Global Mobility Management in Multi-access Networks
1 like198 views
This document provides an overview of global mobility in multi-access networks. It discusses the need for global mobility management to allow users to access services anywhere, anytime independently of location, terminal, or access technology. It presents several use cases to illustrate global mobility concepts and challenges. It also reviews key notions in mobility management including nomadism, session continuity, and handover latency. Additionally, it examines major approaches to mobility management such as 3GPP, WiMAX, Wi-Fi, and Mobile IP and their benefits and limitations for supporting global mobility.
Global Mobility Management in Multi-access Networks
- 1. Campus da FEUP Rua Dr. Roberto Frias, 378 4200 - 465 Porto Portugal T +351 222 094 000 F +351 222 094 050 www@inescporto.pt www.inescporto.pt © 2007 Global Mobility in Multi-access Networks Rute Sofia (rsofia@inescporto.pt) http://telecom.inescporto.pt/~ian 2008 Fev 27
- 2. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 2 Overview • What is global mobility • The global perspectice • Some use cases • Notions • Is it really needed? Yes vs. No factions • Main requisites • Main mobility management approaches of today • Summary/Conclusions • Some interesting research topics
- 3. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 3 What is Global Mobility? • Today, end-users have a highly nomadic lifestyle (e.g. frequent travelling; different residential locations; different access locations) – Least service disruption: access services, anywhere, anytime. • End-users hold in average a cellular and a fixed access subscription – One end-user may hold N subscriptions coming from X providers – Multi-access technologies and architectures are involved (wireless/fixed/cellular) • Wireless introduced the possibility to extend Internet access into new areas – Wireless extends the fixed networks – Wireless works as new backhaul – Wireless allows bypassing local-loop legislation – Potential for new services and new markets • Multi-access architectures (wireless/fixed/cellular) – One or more operators involved; different access technologies Global Mobility: Internet access anywhere, anytime, independently of location, terminal and access technology
- 4. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 4 User on the Move (Train, car…) User Plane Session stays; re-registration not required – SESSION CONTINUITY Provider A Domain 1 Domain 2 Domain 3 Domain 4 Domain 5 Domain 6 Fixed Wireless Mobile Micro-Mobility Micro-Mobility Micro-Mobility Provider B Provider C Partial Macro-Mobility The Global Perspective Today User on the Move (Train, car…) Session breaks; re-registration required – nomadism but NO session continuity User Perspective Internet Services (Google, Skype, …) IMS Services (Instant Messaging, VoIP) Service Provider Perspective
- 5. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 5 User on the Move (Train, car…) User Plane Session stays; re-registration not required – SESSION CONTINUITY Provider A Domain 3 Domain 4 Domain 5 Domain 6Fixed Wireless Mobile Provider B Provider CAccess/Network Provider Perspective A GLOBAL SOLUTION…Mobility Management Functionality Global Macro-Mobility Any Access Any Service Seamless to the User The Global Perspective Tomorrow User on the Move (Train, car…) User Perspective Internet Services (Google, Skype, …) IMS Services (Instant Messaging, VoIP) Service Provider Perspective No visible interruption: session continuity Domain 1 Domain 1 Domain 1 Domain 2 Domain 2 Domain 2
- 6. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 6 Use Case 1 Hotspots /Internet-on-the-go MN on the move Provider A Administrative Domain 1 Provider B Administrative Domain 2 Same IP subnet (L2 mobility scheme) Administrative Domain 3 Administrative Domain 4 Wireless Mobile Hotspot 1 Correspondent Node 1 Different IP subnets Hotspot 2
- 7. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 7 Mobile Access Fixed Access Use Case 2 Multihoming • End-user holds a mobile terminal with UMTS and WLAN interfaces • On the street relies on UMTS • At home prefers WLAN/DSL (higher throughput) -broadband access uses Wi-Fi as first- hop • Two interfaces = two different IP addresses – Automatically activate an interface – Automatically divert sessions from one interface to another? • And if receiver has two different interfaces, which one to rely upon? GGSN UMTS Access WLAN Access Access Router Internet
- 8. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 8 Use Case 3 FON, viral networks Internet Access provider • Fonera community – Linus: buys social router (34,4 € plus 5 € dispatch) for own use; shares broadband access for free – Bill: buys router and re-sells access; gets 50% of sales from vouchers – Alien: gets access by means of vouchers • Viral network – Spread without operator control – Backhaul is not changed! • Limited mobility – Just nomadism • Simple AAA – WEP, WPA – No incentive for malicious usage “Bill” user1 “Alien” user3 Pays 3€ per 24h usage “Linus” user2@somewhere “Linus” user1@home
- 9. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 9 Mobility Management Notions (1/2) Nomadism, aka Nomadicity • The same service look & feel delivered independently of user location or terminal • Change of attachment point requires service re-activation •Sessions (e.g., TCP) have to be re-initiated • Requires global form of addressing (identification) and global AAA • Example: 3G global roaming Session Continuity • Sessions are seamessly diverted (handed over) from an old identifier to a new • Change of attachment point does not disrupt active sessions • Examples: L2 mobility support (micro-mobility); Mobile IP (v4 and v6) Optimized Global Mobility Management requires support for both nomadism and session continuity
- 10. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 10 Mobility Management Notions (2/2) • Global handover = Layer 2 handover + Layer 3 handover • Layer 2 handover building blocks – E.g.: in 802.11b, 500~2000ms, mainly scanning of available APs • Layer 3 handover building blocks • Movement detection (1-3s; Neighbor Discovery default values) – Current router reachability – Validity of current address – Availability of new router on the network – IP address configuration (stateless, stateful) • Duplicate address detection (DAD) time (~1s) • Layer 3 signaling procedures (example based on MIPv6) • Binding update (in microseconds) • Route optimization (in microseconds) 0 0.5 3 4 4.05 Layer 2 handover Movement detection DAD Layer 3 signaling Global Handover latency components (example based on Wi-Fi+MIPv6) Time (sec)
- 11. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 11 Do we Really Need Global Mobility Management? Yes vs No Factions The “No” Faction • Only some applications require session continuity •No need for it in the near future • Mobility management is not a service per se • Mobility management must be centralized to work • Mobility management is a layer 2 issue and has been dealt with • Main stakeholders: 3GPP The “Yes” Faction • End-user nomadic lifestyle is increasing •Global mobility management is necessary to convey adequately all subscribed services • Global mobility provides basis for new services •New business models and opportunities will come along • A single operator should not take care of all mobility management • Stakeholders: virtual and micro operators; wireless-based operators, end-users
- 12. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 12 Main Requisites • Support for multi-access networks – Fixed, mobile, wireless – Multi-mode terminals – One or more operators involved – Simultaneous handover support (source and receiver) • Services provided anywhere, anytime – Zero disruption! – Independently of access technology and terminal – Agnostic to ANY application (current OR FUTURE) • Highly nomadic lifestyle requires… – Adequate global AAA (or incentive to act honestly) – Adequate global resource management – Seamless handovers (session continuity) – Ease of use (user-friendly approaches) Can we do it with based upon current approaches? One solution serves all, or shall approaches be combined?
- 13. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 13 Major Mobility Approaches of Today Transport (OSI L4) Network (OSI L3) Data/Link (OSI L2,L1) Application (OSI L5-L7) *SCTP: Stream Control Transport Protocol SIP/IMS M-SCTP* HIP Mobile IP Layer 2 Mobility (micro-mobility) GPRS/UMTS Wi-Fi WiMAX “Mobile” Ethernet
- 14. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 14 3GPP Mobility Management • Roaming agreements between different providers • Connectivity: GTP tunneling – only one IP address is used anywhere, anytime! • Normally, Home GGSN is the mobility anchorpoint • In some cases, the Visited GGSN may be the chosen anchorpoint! • E.g., due to resources or proximity of services • No session continuity possible with GTP! • Current solution discussed – Mobile IP •GGSN:Gateway GPRS Support Node •SGSN: Serving GPRS Support Node •UE: User Equipment •RNC: Radio Network Controller •HLR: Home Location Register •PDP: Packet Data Protocol •PLMN: Public Land Mobile Network •NSP: Network Service Provider •BG: Border Gateway •GTP: GPRS tunneling protocol •ER: Edge Router • Potential Benefits – Long experience in mobility management – ETSI TISPAN expands IMS to fixed access • Key Issues – Requires specific client – Only supports nomadism – Only supports VoIP/IM – Not suited for Internet- based services
- 15. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 15 WiMAX Mobility management Main Features • 802.16-2004 defines the fixed WiMAX – Only nomadism • 802.16e-2005 defines the mobile WiMAX – Mobility: nomadism and session continuity – Intra-ASN handovers supported by L2 – Inter-ASN supported by Mobile IP • ASN-GW as mobility anchor point • Potential Benefits – Open system – Contemplates global mobility – Considers interoperability with other technologies • Key Issues – Coverage, collisions – Specific security system has implications in interoperability – Layer 2 features may require adaptation from upper layer mechanisms •UE: User Equipment •NAP: Network Access Provider •NSP: Network Service Provider •ASP: Application Service Provider •MSS: Mobile Subscriber Station •CoA: Care-of-Address •BS: Base Station
- 16. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 16 Wi-Fi Mobility management Internet Mobile Provider WIMAX Provider AP1 AP2 AP3 AP14 DSL Provider Micro-handover Macro-handover Macro-handover Main Features • Handover triggered by signal threshold • Scanning causes major latency of handover (300~400 ms) – Some optimized algorithm can reduce scanning latency to ~50ms (still too large for real-time applications) • Potential Benefits – Easy to implement – High deployment penetration – Easily extends other access technologies • Key Issues – MAC Layer vendor-specific interpretated – Low coverage – collisions – No global mobility method – Weak security
- 17. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 17 Network layer Mobility Management (1/5) Mobile IP • HA: Home Agent – Mobility anchor point – Control plane: signalling, registration, data redirection – Data plane: routing • MN: Mobile Node – MIPv6/v4 client – If MIPv6, integrated in the IPv6 stack – not an additional package! • CN: Correspondent Node – Third-party with whom MN communicates MN Seamless handover Provider A Administrative Domain 1 Provider B Administrative Domain 2 Administrative Domain 3 Administrative Domain 4 Different IP subnets MN Wireless Mobile Internet Services CN 1 CN 2 Different IP subnets Same IP subnet (L2 mobility scheme)
- 18. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 18 Visited Network Foreign Agent (FA) Mobile Node (MN) Assigns (permanent) HoA to MN If MN has registered its new CoA, HA intercepts Packets to MN HoA and tunnels them to FA Home Address (HoA) MN Internet Registers CoA with HA via FA Home Network Home Agent (HA) Assign CoA to MN and forwards traffic arriving on from the tunnel to the MN 1. Mobile Node moves 2. FA advertises IP address 3. MN registers CoA with HA via FA 4. HA intercepts all packets to HoA of MN and tunnels them to CoA (FA) 5. FA decapsulates packets and forwards them to the MN 6.In the other direction, packets do not need to be tunneled to the HA -> triangular routing Network Layer Mobility Management (2/5) Mobile IPv4 1 2 3 4 5 CoA: Care-of-address
- 19. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 19 VN MN (HoA=IP1) MN (CoA=IP2) Internet CN HN Home Agent (HA) Network Layer Mobility Management (3/5) MIPv6 (Mobile IPv6) Access Router2 3 3 1 1. MN moves 2. gets CoA from VN access router (automatically) 3. MN registers CoA with HA AND with active CNs 4. HA intercepts packets sent to HoA and redirects them to CoA (v6-in-v6* tunneling) 5. active sessions get actively diverted GRE tunneling: Global Routing Encapsulation HN: Home Network VN: Visited Network HoA: Home Address
- 20. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 20 ADVANTAGES TO MIPv4 • Integrated into IPv6 •Mandatory in every implementation • Larger address space • No foreign agent • Address autoconfiguration • Integrated Neighbor/Router discovery/unreachability • Improved Security •MIPv6 return routability procedure •IPsec used to protect MIPv6 exchanges • Automatic home agent discovery • Support for route optimisation • More efficient routing •CN->MN packets sent directly •MN->CN packets sent directly MIPv6 OPTIMISATION APPROACHES • Fast handover support •Reduction of movement detection time •Signaling load minimisation/network registration time reduction •Fast handovers for MIPv6, FMIPv6 (RFC 4068) •Hierarchical MIPv6, HMIPv6 (RFC 4140) •F-HMIPv6 (draft in progress) • Smooth handover support •Multihoming issues (IETF WG Monami6) •Simultaneous bindings (draft in progress) Default MIPv6 is slow for real-time and interactive applications handovers, even ignoring the data link layer handover delay Network Layer Mobility Management (4/5) MIPv6 Advantages/Open Issues
- 21. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 21 HMIPv6 • Reduces the MIPv6 signaling load – separates local from global mobility • Two different CoAs: Local CoA (LCoA), Regional (RCoA) •MN does not need to frequently change CoA (but still exchanges regular binding updates) • MAP works as local HA: intercepts packets to the MN and tunnels them to the LCoA VN 1 MN Internet HN HA CN VN 2 LCoA 1 LCoA 2 MN H@ => RCoA MAP FMIPv6 • Reduces the latency/packet loss between handovers • Anticipates the L3 handover (L2 trigger) • Access router forwards data packets to new location during the L3 handover • Assumes that old AR knows neighbor AR informationMN@PCoA RCoA LCoA1 LCoA2 VN 1 Internet HNHA CN VN 2 1. L3 Information about new AR obtained as soon as MN “discovers” new possible L2 attachment 3. CN traffic redirected while MIPv6 binding exchange for NCoA is being performed MN@NCoA 2. Tunnel established by previous access router, between PCoA and NCoA Network Layer Mobility Management (5/5) SOME MIPv6 Extensions
- 22. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 22 Between network and transport Layers (1/3) HIP: Host Identity Protocol (IETF RFC 4423) –Originally intended as security protocol •Supports multihoming •Embodies some of the properties required for mobility, e.g., identifier independent of link/network layer addresses •Communication secured by means of IPSec –Bump in the Stack •New Layer (host layer), between network/transport layers •New Host Identifier (HI) •For applications, HI is used; for routing, IP –End-host identified by a public key •128-bit hashed key (called HIT tag) identifies host on a session •32-bit local scope identifier used for compatibility to IPv4 •Binding to IP address is dynamic and can have a 1-N relation –Host Layer •New namespace – extensions for DNS required •New signaling protocol, 7 message types
- 23. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 23 Between network and transport Layers (2/3) HIP: The Overall Picture Above the HI Layer, communication uses HI for identity Below the HI layer, HI is replaced by IP for routing SPD: Security Policy Database SAD: Security Association Database
- 24. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 24 Between network and transport Layers(3/3) HIP: Summary – Potential Benefits • Security/anonimity • Multihoming support • Solution that can support global mobility (agnostic handovers to ANY application) – Key Issues • Requires OSI stack update – new layer between network/transport • Requires DNS extensions (two new resource records) • DNS – HIP mapping is problematic – Mobility: frequent handovers imply frequent mappings of HI to locator – Charging/local identifier association is tricky – Mapping global tag to other identifiers problematic within a service/access provider
- 25. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 25 Session Layer Mobility SIP Session Identification Protocol (1/2) Pre-Call Stage – SIP Proxy Server • Relays call signalling • Keeps no session state – SIP Redirect Server • Redirects callers to other servers – SIP Registrar • Accept registration requests • Keeps state about user (Location Server) – SIP Address • Globally reachable • Address data format, e.g., user@somewhere.org – Must include hostname; may include username, port number, others – Non-SIP URLs can be used (e.g., mailto)
- 26. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 26 Session Layer Mobility SIP Summary (2/2) • Potential Benefits – SIP is generalized as signaling solution – Can deal with NATs (STUN) – Allows different levels of mobility (personal, terminal, service) – Can be used together with other solutions – Signalling path != data path – SIP components do not interfere in the data path • Key Issues – Does not support TCP-based applications – Does not provide global session continuity • But redirection with minor delay is possible (for SIP-based applications) Mid-Call Stage
- 27. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 27 “Mobile” Ethernet (1/4) • Mitsubishi/NIICT paper/solution (2001) • Relies on a virtual MAC address scheme – Mostly thought for multihoming cases • Architecture “glues” different segments forming a single IP subnet – Signaling required to update MAC tables split (inter/intra segment) • Layer 3 broadcasts (e.g., ICMPv6) treated as unicast • Neighbor discovery servers hold mapping between IP/ virtual MAC • Specific mobility management frames used
- 28. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 28 “Mobile” Ethernet (2/4) - Registration • MN (MD) enters AP range • performs registration request • AP forwards request to Edge Switch • Edge Switch requests a Virtual MAC address to a Neighbor Discovery Server (NDS) • Edge Switch creates entry for obtained virtual MAC • Switches in the Mobile Ethernet only learn the virtual MAC
- 29. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 29 Mobile Ethernet (3/4) - Handover • MN (MD) moves between different APs • Signaling server detects handover and sends HO recommendation to Edge Switch • Old Edge Switch sends state about virtual MAC to new Edge Switch • MN requests registration to new switch • New edge switch updates virtual MAC entry
- 30. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 30 Mobile Ethernet (4/4) • Potential Benefits • Glues segments together • No need to change IP address – agnostic to application • Good support for multi-homed devices • May help to lower latency (anticipate the global handover) • Key Issues • Virtual MAC addresses are still flat – weak scalability (in terms of state/signaling overhead) • Does not solve problems related to IP changes • Network segments that are far apart cannot be glued • Network segments belonging to different providers cannot be glued • Mapping between IP/MAC required; • Several servers involved: NDS, SS • Frequent handovers between two MACs (ping-pong effect) may not be adequately supported
- 31. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 31 Conclusions • Current operational solutions fall short in providing global mobility • Micro-mobillity is limited in scope and range • Globally, only support for nomadism is considered • Exception: mobile WiMAX considers MIP for global management • There is NO “one serves all” solution • SIP does not support TCP-based applications • HIP requires a “bump-in-the-stack” – implies changing every single IP device and holds no clear benefit to other approaches • MIP seems to be the most flexible solution (agnostic to applications, technology, user) BUT requires IPv6 support in devices (MIPv4 is not an option) • Main approaches have common points • Unique identifier or way to dynamically map identifiers (locator/IP address) • Mobility anchorpoint • Most of the current solutions can be used together to create a global solution • E.g., SIP with MIP; SIO with HIP; HIP with MIP Before choosing a solution, requirements must be analysed! A global mobility management solution MUST be agnostic to application (current or future), network technology, and end-user!
- 32. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 32 Some Interesting Research Topics – Which OSI Layer(s) are best suited to support mobility convergence? – Multi-access handover issues • Resource management • Dealing with simultaneous movement • AAA, single sign-on – Handover optimization • Influencing handover decision – New metrics – Consider user policies – Context-awareness revisited – Mobility patterns & models • Mobility patterns influence architecture requirements • Adequate support for de-centralized control requires a better analysis of potential mobility patterns
- 33. © 2008 This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported License. http://creativecommons.org/licenses/by-nc-nd/3.0/ INESC-Porto, 2007.02.27 33 Thanks !