Wireless Networks
`
`- Catching the mobile future -
`
`Microsoft Corporation
`Exhibit 1013-00001
`
`Microsoft Corporation
`
`

`

`974
`
`WA 5.1
`
`WCN
`
`THE WALKSTATION PROJECT ON MOBILE COMPUTING
`
`Frank Reichert
`
`Royal Institute of Technology, Dep. of Teleinformatics
`Elecnum 204, S-164 40 Stockholm, Sweden
`tel. +46—8—752 1434/ 1440
`fax. +46—8—751—l751
`reichert@il.kth.se
`
`Abstract: Wireless LANs (WLAN) will soon be as
`popular as Ethernet is today. First products for DECT
`and 2.4 GHz systems will be available this year.
`How can these WLANs be connected to the Inter-
`net? Which problems arise when mobile users appear
`as guests in foreign networks?
`The Walkstation project is realizing a testbed for
`such a mobile computing environment using a modi-
`fied set of Internet protocols. These protocols are based
`on the work of the intemet working group on mobile
`IP (Internet protocol).
`Other aspects of the Walkstation project cover low
`power VLSI, cellular infrastructures, network integra-
`tion, and mobility aware applications.
`
`1. INTRODUCTION
`
`The popularity of the Internet and its
`attractive applications make it a natural choice
`for mobile communication and computing.
`
`Currently every 30 minutes [1] a new net—
`work is connected to the lntemet. The Internet
`combines about 32.000 networks in over 80
`countries in the middle of 1994. Some of these
`
`networks will be wireless networks for support
`of mobile clients.
`
`The World Wide Web (WWW) [3,4] is an
`example for a new world-wide multimedia infra—
`structure based on the Internet. Easy—to—usc
`informatiOn browsers made the WWW traffic
`
`grow within one and a half year from O to 600
`Gbit/month [2] which is about 5% of the total
`traffic on the N'FS backbone. This is one exam-
`
`ple of many attractive applications which are
`available for hosts using the Internet.
`
`By their nature networks for mobile clients
`require more effort
`for communication than
`wired LANs. Suitable communication architec-
`
`tures have to deal with problems such as lower
`link quality, changing location of mobile clients,
`and security problems.
`
`The reminder of this paper will describe
`
`the efforts of the Internet working group on
`Mobile-1P to integrate mobile communication
`into the Internet as well as work done in the
`
`Walkstation project [5,6] at the Royal Institute of
`Technology in Stockholm, Sweden.
`
`2. MOBILE IP
`
`The goal of the working group on Mobile-
`IP (established in June 1992) is to “allow trans—
`
`parent routing of IP datagrams to mobile nodes
`in the Internet” [7]. This requires the Internet
`protocol (IP) enhancements to be ( l) transparent
`for the mobile user, (2) for fixed users, and (3)
`for existing routers in the network.
`While most mobile networks
`
`handle
`
`mobility at the data link layer the proposal of the
`Mobile-1P group handles all at the network layer.
`This is important in order to get a standard suita—
`ble for existing and future networks independent
`of the underlaying network technology.
`
`The Mobile—1P proposal achieves these
`goals by providing mobile hosts with a penna—
`nent “home address”, regardless of its current
`point of attachment to the Internet, and a “care-
`of—address”, which provides information about
`the current point of attachment.
`
`Additional nodes at the home site, called
`“home agents”, forward packets to the mobile
`host by tunneling encapsulated packets through
`the Internet. At the destinatiOn they are decapsuv
`lated either by the mobile host directly or by an
`additional node called “foreign agent”.
`
`The simplicity of the protocol allows easy
`implementation and early field trials.
`
`Detection of the Network
`Before a mobile node with a wireless inter—
`
`© IEEE / ICCC
`
`Microsoft Corporation
`Exhibit 1013-00002
`
`Microsoft Corporation
`
`

`

`WCN
`
`WA 5.1
`
`975
`
`As a guest in a foreign network the mobile
`host registers with a “foreign agent” (fig. 2).
`This foreign agent consults the corresponding
`home agent Whether it
`is allowed to provide
`service. By receiving this registration the home
`agent can now update the location of the mobile
`host in its database. A confirmation message to
`foreign agent results in a successful registration
`of the mobile host.
`
`Packet Forwarding
`
`The cooperation of the home agent and the
`foreign agent after a successful
`registration
`allows the mobile node to exchange packets with
`nodes in the Internet.
`ha.home.se
`
`
`rne® mh.home.se
`gm
`tan.us
`
`
`
`MH.HOME.SE
`
`fa.foreign.us
`
`((C
`
`>
`
`3—
`
`ritaj©it.kth.se
`
`Figure 3: Packet Forwarding
`
`A fixed user sends a message to the perma-
`nent home address of the mobile host. The mes—
`
`sage appears on the home network where it is
`received by the home agent in place for the
`mobile host which is currently registered as a
`guest in a foreign network.
`
`The home agent encapsulates the packet
`into another IP packet which is sent to the for-
`eign agent. The foreign agent recognizes the
`packet for the mobile host inside, decapsulates it
`and sends it to the mobile host.
`
`The forwarding process was transparent
`for the fixed user, the mobile user and the net-
`
`work.
`them.
`
`.No special actions were required from
`
`Reconfirmation
`The maximum time of service for a mobile
`
`face can Start to communicate it must detect the
`
`network. Contact points to the network are
`offered by agents which act as base stations
`offering Internet connectivity.
`
`“NH BEACON
`
`FOREIGN
`
`A FA
`
`???
`
`MH.HOME.SE
`
`FOREIGN'i .US
`
`Figure 1: Foreign Agent Beaconing
`
`A Mobile—[P agent announces its presence
`by regularly sending beacon messages (fig. 1).
`These beacons indicate not only that Internet
`service is offered but which agent is offering it.
`
`Beacons can be requested by the mobile
`node through solicitation messages to which the
`agent answers by sending a beacon.
`
`these messages are
`In the current draft
`based on ICMP Router Discovery [8].
`
`Registration
`
`Being aware of a contact point to the Inter-
`net the mobile host can now request service by
`registering with that agent.
`
`If‘ the node is in the “home network” then
`
`it will have to register with the “home agent”,
`which maintains a list of mobile hosts for which
`
`it is responsible.
`
`HOME
`AGENT HA
`
`FOREIGN
`AGENT
`
`MH.HOME.SE
`HA-FA-Confirm
`
`Figure 2: Registration at a foreign agent
`
`host is supervised by timers at the foreign and at
`
`Microsoft Corporation
`Exhibit 1013-00003
`
`Microsoft Corporation
`
`

`

`976
`
`WA 5.1
`
`WCN
`
`the home agent (fig. 4). A mobile node has to rev
`register from time to time in order to avoid
`denial of service due to an expired timer.
`FA Service
`
`HA Service
`Life Time
`
`
`
`((C
`
`LifeTime
`
`
`
`MH.HOME.SE
`FOREIGN AGENT
`
`T0 FA
`
`l
`
`OR
`TO HA
`
`HUME
`AGENT
`
`with a new foreign agent. It sends a registration
`message to the new foreign agent which asks the
`home agent for service confirmation. The home
`agent generates two messages, one to deny serv-
`ice to the previous foreign agent and one to
`allow service to the new foreign agent
`
`The home agent is now aware of the new
`location of the mobile node, while the previous
`foreign agent will stop to forward packets to the
`mobile node.
`
`Packets still arriving at the previous for-
`eign agent are decapsulated and sent back to the
`home network.
`
`Deregisrmtion
`
`Figure 4: Communication monitoring
`
`These timers help to detect nodes which
`have moved out of the service area of an agent,
`or which were simply switched off.
`
`A mobile node can stop service by a
`deregistration procedure. This
`is useful
`for
`mobile nodes which are being switched off, and
`for hand-over procedures.
`
`Hand-over
`
`Open Questions
`
`A mobile host may wish to register at a
`new agent due to movement or other reasons. A
`move might occur so fast that the mobile host
`loses contact to the previous foreign agent. It can
`therefore not tell the previous foreign agent that
`it has disconnected and reconnected at a new
`
`agent.
`
`PREVIOUS
`
`NEW
`
`Foreign
`Agent
`
`F2323? ((g
`
` FA-MH Con irm
`
`Figure 5: Handover between foreign agents
`The hand—over in Mobile—1P handles this
`
`with an enhanced registration procedure (fig. 5).
`The mobile node detects that it should register
`
`Route optimization allows exchange of
`packets without the use of the home agent. Two
`nodes located very close to each other, e.g., in
`US, need the help of a home agent, e.g., in Swe—
`den to communicate with each other. Different
`
`proposals are discussed to optimize this ineffi—
`cient use of bandwidth. Solutions discuss mobil—
`
`ity aware hosts, which support Mobile-1P, and
`“caching agents” which participate in the proto—
`col in order to redirect packets.
`
`Securitv mechanisms are currently stud-
`ied to mutually authenticate mobile nodes, for-
`eign agents and home agents. Attacks by
`eavesdropping are much easier for wireless sys-
`tems than for wired links. Proposed is an MD5
`authentication scheme. The key distribution
`problem is however still an open issue.
`
`Scalability from wide area networks down
`to local area networks may not be satisfying. In a
`local environment with multiple cells traffic
`would be forwarded from a home agent in one
`cell to a foreign agent in another cell. If cells are
`as small as rooms a lot of forwarding can take
`place inside the same department of an organiza—
`tion loading the network at least twice for each
`message.
`
`Microsoft Corporation
`Exhibit 1013-00004
`
`Microsoft Corporation
`
`

`

`WCN
`
`WA 5.1
`
`977
`
`Signalling effort for the support of mobile
`nodes may take a considerable part of the net—
`work bandwidth if nodes are very mobile like for
`a car or for pedestrians using pico~cells (e.g.,
`radius < 50 meters). Measurements and simula-
`
`tions have to verify Whether such signalling can
`overload a network.
`
`3. THE WALKSTATION PROJECT
`
`involves
`project
`The Walkstation II
`researchers from different areas in order to find a
`solution for mobile multimedia communication
`
`via a global system approach. Beyond the inte-
`gration of mobile stations into existing fiXed net-
`works (Internet), areas being studied include
`mobile adaptive applications, mobile systems
`management and security, future communication
`architectures, use of multiple wireless interfaces,
`dynamic resource allocation and media access,
`direct sequence CDMA (code division multiple
`access) and low power VLSI integration.
`
`Low power VLSI
`
`An important task is the investigation of
`new low cost, digital, highly integrated radio
`interfaces which are small in size and have low
`
`power conSumption. A current prototype of a
`Walkstation radio interface utilizes a Direct
`
`rently we are studying the performance of indoor
`packet radio systems using diversity reception.
`
`Network Aspects
`
`In the Walkstation project the variety of
`networks (TDMA/CMDA—based, GSM, Mobi—
`tex, DECT, etc. [10]) and the difference of envi~
`ronments (indoor picocell, outdoor microcell,
`urban maeroeell and rural areas) are harmonized
`
`at the network layer (fig. 6). The goal is to pro~
`vide mobility transparent communication serv-
`ices to traditional, non—adaptive applications
`such as electronic mail, file—transfer, virtual ter—
`
`minal, information browsers and other applica-
`tions common to users on the Internet.
`
`DECT, 802.11T
`HiperLan
`Infrared
`<-——->
`
`Mobilex
`GSM, NMT
`DECT
`4-——>
`
`Mobile Station
`
`
`
`
`
`Wireless LAN
`ROUTER
`
`Wide Area
`Base Static
`
`INDOOR
`LAN
`
`OUTDOOR
`WAN
`
`Sequence CDMA receiver [9] and transmitter
`operating with 500 kbit/scc at 2 GHz.
`
`Global Coverage
`Local Coverage
`ngh Peak Power
`Low Power
`High Datarale (a: 1 Mbps) Low Datarate (a 10 - 100 kbp
`
`The simplicity of both the analog and dig—
`ital parts should allow both low power operation
`and small area for a final BiCMOS implementa-
`tion. The next prototype aims at 2 Mbit/sec at
`2.4 GHz. It allows the parallel reception of mul-
`tiple spreading codes as well as hardware recon-
`figuration for
`flexible bandwidth and new
`modulation schemes.
`
`Cellular Aspects
`
`In this task the focus is on design princi-
`ples for distributed radio access schemes and
`performance evaluation of these schemes in
`wireless information systems. The main empha-
`sis is on packet switched environments where
`both moderate speed interactive data communi—
`cation systems as well as high—speed ATM—based
`systems for wireless access are considered. Cur—
`
`Figure 6: Integration of mobile networks
`Not all differences of mobile networks can
`
`be hidden from user applications. Parameters
`such as available bandwidth, latency and setup
`time influence the responsiveness and perform—
`anee of multi-media applications. New adaptive
`applications have to react to the available quality
`of service by, e.g., changing the presentation of
`user objects from full graphical display to simple
`character based interaction.
`
`A device which can act either as a modem
`
`for mobile hosts or as a complete base station
`has been realized by the same kind of hardware
`but with different kinds of communication soft—
`
`ware. This device, called MINT (Mobile IP
`Router) [11,12], consists mainly of three parts,
`one for the connection to the host or backbone
`
`Microsoft Corporation
`Exhibit 1013-00005
`
`Microsoft Corporation
`
`

`

`' 978
`
`WA 5.1
`
`WCN
`
`net (Ethernet), one for connecting to wireless
`LANs (radio or infrared), and a processing part
`for
`the communication protocols. By using
`Ethernet as an input channel,
`it should not be
`necessary to install special purpose wiring dedi-
`cated to mobile communication.
`
`To realize a large scale field trial, such a
`new type of router (MINT) must be low cost,
`small in size, and have a low power consump—
`tion while offering high performance. The cur-
`rent version of the MINT is a 68030 based router
`
`with multiple interfaces, including as host inter-
`faces: SCSI, Ethernet, serial, parallel; and as
`wireless interfaces: IR, microwave-radio, or a
`
`second ethernet (supported for software devel—
`opment).
`
`Network Measurements
`
`Traffic characteristics are still not well
`
`understood for integrated mobile data communi-
`cation. Multimedia applications, such as tel-
`eteaching or information browsers will generate
`traffic of a nature different from that of regular
`data and telecommunications. By capturing data
`essentially generated by multimedia applications
`driven by diverse and distributed groups of
`users, we will be able to study characteristics of
`traffic generated in a mobile data network.
`
`The Walkstation project provides an excel-
`lent opportunity to investigate the validity of a
`number of currently used traffic models. Recent
`packet traffic measurements for a wide spectrum
`of packet based networks have convincingly
`demonstrated the fractal nature of packet traffic
`and also the significance of these properties in
`the analysis, design and modelling of packet
`based networks [13,14].
`
`4. SUMMARY
`
`The simplicity of the MobilewIP protocol
`will enable a rapid integration of mobile hosts
`into the Internet. Some open questions like scal-
`ability and security are still under study.
`
`The Walkstation project is one initiative to
`study future mobile communications in a system
`approach covering aspects
`from low power
`VLSI up to mobile applications.
`
`[1]
`
`[2]
`
`[3]
`
`[4]
`
`[5]
`
`[6]
`
`[7]
`
`[8]
`
`[9]
`
`5. REFERENCES
`
`Internet Society News, No.1, Vol.3, 1994, pp. 8—9
`
`“Special Report - World Wide Web and Mosaic,
`“Internet Society News, No.1, Vol.3, 1994, pp.15-
`19
`
`B. Kelly, “Becoming an Information Provider On
`the World Wide Web, “Internet Society News, No.1,
`Vol.3, 1994
`
`T. Bemers-Lee, “World Wide Web Initiative“,
`HTML-document at URL; http://info.cern.ch/
`hypertext/WWWfIheProj ect.html
`
`Y. Ismailov, T. Ohsawa, P. Pruthi, F. Reichert, "The
`Walkstation Project - A flexible platform for experi—
`ments in mobile multimedia environments," IFIP
`Congress 94, Hamburg, Germany, Aug. 1994
`
`G. Maguire, F. Reichert, M. Smith, "A Multiport
`Mobile Internet Router," 44th IEEE Vehicular Tech-
`nology Conf., Stockholm, Sweden, June 1994
`
`WA: Simpson, “IP Mobility Support, “IETF Mo-
`bile IP Working Group, 5th draft, 1994, available
`via anonymous FI‘P from playground.sun.com as
`pub/mobile-ip/draft-ietf-mobileip-protocol—05.txt,
`or inW with URL=file://playground.sun.com/
`pub/mobile—ip/draft—ietf—mobileipprotocol-05.txt
`
`S. Deering, “ICMP Router Discovery Messages",
`Internet RFC 1256, September 1991.
`
`Daniel Kerek, Hannu Tenhunen, Gerald Maguire
`Jr.,Frank Reichert, “Direct Sequence CDMA Tech-
`nology and its Application to Future Portable Multi—
`media Communication Systems”, IEEE ISSSTA’94:
`Third International Symposium on Spread Spec-
`trum Techniques & Applications. Oulu, Finland,
`July 1994
`
`[10]
`
`[11]
`
`[12]
`
`[13]
`
`[14]
`
`B. Goldenberg, W. Sonnenfeldt, ”The Sky’s the
`Limit for Wireless WAN5,” Data Communications
`Magazine, Vol. 22, No. 3, Feb. 1993
`
`R. Hager, A. Klemets, G.Q. Maguire Jr., M.T.Smith,
`and F. Reiehert, “MINT - A Mobile Internet Rout-
`er”, 43rd IEEE Vehicular Technology Conference,
`New Jersey, USA, May 1993.
`
`Anders Klemets, Gerald Q. Maguire, Frank Re—
`ichert, Mark T. Smith, “MINT-A Mobile Internet
`Router", First IEEE International Symposium on
`Global Data Networking, Cairo, Egypt, Dec. 1993.
`
`A. Erramilli and W. Willinger, “Fractal Properties in
`Packet Traffic Measurements,” Proc. of ITC Semi-
`nar, St. Petersburg, 1993
`
`A. Erramilli, R.P.Singh, P.Pruthi, “Chaotic Maps As
`Models of Packet Traffic,” to be published [TC ‘94.
`
`Microsoft Corporation
`Exhibit 1013-00006
`
`Microsoft Corporation
`
`