HIPERLAN

HIPERLAN
Introduction
• HIPERLAN stands for High Performance Local Area Network.
• Standardized by ETSI in 1996 as a Wireless LAN (WLAN) supporting:
o Node mobility
o Both ad-hoc and infrastructure-based topologies.
• Designed to provide high-speed wireless communication with time-sensitive (QoS) data
transfer capabilities.
Standardization Background
HIPERLAN 1 was the first of four HIPERLAN types originally planned:
HIPERLAN 1 – Wireless LAN for mobile nodes.
HIPERLAN 2 – Broadband wireless LAN (evolved into ETSI BRAN).
HIPERACCESS – Wireless access for broadband connectivity.
HIPERLINK – Point-to-point high-speed wireless interconnect.
HIPERLAN
Evolution & Impact
• Although HIPERLAN 1 and Wireless ATM were not commercially
successful, their QoS mechanisms and MAC design principles greatly
influenced later standards such as HiperLAN2 and IEEE 802.11e.
• The focus shifted from HIPERLAN 1 to HiperLAN2, integrating
developments from:
• ETSI BRAN (Broadband Radio Access Networks)
• Wireless ATM research
Key Takeaway:
• HIPERLAN 1 laid the foundation for QoS-enabled wireless broadband
systems, serving as a technological bridge between early WLANs and
later high-performance standards like HiperLAN2.
Overview of HIPERLAN 1 (ETSI 1998b)
Definition
• HIPERLAN 1 (High Performance Local Area Network) — a wireless LAN
standardized by ETSI (1998b).
• Targeted at 23.5 Mbit/s data rate over 5.1–5.3 GHz, ~50 m indoor range at 1
W TX power.
Design Goals
• Provide QoS support through priorities + packet lifetime.
• Enable mobility, ad-hoc + infrastructure operation.
• Integrate mechanisms for:
• Forwarding and topology discovery
• Encryption & confidentiality
• Network identification
• Power conservation
Network Compatibility &
Identification
MAC Service Model
• Compatible with IEEE 802.x MAC services → easy LAN integration.
• Uses 48-bit MAC addresses (same as Ethernet).
Identification & Coexistence
• Unique HIPERLAN ID scheme allows multiple overlapping HIPERLANs to
coexist without interference.
Security
• Encryption/decryption with identical keys + initialization vectors
required for successful reception.
• Ensures confidentiality of user data (but limited key-management
defined).
Forwarding and Power Conservation
MAC-Layer Relaying
• Supports multi-hop forwarding between relays → extends coverage
beyond single radio range.
• Operates entirely at MAC layer, transparent to higher layers (no IP
routing needed).
Energy Management
• p-savers: nodes that sleep most of the time, wake per pattern.
• p-supporters: neighbors that store packets and forward only when p-
saver is awake.
• Requires buffering + scheduling at supporters.
• Achieves significant power saving at the cost of latency + state
maintenance.
QoS Support and Prioritization
QoS Mechanisms
• Priorities: each data packet assigned a priority level; higher priority →
faster channel access.
• Packet Lifetime: limits delivery time; packets exceeding lifetime are
dropped to preserve QoS.
• These mechanisms provide delay guarantees and traffic
differentiation.
Observation
• Although HIPERLAN 1 pioneered QoS-aware WLAN design, later IEEE
802.11 implementations omitted priorities (PCF rarely used) yet
succeeded commercially due to simplicity.
EY-NPMA: Channel Access
Mechanism
Core Access Protocol:
Elimination-Yield Non-Preemptive Priority Multiple Access (EY-NPMA)
• Three Phases of Access:
1. Prioritization:
Competing nodes determine the highest-priority packet ready for transmission.
2. Contention:
If several nodes share the same top priority, contention begins.
→ Elimination: remove most contenders.
→ Yield: one node remains.
3. Transmission:
The winning node sends its packet (non-preemptive).
Essence:
• Provides collision-minimized, QoS-aware medium access.
• Balances efficiency (few collisions) and fairness through structured contention.
EY-NPMA Access Modes in HIPERLAN
1
Three possible access conditions:
1. Synchronized Channel Condition
• Many nodes contend simultaneously.
→ All three phases executed: Prioritization → Contention (Elimination + Yield) →
Transmission.
• Ensures fair and deterministic resolution under heavy load.
2. Channel-Free Condition
• Channel idle ≥ 2 000 high-rate bit-periods + a random dynamic extension.
• Dynamic extension: random delay = 0 – 3 × 200 bit-periods (equal likelihood).
• Only Transmission phase used → minimizes latency when the medium is truly idle.
3. Hidden Elimination Condition
• Used when hidden terminals exist (nodes cannot sense each other).
• Additional elimination sub-cycles reduce the probability of simultaneous transmissions.
Purpose of Dynamic Extension
Why add the random extension?
• Prevents synchronized channel grabs by stations aligned at higher protocol layers.
• Each node independently waits a random multiple of 200 bit-periods before
transmitting.
• Effect:
• Randomization ⇒ lower collision probability.
• Higher fairness across synchronized senders.
Example
If the medium was idle ≥ 2 000 bit-periods:
• Station A adds 0 × 200 bp = immediate transmit.
• Station B adds 2 × 200 bp = wait 400 bp.
• Station C adds 3 × 200 bp = wait 600 bp.
→ only one starts first → others defer
 EY-NPMA Contention Phase
Structure
Contention Phase = Elimination + Yield
Sub-Phase Goal Result
Rapidly reduce number of Leaves a small, constant subset of
Elimination Phase contenders using short bursts & nodes, nearly independent of initial
random backoff. contenders.
Random waiting among remaining Exactly one winner → ready for
Yield Phase
nodes. Transmission.

Operation Sequence:
Synchronization → Prioritization → Elimination → Yield → Transmission.
Key Parameters defining timing:
IPS: Prioritization slots
IPA: Acknowledgment interval
IES: Elimination slots
IESV: Variance in elimination duration
IYS: Yield slots
Each measured in bit-periods, determining the duration of its sub-phase.
 Phases of the HIPERLAN 1 EY-NPMA access scheme

Figure gives an overview of the three main phases and some more details which will be
explained in the following sections. For every node ready to send data, the access cycle starts
with synchronization to the current sender.
The first phase, prioritization, follows. After that, the elimination and yield part of the
contention phase follow. Finally, the remaining node can transmit its data. Every phase has a
certain duration which is measured in numbers of slots and is determined by the variables
IPS, IPA, IES, IESV, and IYS.
Prioritization Phase in HIPERLAN 1
(EY-NPMA Stage 1)
Purpose
• Ensures that higher-priority packets always gain access before lower ones.
• Prevents low-priority nodes from occupying the channel while high-priority traffic is
waiting.
• Works even under heavy load — priority enforcement is absolute, not probabilistic.
Priority Classes
• HIPERLAN 1 defines five priority levels:
• Slot 0 → Highest priority
• Slot 4 → Lowest priority
Two Sub-steps:
1. Priority Detection — nodes listen to detect higher-priority traffic.
2. Priority Assertion — if medium idle for the node’s priority duration, node asserts its
priority by sending a burst.
Timing & Operation of Priority Detection
Slot Timing
• Each slot = IPS = 168 high-rate bit-periods.
• Bit-period ≈ smallest timing unit of PHY (≈40 ns at 25 Mbit/s).
Operation Flow:
1. Priority Detection:
• Node with priority p listens for p slots (0 to p–1).
• If medium idle for all these slots → node continues.
• If activity detected → node stops and waits for next access cycle
[Link] Assertion:
• After idle detection, node transmits a burst for IPA = 168 bit-periods.
• Burst signals its claim to the medium at priority p.
Burst Bit Sequence (repeated pattern):
11111010100010011100000110010110
End Condition:
• Prioritization phase ends when at least one node asserts its priority with a burst.
• Phase duration depends on the highest priority present, not fixed length.
Example Scenario: Three Nodes
Competing
Node Packet Priority (p) Action During Slots Outcome
Listens to slots 0–1 →
Node 1 2 Survives
idle; sends burst in slot 2
Listens to slots 0–1 →
Node 2 2 Survives
idle; sends burst in slot 2
Listens to slots 0–3 →
Node 3 4 Stops attempt
detects burst in slot 2

Result:
•Nodes 1 & 2 survive (both have top priority 2).
•Node 3 (priority 4) defers.
•Prioritization phase took 3 slots (0, 1, 2).
Key Insight:
Only nodes of the highest active priority continue to the next
EY-NPMA stage (Contention → Elimination + Yield).
Characteristics & Design
Significance
Properties
• Dynamic phase length: depends on the highest priority in the network.
• Fairness preserved: among equal-priority nodes, contention continues later.
• Deterministic behaviour: ensures latency bounds for top-priority traffic.
• Advantages
i. Guarantees access for urgent packets.
ii. Low control overhead (simple burst signalling).
iii. Provides strict QoS differentiation at MAC layer.
• Limitations
i. Burst-based signalling sensitive to synchronization and noise.
ii. Only five fixed priorities — limited granularity for complex QoS profiles.
Summary:
• The Prioritization Phase acts as a priority gatekeeper, filtering contenders and enforcing
deterministic, class-based channel access.
Elimination phase Elimination Phase (EY-NPMA Stage 2)

Purpose
• The Elimination Phase resolves contention among nodes that survived the
Prioritization Phase.
• Objective: reduce the number of contending nodes to a small, near-constant set —
independent of initial load.
• Uses random-length burst transmissions to statistically “thin out” contenders.
Timing Basics
• Time divided into Elimination Slots, each of IES = 212 high-rate bit periods.
• Nodes send elimination bursts with random durations (0–12 slot intervals).
• After bursting, nodes perform Elimination Survival Verification (ESV) for IESV = 256 bit
periods.
Outcome:
• Only nodes that sense the channel idle during their verification survive and proceed to
the Yield Phase.
Elimination Burst Probabilities
• Each contending node independently selects a burst length n (in slot
intervals) according to a geometric-like probability distribution:

Interpretation:
• Shorter bursts are more likely, longer bursts are rare.
• Longer bursts statistically dominate the channel → suppress shorter
ones.
• After each node’s burst, it listens for IESV:
• If channel idle → survives.
• If activity detected → eliminated.
Elimination Phase Timing and
Operation
• Step-by-step Process:
i. Each contender chooses random burst length n via PE(n).
ii. Transmits burst (same bit sequence as in priority assertion).
ii. Listens during IESV period.
[Link] channel remains idle → node survives; otherwise → eliminated.
Phase Duration = (Longest elimination burst among nodes) + IESV
Result:
• A few nodes survive → proceed to the Yield Phase.
• Ensures network scalability: duration and survivor count remain
predictable even under high contention.
Yield Phase (EY-NPMA Stage 3)
• Purpose
• Final step in EY-NPMA after Elimination Phase.
• Used by surviving nodes to passively listen before transmission — no further
bursts sent.
• Objective: minimize remaining contenders and allow a single transmitter to
proceed.
• Timing Basics
• Medium divided into Yield Slots of duration IYS = 168 high-rate bit periods.
• Each node randomly selects a Yield Listening Period (n) = 0 to 9 slots (equal
probability).
• Probability distribution:
Operation of the Yield Listening
Process
Step-by-Step Sequence
1. Each remaining node chooses a random n ∈ [0, 9] slots.
2. Listens silently for its n slots.
3. If medium stays idle for entire listening period → node survives.
4. If any transmission sensed → node withdraws from this cycle.
Phase Duration
• Determined by the shortest yield-listening period among contenders.
• The node(s) with smallest n finish first → likely to start transmitting first.
Result:
• At least one node survives and begins data transmission.
• Other nodes detect the transmission and defer.
Transmission Phase (EY-NPMA Stage
4)
Definition
• After successfully passing Prioritization, Elimination, and Yield phases, a node gains
access to the medium and begins data transfer.
• This is the final step of HIPERLAN 1’s EY-NPMA cycle.
• Data Unit Sent
• Node transmits a Low Bit-Rate / High Bit-Rate HIPERLAN 1 CAC Protocol Data Unit
(LBR–HBR HCPDU).
• The PDU encapsulates both control and data fields, optimized for reliable, QoS-
aware transmission.
Transmission Type
• Can be either:
• Unicast → to a specific destination node.
• Multicast → to multiple nodes within HIPERLAN cell.
Unicast Transmission &
Acknowledgement Mechanism
For Unicast Mode:
• After sending an LBR–HBR HCPDU, the transmitter expects an immediate
acknowledgement (ACK) from the receiver.
• The ACK is sent as an Acknowledgement HCPDU (AK-HCPDU).
Acknowledgement PDU (AK-HCPDU):
• A Low Bit-Rate HCPDU containing only the LBR part (control section).
• Confirms successful reception of the transmitted frame.
• If ACK not received within a timeout window → node assumes loss and re-enters
contention in the next cycle.
Timing Characteristic:
• Immediate ACK reduces latency and supports fast re-synchronization among nodes.