvious solution would be to put a stateful firewall between the wireless network and the Internet. This firewall would block incoming traffic not associated with an outbound request. The IETF philosophy has been to have the network as open as possible / honor the end-to-end principle. An attendee on the meeting network should be an Internet host, and should be able to receive unsolicited requests. Unfortunately, keeping the network working and stable is the first priority and a stateful firewall may be required in order to achieve this.

Many of the causes of performance degradation described in earlier sections are also observable for wireless media other than 802.11. For instance, problems with power save, excess media occupancy, and poor reliability will also affect 802.15.3 and 802.15.4. Unfortunately, 802.15 media specifications do not yet include mechanisms similar to those developed for 802.11. In fact, the design philosophy for 802.15 is oriented towards minimality, with the result that many such functions are relegated to operation within higher layer protocols. This leads to a patchwork of non-interoperable and vendor-specific solutions. See for some additional discussion, and a proposal for a task group to resolve similar issues, in which the multicast problems might be considered for mitigation.

This section will provide some recommendations about the usage and combinations of the multicast enhancements described in and . Future protocol documents utilizing multicast signaling should be carefully scrutinized if the protocol is likely to be used over wireless media. Proxy methods should be encouraged to conserve network bandwidth and power utilization by low-power devices. The device can use a unicast message to its proxy, and then the proxy can take care of any needed multicast operations. Multicast signaling for wireless devices should be done in a way compatible with low-duty cycle operation. (FFS)

This section suggests two discussion items for further resolution. The IETF should determine guidelines by which it may be decided that multicast packets are to be sent wired. For example, 802.1ak works on ethernet and wifi. 802.1ak has been pulled into 802.1Q as of 802.1Q-2011. 802.1Q-2014 can be found here: http://www.ieee802.org/1/pages/802.1Q-2014.html. If a generic solution is not found, guidelines for multicast over wifi should be established. Perhaps a reliable registration to Layer-2 multicast groups and a reliable multicast operation at Layer-2 could provide a generic solution. There is no need to support 2^24 groups to get solicited node multicast working: it is possible to simply select a number of trailing bits that make sense for a given network size to limit the amount of unwanted deliveries to reasonable levels. IEEE 802.1, 802.11, and 802.15 should be encouraged to revisit L2 multicast issues. In reality, Wi-Fi provides a broadcast service, not a multicast service. On the physical medium, all frames are broadcast except in very unusual cases in which special beamforming transmitters are used. Unicast offers the advantage of being much faster (2 orders of magnitude) and much more reliable (L2 ARQ).

This document does not introduce any security mechanisms, and does not have affect existing security mechanisms.

This document does not request any IANA actions.

This document has benefitted from discussions with the following people, in alphabetical order: Pascal Thubert