The Fundamental Need for Micro-Mobility
Standard Mobile IP (IPv4 or IPv6) operates as a macro-mobility protocol. It was architected under the assumption that a mobile node’s physical movements across distinct subnets would be relatively infrequent—for example, a user disconnecting their laptop from a corporate LAN in New York and reconnecting it hours later to a hotel Wi-Fi network in Tokyo. In this macro scenario, the latency incurred by updating the global Home Agent (HA) across the internet is acceptable.
However, modern cellular networks and dense urban Wi-Fi meshes completely violate this assumption. In a dense city center, microcells might only cover a radius of 200 meters. A user riding a bus could cross a physical subnet boundary every thirty seconds. If standard Mobile IP were deployed in this environment, the consequences would be catastrophic:
- Global Signaling Storms: Every single crossing of a microcell boundary would force the Mobile Node (MN) to acquire a new Care-of Address (CoA) and transmit a Binding Update across the global internet to its Home Agent, potentially located continents away. Millions of commuters doing this simultaneously would effectively launch a Distributed Denial of Service (DDoS) attack against the core internet routing infrastructure.
- Unacceptable Latency: The Round-Trip Time (RTT) required to send a Binding Update to a distant HA and receive an acknowledgment might exceed hundreds of milliseconds. By the time the HA updates its tunneling pointer, the user on the bus may have already exited the new microcell and entered a third one.
- Severe Packet Loss: During this prolonged update latency, all packets destined for the MN are still being tunneled by the HA to the old, stale CoA. These packets arrive at the old Base Station, find the MN missing, and are permanently dropped, rendering real-time applications like VoIP or video streaming impossible.
Micro-Mobility Protocols were engineered to solve this exact problem by introducing a hierarchical boundary. They create a localized, autonomous “mobility domain.” These protocols hide the chaotic, high-frequency local movements from the global HA. To the global internet, the MN appears perfectly stationary at a regional gateway. All local handoffs are handled instantly by altering internal routing tables within the domain, ensuring ultra-low latency handovers and completely eliminating global signaling overhead.
1. Cellular IP (CIP)
Cellular IP operates on a radical premise: within the localized micro-mobility domain, it discards standard IP routing protocols (like OSPF or BGP) entirely. Instead, it relies on a MAC-layer tracking system that mimics the operation of a traditional cellular network.
Architecture and Routing Topology
A Cellular IP domain consists of a single Gateway Router (which interfaces with the global internet and acts as the Foreign Agent) connected to numerous Base Stations via a tree-like topology of specialized Cellular IP nodes.
- To the outside world, the MN retains a single, static Care-of Address assigned by the Gateway.
- Inside the domain, packets are not routed based on IP subnets. Instead, Cellular IP nodes use the MN’s physical MAC address or a unique internal identifier to track its location.
Paging and Routing Caches
Cellular IP manages mobility using two highly specialized, time-sensitive databases maintained in the RAM of every node:
- Routing Cache (Active Mode): Used exclusively for active data transmission. When an MN transmits an uplink packet toward the Gateway, every intermediate node inspects the source MAC address and creates a temporary reverse-path mapping in its Routing Cache. Downlink packets arriving from the internet simply follow this precise, hop-by-hop breadcrumb trail backward to the MN. Because the MN is moving, these cache entries are highly volatile and expire in milliseconds if not constantly refreshed by uplink traffic.
- Paging Cache (Idle Mode): Maintaining active Routing Caches consumes significant battery power. When an MN goes idle, it stops sending data. To remain reachable without draining the battery, the idle MN sends periodic, lightweight “Paging Update” packets. These updates build a more generalized, longer-lasting path back to the MN. If a downlink packet arrives at the Gateway for an idle MN, the network uses the Paging Cache to broadcast a localized “page” to wake the MN up, transitioning it back to active mode.
Handoff Execution
Handoffs in Cellular IP are driven entirely by the mobile node measuring beacon signals, independent of the core network.
- Hard Handoff: The MN physically breaks its radio link with the old BS, connects to a new BS, and instantly transmits a “Route Update” packet uplink. The intermediate Cellular IP nodes intercept this packet and dynamically redirect the downlink breadcrumb trail to point to the new BS branch.
- Semi-Soft Handoff: Designed to eradicate packet loss for sensitive real-time traffic. As the MN approaches the cell boundary, it sends a route update to the new BS before breaking the physical link with the old BS. The Cellular IP domain dynamically bifurcates the downlink traffic at the nearest crossover node. Identical packets are streamed to both the old and new BS simultaneously. During the critical milliseconds while the radio physically retunes, data is safely buffered at the new BS, guaranteeing zero packet loss during the transition.
2. HAWAII (Handoff-Aware Wireless Access Internet Infrastructure)
HAWAII takes a fundamentally different architectural approach. While Cellular IP discards IP routing in favor of proprietary MAC-layer breadcrumbs, HAWAII utilizes standard, unmodified IP routing protocols but manipulates the routing tables dynamically to achieve micro-mobility.
Architecture and the Domain Root Router
- The HAWAII domain is managed by a Domain Root Router (DRR), which serves as the supreme ingress point from the global internet.
- When the MN enters the HAWAII domain, it acquires a Co-located Care-of Address (CoA) from the DRR. Crucially, the MN retains this exact same IP address as it roams across multiple different internal subnets within the domain.
Dynamic Routing Table Injection
In a standard IP network, if an MN moves to a new subnet but keeps its old IP address, packets will be dropped because the network prefix is topologically invalid. HAWAII bypasses this limitation through dynamic host-specific route injection.
- When the MN moves to a new Base Station, it sends a path setup message uplink.
- Instead of traveling all the way to the global HA, this setup message is intercepted by a Crossover Router located deep within the domain’s hierarchy.
- The Crossover Router dynamically injects a highly specific, host-specific route (a
/32subnet mask route) into its standard IP routing table, pointing specifically to the MN’s IP address. - Because standard IP routing always prefers the most specific match (Longest Prefix Match), any downlink packets arriving at the Crossover Router destined for the MN will hit this
/32route and be instantly deflected down the new physical path toward the new BS, ignoring the standard subnet topology rules.
Handoff Mechanisms: Forwarding vs. Non-Forwarding
HAWAII defines multiple path setup schemes depending on the network’s priority regarding speed versus packet ordering:
- Forwarding Scheme: Packets that arrive at the old BS just as the MN leaves are temporarily forwarded over the wired backbone network to the new BS. This prevents packet loss but can cause severe out-of-order delivery, which degrades TCP performance.
- Non-Forwarding Scheme: The path is updated rapidly at the crossover router, and any “in-flight” packets that unfortunately arrive at the old BS are simply dropped. This is faster and maintains packet ordering, but relies entirely on higher-layer protocols (like TCP Fast Retransmit) to recover the dropped packets.
3. Hierarchical Mobile IPv6 (HMIPv6)
Hierarchical Mobile IPv6 (HMIPv6) is the official Internet Engineering Task Force (IETF) standard for micro-mobility, specifically engineered to extend and optimize standard Mobile IPv6. It leverages the massive address space of IPv6 to create a clean, elegant proxy architecture.
Architecture: The Mobility Anchor Point (MAP)
The cornerstone of HMIPv6 is the introduction of a new architectural node called the Mobility Anchor Point (MAP). The MAP sits at the geographical boundary of the regional network (e.g., controlling a city or a large corporate campus) and acts as a localized proxy for the global Home Agent.
Dual Care-of Addresses
Under HMIPv6, the MN configures two distinct, simultaneous IPv6 Care-of Addresses using Stateless Address Autoconfiguration (SLAAC):
- Regional Care-of Address (RCoA): Formed using the network prefix advertised by the MAP. This address represents the MN’s overarching location at the regional level.
- On-Link Care-of Address (LCoA): Formed using the specific prefix advertised by the local Access Router (the immediate cell tower). This address represents the precise, physical location of the MN.
The Micro-Mobility Process
- Global Registration: When the MN enters the region, it registers its RCoA globally with its Home Agent (HA) and all active Correspondent Nodes (CNs). To the outside world, the MN appears to be permanently parked at the MAP. The MAP intercepts all global traffic destined for the RCoA.
- Local Registration: The MN registers its LCoA exclusively with the local MAP. The MAP creates an internal binding table linking the RCoA to the LCoA.
- The Local Handoff: When the MN physically moves to a new Access Router within the same city, its LCoA changes. The MN sends a rapid, localized Binding Update instructing the MAP to tunnel the RCoA traffic to the new LCoA.
- Signaling Isolation: Crucially, because the global RCoA has not changed, the global Home Agent and all Correspondent Nodes are completely oblivious to the movement. Absolutely zero signaling traffic escapes the regional domain. The MAP seamlessly intercepts the packets arriving for the RCoA and rapidly tunnels them to the new LCoA, providing highly scalable, secure, low-latency micro-mobility.