ICEBERG: From POTS to PANS Anthony D. Joseph Randy H. Katz Reiner E. Ludwig B. R. Badrinath UC Berkeley Stanford March 11, Cellular “Core” Network Bridge to the Future S. S. 7

ICEBERG: Internet-based core for CEllular networks BEyond the thiRd Generation June June 2001, joint with Ericsson High BW IP backbones plus diverse access networks –Different coverage, bandwidth, latency, and cost characteristics –Real-time services across diverse access networks –3G cellular: UMTS/IMT2000 –Next generation wireless LANs: Bluetooth –Home networking: DSL / Cable modem

Transparent Information Access Policy-based Location-based Activity-based Empower users! Speech-to-Text Speech-to-Voice Attached- Call-to-Pager/ Notification -to-Speech All compositions of the above! Universal Inbox

Smart Spaces Walk into a A/V room and control everything with your own wireless PDA –Services for each device –Automated discovery and use –Automated UI generation –Composite behaviors Phones as well as PDAs –Speech-enabled control

Potentially Any Network Service (PANS) 2-way Paging WIP GSM/CDMA PSTN IP Iceberg Access Points (More than gateways) Impedance matching Provide policy engine Handles routing, security IAP Same service in different networks Service handoff between networks E.g., “follow me” service E.g., any-to-any service High BW IP core Diverse access links

Important Trends Multimedia / Voice over IP networks –Lower cost, more flexible packet-switching core network –Simultaneous delay sensitive and delay insensitive flows (RSVP, Class-based Queuing, Link Scheduling) Intelligence shifts to the network edges –User-implemented functionality Programmable intelligence inside the network –Proxy servers intermixed with switching infrastructure –TACC model & Java code: “write once, run anywhere” –Rapid new service development, Speech-based services –New challenges for network security and management Cellular networks for the 21st century –High BW data (384 Kb/s-2 Mb/s): Reliable Link Protocols

ICEBERG Project Goals Demonstrate ease of new service deployment –Packet voice for computer-telephony integration –Speech- and location-enabled applications –Complete interoperation of speech, text, fax/image across the four P’s: PDAs, pads, pagers, phones) –Encapsulating legacy servers and supporting new, “thin” clients Demonstrate new system architecture to support innovative applications –Personal Information Management »Universal Messaging: , news, fax, voice mail »Notification redirection: e.g., , pager –Home networking and control of “smart” spaces »Build on experience with A/V equipped rooms in Soda Hall, transfer to home environment

ICEBERG Project Goals Understand the implications for cellular network design based on IP technology –Cellular / IP interworking functionality –Scalability: 100,000s of simultaneous users in the SF Bay Area –“Soft” QoS for wide-area, delay-sensitive flows Understand how to securely –Encapsulate existing applications services like speech-to-text –Deploy and manage computational resources in the network –Integrate other kinds of services, like mobility and redirection, inside the network

Outline Example Services Trends and Goals Experimental Testbed Project Approach Research Areas –Cellular / IP integration –Wireless link management –Multi-modal services Summary

Experimental Testbed SimMillennium Network Infrastructure GSM BTS Millennium Cluster WLAN / Bluetooth Pager IBM WorkPad CF788 MC-16 Motorola Pagewriter Soda 326 Soda “Colab” 405 Soda Velo Smart Spaces Personal Information Management H.323 GW Nino

Project Approach Make it real: build a large-scale testbed –Time travel: bring the future to the present –Collect “real” information about systems –Users develop new/interesting applications Understanding three key research areas –Cellular / IP integration: Mobility Management, Universal Inbox –Wireless link management »Packet Scheduling in GPRS/W-CDMA, Reliable Link Protocols –Multi-modal services: Speech control / Information dissemination ProActive Infrastructure: NINJA –Computing resources spread among switching infrastructure –Computationally intensive services: e.g., voice-to-text –Service/server discovery, security, authentication, and billing

Internet-Scale Systems Research Group Computing and Communications Platform: Millennium/NOW Distributed Computing Services: NINJA Active Services Architecture MASH Media Processing Services Distributed Videoconferencing Room-scale Collaboration TranSend Extensible Proxy Services ICEBERG Computer-Telephony Services Speech and Location Aware Applications Personal Information Management and “Smart Spaces” 5 faculty, ~35 students

Outline Example Services Trends and Goals Experimental Testbed Project Approach Research Areas –New service requirements: Multi-modal user interfaces –Generalized Information Redirection –Cellular / IP integration –Wireless link management Summary

New Service Requirements Encapsulation of complex data transformations –Speech-to-text, text-to-speech Dynamic service composition –Voice mail-to- , -to-voice mail Location-aware information services –E.g., traffic reports Multicast-enabled information services –Multilayered multicast: increasing level of detail as number of subscribed layers increases –Reliable information delivery over low bandwidth links

Multi-Modal User Interfaces Speech is the ubiquitous access method –Access from millions of phones (analog to digital cellular) Rapid support for new devices (new device in 2 hrs!) Simja Server Service Entity Room Control Entity Barbara Entity Emre Room (MASH) UDP RMI Gateway Cell Phone IP-Pad (BTS) RTP RMI

Interactive Voice Response to A/V Devices Application Dynamic data transcoding –Source and target data format independence / isolation Room Entity Microphone Cell phone A/V Devices Response to Client Automatic Path Creation Audio ICSI Speech Recognizer Text NLP Cmd Control/Metadata

Generalized Redirection Agents Users (will) have many communication devices Dynamic policy-based redirection –User- or service-specified policies –Universal Inbox: service, to pagers, etc. –Use APC to perform dynamic data transcoding Service mobility as a first class object

OfficePSTN: FaxPSTN: DeskIP: rover.cs.berkeley.edu:555 LaptopIP: fido.cs.berkeley.edu:555 PCS: Home: OfficePSTN: FaxPSTN: DeskIP: rover.cs.berkeley.edu:555 LaptopIP: fido.cs.berkeley.edu:555 PCS: Home: An Entity has a universal name and a profile; Entities are people, services or processes Universal Names: Globally unique IDs Profile: set of domain-specific names Service Mobility as a First-Class Object

Iceberg Inter-Domain Naming Protocol IDNP Server IDNP Server Caller’s network, Interactive, CallerID certificate) IDNP Server IDNP Server Profile Policy System State System State Replicated Information: Real-time Lazy Epidemic minutes/hours days/weeks weeks/months IAP

Voice Mail store Laptop (VAT) Univ-Inbox Service store Universal Inbox Service IDNP Server 1 IAP 1 IAP 2 IAP 3 IAP 4 GSM IAP 5 IP Core Network PSTN IDNP Server n

Cell-Phone to Cell-Phone 1. Dial Number Univ-Inbox Service 2. Intercept Call IDNP Server 3. Access Directory Service 5. Complete Call-Setup Data Path (Null) 4. Create Data-Path

Cell-Phone to POTS-Phone 1. Dial Number Univ-Inbox Service 2. Intercept Call IDNP Server 3. Access Directory Service Data Path 4. Create Data-Path GSM PCM 5. Complete Call-Setup

Cell-Phone to 1. Dial Number Univ-Inbox Service 2. Intercept Call IDNP Server 3. Access Directory Service 5. Complete Call-Setup Voic Service Data Path 4. Create Data-Path GSM PCM 6. Another Path PCM Text Send

Cellular / IP Integration Integrating a GSM BTS with an IP core network –Mapping IP signaling to SS7 radio management –Call admission and handoff Mobility management interworking –Mobile IP uses home agent / foreign agent –GSM uses Home Location Register / Visiting Location Register –Handoff between Mobile IP and GSM networks –Scalability, security of Mobile IP?

GSM BTS-IP Integration RBS 2202 UPSim Ethernet IP-PAD Traffic Signaling E1 Control Signaling GSM Phone E1: 13kb/s 12kb/s VAT Internet PC Interactive Voice Response Infocaster H.323 GW NetMeeting Uses OM & TRAFFIC to simulate BSC, MSC, and HLR functionality PSTN 2 TRX GPC board Thor-2 Performs rate adaptation function of ZAK/TRAU

Wireless Link Management Modeling GSM data links –Validated ns modeling suite, now using BONES simulator –GSM channel error models from Ericsson QoS and link scheduling for next generation links –High Speed Circuit Switched Data (HSCSD), General Packet Radio System (GPRS), and Wideband CDMA (W-CDMA) –RSVP signaling integration with bottleneck link scheduling Reliable Link Protocols –Wireless links have high error rates (> 1%) –Reliable transport protocols (TCP) interpret errors as congestion –Solution is ARQ protocol, but retransmissions introduce jitter

RLP-TCP Collection & Analysis Tools RLP and TCP interaction measurement / analysis –Both are reliable protocols (link and transport layers) –Trace analysis tool to determine current interaction effects –Tools for design of next generation networks (e.g., frame length) BTS TCP: End-to-End Reliability RLP: Wireless Reliability GSM-IP Gateway GSM Network TCP stats RLP stats TCP / RLP stats Post-processing tool (300 bytes/s)

TCP and RLP Data Plot Sent 30,720 bytes from mobile host to stationary host Dynamic interface - Zoom, scale - Add/delete items

Summary Iceberg testbed will be mostly completed by summer –Testbed will enable development of new protocols Lots of on-going design work –Automatic path creation –Service handoff: Passing metadata across/through networks –IVR: More applications and devices (WindowsCE) –Service location and discovery »Query model and security

ICEBERG: From POTS to PANS Anthony D. Joseph Randy H. Katz Reiner E. Ludwig B. R. Badrinath UC Berkeley Stanford March 11, Cellular “Core” Network Bridge to the Future S. S. 7