White Paper: Powerline Networking

 

System Architecture

The above figure shows the architecture of a powerline networking system (in particular a HomePlug AV system).  The Higher Layer Entities (HLEs) above the H1 (Host) Interface may be bridges, applications or servers that provide off-chip services to clients below the H1 Interface. The Data Service Access Point (SAP) accepts Ethernet format packets, so all IP based protocols are easily handled.

The Architecture defines two planes as shown in Figure 1. The data plane provides the traditional layered approach with the M1 interface between the Convergence Layer (CL) and the MAC, and the PHY interface between the MAC and the PHY. In the control plane, the MAC is a monolith without conventional layering. In Figure 1 it is labeled as the Connection Manager (CM) since that is its primary function. The approach adopted for the control plane was chosen to provide more efficient processing and to provide implementers greater flexibility for innovation. Although part of the control plane in all stations, the Central Coordinator (CCo) entity will be active in one and only one station in a single HPAV network.

 

The Physical Layer

The Physical Layer (PHY) operates in the frequency range of 2 – 28 MHz and provides a 200 Mbps PHY channel rate and a 150 Mbps information rate. It uses windowed OFDM and a powerful Turbo Convolutional Code (TCC), which provides robust performance within 0.5 dB of Shannon Capacity. Windowed OFDM provides flexible spectrum notching capability where the notches can exceed 30 dB in depth without losing significant useful spectrum outside of the notch. Long OFDM symbols with 917 usable carriers (tones) are used in conjunction with a flexible guard interval. Modulation densities from BPSK (which carries 1 bit of information per carrier per symbol) to 1024 QAM (which carries 10 bits of information per carrier per symbol) are independently applied to each carrier based on the channel characteristics between the transmitter and the receiver.

This diagram shows a block diagram representation for the physical layer of a powerline transmitter and receiver (specifically a Homeplug AV specification).

On the transmitter side, the PHY layer receives its inputs from the Medium Access Control (MAC) layer. There are separate inputs for HPAV data, HPAV control information, and HomePlug 1.0 control information (the latter in order to support HomePlug 1.0 compatibility). HPAV control information is processed by the Frame Control Encoder block, which has an embedded Frame Control FEC block and Diversity Interleaver. The HPAV data stream passes through a Scrambler, a Turbo FEC Encoder and an Interleaver. The outputs of the three streams lead into a common OFDM Modulation structure, consisting of a Mapper, an IFFT processor, Preamble and Cyclic prefix insertion and a Peak Limiter. This output eventually feeds the Analog Front End (AFE) module which couples the signal to the Powerline medium.

At the receiver, an AFE operates in conjunction with an Automatic Gain Controller (AGC) and a time synchronization module to feed separate data information and data recovery circuits. The HPAV Frame Control is recovered by processing the received stream through a 3072-point FFT, a Frame Control Demodulator and a Frame Control Decoder. The HomePlug 1.0 Frame Control, if present, is recovered by a 384-point FFT. In parallel, the data stream is retrieved after processing through a 3072-point FFT for HPAV, a demodulator with SNR estimation, a De-mapper, De-interleaver, Turbo FEC decoder, and a De-scrambler for HPAV data.

The HPAV PHY provides for the implementation of flexible spectrum policy mechanisms to allow for adaptation in varying geographic, network and regulatory environments. Frequency notches can be applied easily and dynamically, even in deployed devices. Region-specific keep-out regions can be set under software control. The ability to make soft changes to alter the device’s tone mask (enabled tones) allows for implementations that can dynamically adapt their keep-out regions.

 

MAC Protocols and Services

HPAV provides connection-oriented Contention Free (CF) service to support the QoS requirements (guaranteed bandwidth, latency and jitter requirements) of demanding AV and IP applications. This Contention Free service is based on periodic Time Division Multiple Access (TDMA) allocations of adequate duration to support the QoS requirements of a connection.

HPAV also provides a connectionless, prioritized Contention based service to support both best-effort applications and applications that rely on prioritized QoS. This service is based on Collision Sense Multiple Access/Collision Avoidance (CSMA/CA) technology which is applied to only traffic at the highest pending priority level after the pending traffic with lower priority levels has been eliminated during a brief Priority Resolution phase at the beginning of the contention window.

To efficiently provide both kinds of communication service, HPAV implements a flexible, centrally-managed architecture. The central manager is called a Central Coordinator (CCo). The CCo establishes a Beacon Period and a schedule which accommodates both the Contention Free allocations and the time allotted for Contention-based traffic. As shown in Figure 3, the Beacon Period is divided into 3 regions:

• Beacon Region
• CSMA Region
• Contention-Free Region

The Central Coordinator broadcasts a beacon at the beginning of each Beacon Period; it uses the beacon to communicate the scheduling within the beacon period. The beacons are extremely robust and reliable. The schedules advertised in the Beacon are persistent—i.e., the Central Coordinator promises not to change the schedule for a number of Beacon Periods—and the persistence is also advertised in the beacon so that the transmitting station for a connection can confidently transmit during its persistent allocation(s) even if it has missed several beacons within the advertised persistence of the schedule. This provides additional continuity even if a few beacons are missed. The CSMA periods are also persistent so that stations wishing to send CSMA traffic can do so even if they miss a few beacons.

The MAC layer provides both Contention (CSMA) and Contention Free (CF) services through the respective regions in the Beacon Period. The Central Coordinator-managed Persistent Contention Free (PCF) Region enables HPAV to provide a strict guarantee on Higher Layer Entity (HLE) QoS requirements. An HLE uses the Connection Specification (CSPEC) to specify its QoS requirements. The Connection Manager (CM) in the station evaluates the CSPEC and, if appropriate, communicates the pertinent requirements to the Central Coordinator  and asks the Central Coordinator for a suitable Contention Free allocation. QoS features specified in the CSPEC include:

• Guaranteed bandwidth
• Quasi-Error free service
• Fixed Latency
• Jitter control

If the CCo is able to accommodate the connection request, it will ask the stations to “sound” the channel. This allows the stations to perform the initial channel estimation (i.e., establish a Tone Map specifying the optimal modulation on each OFDM tone). The Tone Map is communicated from the receiver to the transmitter; the channel estimation is also communicated in abbreviated form to the CCo to help it determine how much time should be allocated to the connection. Based on the CSPEC and the channel sounding results, the CCo provides one or more persistent time allocations—Transmit Opportunities (TXOPs)—for the connection within the PCF Region.

The PCF Region also contains time for non-persistent allocations good only in the current beacon period. These non-persistent allocations are used to provide additional short term bandwidth to connections that require it (e.g., because of transient errors or changing channel conditions) to meet their QoS requirements, providing that the transmitting station hears the beacon at the beginning of the Beacon Period. When this time is not used for non-persistent CF allocations, in may be used for CSMA traffic. Again, stations must hear the beacon in order to know whether the time is available for CSMA traffic.

Messaging in HPAV is direct from station to station; however, the Central Coordinator monitors the messages. The header of each message contains information about how much data is pending for transmission on the connection; if this amount becomes large on a given connection, the Central Coordinator may allocate additional non-persistent time to the connection in the PCF Region.

The Persistent CSMA Region provides prioritized contention-based communication. It is used where there is no CSPEC and/or the traffic is of short duration. When operating in 1.0 Coexistence mode, or “Hybrid Mode”, AV coordinates with HomePlug 1.0 devices and permits them to communicate during the CSMA period.

As shown in the diagram below, the Beacon Period is synchronized to the AC line cycle. By synchronizing to the line cycle, HPAV provides stability of the periodic allocations relative to the line cycle. This, in turn, provides better channel adaptation to the synchronous (to the line cycle) interference, resulting in improved throughput. The beacon provides announcements of where the beacon will occur over the next few beacon periods—i.e., beacon persistence—to enable continued communications by stations that miss an occasional beacon.

MAC Control Plane

The Medium Access Control (MAC) Layer contains an integrated Connection Manager (CM). HLEs provide a Connection Specification (CSPEC) that details QoS requirements for application data. For bridged traffic, CSPECs may be generated dynamically by the Auto Connection Service (ACS) or by a higher layer QoS Manager that coordinates QoS over multiple network segments; otherwise the traffic is transmitted as prioritized CSMA traffic.

The Control Plane provides a seamless interface to the application layer. Application requirements are received at the H1 Control SAP in the CSPEC and are interpreted by the CM. The CM is responsible for evaluating the CSPEC and setting up the appropriate connection in conjunction with the CM in the station at the other end of the connection and with the CCo. It is the Connection Manager’s responsibility to ensure that the appropriate AV mechanisms are engaged in order to provide the application with the bandwidth it requires. It must also monitor the level of service that the connection is receiving and take remedial action if the guaranteed QoS is not being provided.

The MAC also maintains a clock that is tightly synchronized to the Central Coordinator’s clock (the Central Coordinator includes a timestamp in the beacon). This means that the entire HPAV network shares a common network clock for use by HLEs that have tight timing constraints (e.g., to synchronize surround sound speakers).

MAC Data Plane

In the Data Plane, the MAC accepts MSDUs (e.g., Ethernet packets) arriving from the Convergence Layer and encapsulates them with a header, optional Arrival Time Stamp (ATS) and Check Sum to create a MAC Frame. The MAC Frames are then en-queued into the appropriate MAC Frame Stream. It is the MAC’s responsibility to ensure that the MSDUs related to a given connection are delivered to the PHY in a timely fashion for transmission during the time allocated for the connection. For this purpose, it maintains individual queues for each connection’s data, for each priority level of CSMA traffic and for each priority level of Control Messages.

Each MAC frame stream is divided into 512 octet segments each of which is encrypted and encapsulated into a serialized PHY Block (PB). As shown in Figure 4, the PBs are packed into an MPDU which is delivered to the PHY. The PHY transmitter applies forward error correction and places the resulting PPDU onto the power line as described in the PHY section above.

As the receiver reconstructs the MSDUs, it selectively acknowledges the PBs; those that are not positively acknowledged are retransmitted during the next TXOP. The Selective Acknowledge (SACK) is an integral part of the TDMA allocation. When all the PBs composing an MSDU have been received correctly, the segments are decrypted and the resulting MSDU is passed to the Convergence Layer for delivery to the appropriate HLE.

Control messages are processed in an analogous fashion.

Since FEC and Selective Acknowledgment (SACK) are performed on relatively small blocks of data, the FEC is more robust and retransmissions are minimized. These two features contribute to HPAV’s ability to operate at near channel capacity.

Central Coordinator

Each Central Coordinator controls an AV Logical Network (AVLN) which consists of several AV stations which all share a common Network Membership Key (NMK). The NMK and other details of security and encryption are discussed below. For now, it is sufficient to know that the NMK provides exclusive access to an AVLN so that the member stations can communicate in a private and secure environment.

As described above, the Central Coordinator provides bandwidth management services for the AVLN. These include admission control (determining whether to admit a new connection when it is requested). If the connection is admitted, the Central Coordinator schedules time allocations for the connection in the PCF Region. It manages this schedule via the beacon, which contains:

• the current schedule and the minimum number of Beacon Periods for which it will remain valid, and/or
• the new schedule and the number of Beacon Periods which will pass before it becomes valid.

When an AV Station is powered on, it listens to the medium. If it hears an existing AVLN, it will attempt to join it. If it does not hear an existing AVLN, it will form its own network by becoming a Central Coordinator and broadcasting a beacon. Eventually, another AV Station will be powered up and the two will associate and form an AVLN. (This is a highly simplified description; in reality, the HPAV specification provides for various cases of encountering more than a single AVLN, encountering HomePlug 1.0 devices, encountering other Powerline Networks and combinations thereof.)

The Central Coordinator attempts to learn the topology of its AVLN and of any neighboring AVLNs. To achieve this, each AV station broadcasts a Discover Beacon periodically (at a time allocated by the Central Coordinator in the non-persistent portion of the PCF Region). This Discover Beacon contains information about the station and the AVLN to which it belongs.

Each station that hears the Discover Beacon adds the information it contains to a Discovered Station List (DSL). While building its DSL, if the station encounters a Discover Beacon from a station in a different AVLN, it adds the information about the other network to a Discovered Networks List (DNL). Periodically the Central Coordinator asks each station for its DSL and DNL and use the collected lists to compose a topology map.

The Central Coordinator uses the topology map it builds from the collected DSLs and DNLs to determine if there is another station in the AVLN that would make a better Central Coordinator than it. The criteria for making this decision, in order of priority, are:

1. User’s Selection
2. Central Coordinator Capability
3. Number of discovered STAs in the Discovered Station List
4. Number of discovered AVLNs in the Discovered Network List

If the current Central Coordinator finds another station that would make a better Central Coordinator, it will negotiate a handover of Central Coordinator functions to the new Central Coordinator. Depending upon the capabilities of the old and new Central Coordinators, the handover may or may not result in existing connections being torn down.

The Central Coordinator may also select another Central Coordinator-capable station to be its backup in case of failure. If the station accepts the backup role, it will monitor the AVLN and, if the Central Coordinator’s beacon is not heard by any stations in the AVLN for a specified number of Beacon Periods, the backup Central Coordinator will assume the role of Central Coordinator and attempt to maintain the existing connections without disruption.

Since a station must be capable of communicating with the Central Coordinator in order to join an AVLN and establish connections, a proxy capability is provided to support stations that are hidden from (i.e., unable to communicate with) the Central Coordinator. This capability provides for the creation of a Proxy Coordinator (PCo) to repeat the beacon information in Proxy Beacons and to relay control messages between the hidden station and the Central Coordinator. Note that only control messages are relayed. The station must be able to communicate directly with any stations with which it wishes to establish a connection. The PCo also transmits a Proxy Beacon during each Beacon Period to convey scheduling and other information to the hidden station.

When all stations are idle, the Central Coordinator causes the AVLN to enter a power saving mode. In this mode, there is only a small CSMA Region (so stations can initiate communication) and a small PCF Region (just long enough for Discover and Proxy beacons). Stations must have their receivers on during these small regions to participate in the AVLN; they may turn their transmitters and receivers off for the remainder of the Beacon Period. This makes it easier for stations to qualify for Energy Star certification.

Convergence Layer

The Convergence Layer (CL) serves as the interface between the HLEs and the MAC in the Data Plane. It accepts data payloads through Service Access Points (SAPs) at the H1 Interface and processes them as needed prior to handing them off to the MAC through the M1 interface. The only Data SAP specified by AV is the Ethernet II-class stack. This stack supports packet formats as specified by IEEE 802.3 with or without IEEE 802.2 (LLC), IEEE 802.1H (SNAP) extensions, and/or VLAN tagging. Using the Ethernet format makes it easy for AVLNs to interface to other LANs.

Among the services the CL provides on the transmit side are classification and auto connection. If requested for a connection, the CL will also associate an Arrival Time Stamp (ATS) with the data payload. On the receive side, the CL provides (optional) smoothing and insures that the received MSDUs are delivered to the appropriate H1 Service Access Point (SAP). On both sides, it provides the Connection Manager sufficient information to monitor the level of QoS being provided by the connection.

When a connection is established, the CM provides the classifier with a set of rules that will enable the classifier to uniquely associate incoming packets with the connection. For example, a set of rules might specify the source and destination MAC addresses and the TCP source and destination ports of the connection.

The classifier examines each packet received at the H1 interface and attempts to match it with a connection using the classification rules that have been provided to it. If it finds a match it will label the packet with the Connection ID (CID) of the appropriate connection, otherwise the classifier will release the packet for transmission in the CSMA region at the appropriate priority level.

If the transmitting station supports the optional Auto Connect Service (ACS), all packets which are released by the classifier without being associated with a connection will be examined by the ACS which will evaluate the data flow(s) between a given source and destination and attempt to identify flows which are worthy of a connection. This evaluation and identification may be based on a mix of the following:

• Policies established by an HLE (or a manufacturer),
• Templates such as traffic associated with ports known to have a particular usage,
• Heuristics such as the volume and regularity of data which is being transmitted.

Until the ACS identifies a connection, it releases the packets for transmission in the CSMA period immediately upon completion of the packet’s inspection.

If the ACS identifies a particular data flow as connection worthy, it behaves in a manner analogous to an HLE and asks the CM to set up a connection, providing classifier rules, etc. When the CM establishes the connection, the Classifier will start associating the packets with the newly established connection and the ACS will no longer see them. The ACS is, however, responsible for servicing the connection in the same way that an HLE would.

At the receiving station, the CL demuxes the received packets. If the packets are associated with a connection for which smoothing (a.k.a. de-jittering) has been requested, the CL will buffer the packets for the appropriate time so that all packets are released to the HLE at a fixed interval after they arrived at the H1 interface at the transmitter, which the receiver knows from the ATS it received with the packet and the synchronized network clock.

On both ends of a connection, the CLs provide sufficient information to the CM that it can monitor the level of QoS being provided to a connection to insure that the guarantees are being met. The CM will take corrective measures specified by the CSPEC if there are any violations to the QoS guarantees.