Pervasive Services Infrastructure Dejan S Milojicic HP Labs
Pervasive Services Infrastructure (Ψ) Dejan S. Milojicic HP Labs, Palo Alto 3/15 2001 http: //www. hpl. hp. com/research/itc/csl/pss/psi 3/18/2001 Y, Pervasive Service Infrastructure 1
Ψ’s Strategic Direction Internet Wireless Service Providers q Global e. Commerce to reach $6. 8 trillion by 2004 (Forrester Research) q Internet Service Gateway market to reach 25 M US homes by 2005, worth $5 B (Parks Assoc. ) q Commerce over mobile phones in W. Europe rise to $37. 7 B in 2004 ($51. 2 M in 1999, IDC) q Worldwide shipments of handheld computers will surpass 5. 7 M, 47% increase (Dataquest) 3/18/2001 Y, Pervasive Service Infrastructure 2
What is Disruptive in this Space q General v v pervasive (ubiquitous, invisible) deployment of computers, smart spaces scale: large number of devices (localized) & services q User perspective v v user interfaces, user intent context awareness diversity of services and clients connectivity q Developer perspective v v versioning and maintaining products (scale thereof) variation in network speed (wireless → wired; wired → memory) the price of each component (reduced cost) service interoperability (everything available on the Internet) 3/18/2001 Y, Pervasive Service Infrastructure 3
Tough Problems Today q q q Wireless speed (need higher bandwidth, lower latency) Unreliable connections (missing disconnected support) Lack of apps and app development tools Lack of solid infrastructure (database, synchronization) Non-scalable solutions Non-trusted solutions (lack of end-to-end security) Lack of quality displays & U/I Few users (innovators and early adopters) Protocols evolving (WAP, XML implementations, APIs) User experience (clumsy handheld, poor I/F) Viability (fractured embedded OS space) 3/18/2001 Y, Pervasive Service Infrastructure 4
Relaxing Assumptions for 2004 (not all problems are Y ’s) q q q Ubiquitous connectivity (3 G+) End-to-end security (Wireless VPN) Device interoperability (Lucent & Novell, Motorola & Lucent, etc) Better battery lifetimes (piezoelectric, solar, bioelectric, etc. ) Handheld, phones, others, … converged Billing, customer care → commodity services Voice U/I (many already work on) Higher quality color displays New Web apps built from ground up Middleware solutions adapt access to existing apps Wireless Web crosses the chasm Consolidation of embedded OS space 3/18/2001 Y, Pervasive Service Infrastructure 5
What are Remaining Problems q q q q Scalability (localized) Service access and adaptivity in dynamic environments Administration (automatic, transparent, dependency aware) One-size fits all (embedded system software solutions) Interoperability (APIs, service brokers, etc. ) Performance e. g. load, latency, caching Maintaining true end-to-end security guarantees 3/18/2001 Y, Pervasive Service Infrastructure 6
Scenario Scene: Helsinki airport, year 2004. Jane and Dave are on a business trip to present a new product to overseas customers. Dave worked the whole weekend at home on an updated presentation. They are stepping off of the plane. Jane: Relax Dave, everything is going to be fine. Your presentation is in good shape. Just fix the typo in the CTO’s name and tune the figures, Dave: Fix the title! Tune the figures! Are you out of your mind? ! I only have the presentation on my phone. I’d need to download and fix it on my laptop. We don’t have time for that before the meeting. Jane: Yes, you can. I can do it on my new phon. DA. Dave: You never told me you got a new, more powerful phon. DA Jane: No, it’s the same as yours, NK 234 y. P, but I’ve updated the software. Dave: No way, my presentation is 4 G. It can’t fit on any phon. DA Jane: Look, I download your presentation. . . Here’s the slide you want, … ok… the CTO’s title is fixed, … now let’s see the figures… it’ll take longer… ok, here are figures, move the 2003 forecast up a little, . . perfect. 3/18/2001 Y, Pervasive Service Infrastructure 7
Scenario, cont. Dave: How did you do it? !! Look, I’m downloading it on my phon. DA and it takes forever… even worse, now it won’t start. Jane: You see those small bumps on the walls, ceiling, and floor. These are embedded servers. Software on my phon. DA offloads networking, memory, processing, storage - you name it - to these servers. Dave: Sigh… Jane: It is called Psi; it’s very simple. Now with respect to these bumps on walls, … hold on a second… my daughter is paging me, first things first…. Yes honey, … I forgot to sign your homework … ooops … Dave: Can your magic Psi help with this too? Jane: Sure it can, see, here it will download her application from her school – it adapts to my phon. DA since the app is designed for desktops– it’s really downloaded on a server there. Looks correct, ok signed! 3/18/2001 Y, Pervasive Service Infrastructure 8
Scenario, cont. Dave: Wow, you are a wizard, installing all these applications, adapting them to your small phon. DA. How long did it take you? Jane: Not a second, Psi does it all, it adapts, it downloads services on demand. Ψ Sigh 3/18/2001 Y, Pervasive Service Infrastructure 9
Pervasive Servers & Clients. . servers… . . in te se rm rv ed er iat s… e IDCs . . clients… . . embedded devices and sensors… 3/18/2001 Y, Pervasive Service Infrastructure infrastructure Ψ embedded 10
Conceptual Layers Services and Applications Middleware Layers E-Speak, Jini, CORBA, ACL Ψ platform: match service requirements to client resources Local OSes & JVMs Client Handheld Device (and Infrastructure Servers) 3/18/2001 Y, Pervasive Service Infrastructure 11
Ψ Vision Adapt any service to any client (anywhere, anytime) q General assumptions v v clients will always be diverse clients will always be less powerful than desktops there will be a large scale of devices in the future services will be accessible on demand from anywhere q Underlying Y assumptions: v v v service adaptivity required three-tier model service splitting 3/18/2001 Y, Pervasive Service Infrastructure 12
Ψ, Important Research Questions q How to enable users to exploit the new pervasive computer infrastructure How to seamlessly offer more services to more clients anytime anywhere? v How to avoid installing and administering increasing number of computers in a pervasive infrastructure? v What are new abstractions and algorithms for computation, communication, storage, user interface → how to write and run new apps? v It is not the strongest of the species that survive, nor the most intelligent, but the ones most responsive to change - Charles Darwin 3/18/2001 Y, Pervasive Service Infrastructure 13
Adaptive Offloaded Services q Goal: increase scalability & performance of (mobile) service delivery to resource-poor devices and enable where not currently possible q Key research questions v v v Splitting service between client and mid-point (intermediate) servers Dynamic adjustment of services: device size, load, roaming Masking performance implications of splitting q Required infrastructure topics v v Distributed run-time support - mid-point(s) and client Sharing, administration, and security at mid-points 3/18/2001 Y, Pervasive Service Infrastructure 14
Adaptive Offloaded Services Illustration 0101 1110 Internet on IDC e Embedded server dependent execution 0101 111010 0 L Ph 010101 111010 010101 1110 DS Which? Back-end service execution Home, office, shop, etc. 3/18/2001 Y, Pervasive Service Infrastructure 15
Adaptive Offloaded Services, Cont. q Splitting services v v v Monitoring (resources, execution, objects) Offloading (migration mechanisms, trigger & placement policies) Service splitting (interaction metrics, graph partitioning, 3 -v. 2 -tier) q Adaptive Offloading v v v Adapting to devices (very constrained, different resources) Adapting to load (multiple services, scalability) Adapting to mobility (roaming, migration of offloaded state) q Performance v v Policies (account for overhead, interaction metric, stability) Mechanisms (offloading overhead, disruption of offloading - lazy) 3/18/2001 Y, Pervasive Service Infrastructure 16
The Java Heap Page Working Set of Embedded Caffeine Mark 3/18/2001 Y, Pervasive Service Infrastructure 17
Services on Demand q Goal: zero-administration and -installation of mobile clients q Key research questions v v v dynamic service composition and deployment on zero-installed devices de/re-coupling of user’s service context between devices tolerance to disconnection of services q Required infrastructure topics v v service brokers and infrastructure device encapsulation APIs 3/18/2001 Y, Pervasive Service Infrastructure 18
Services on Demand Illustration Day 1, John in Sydney Secured Storage Provider Internet Day 2, John in Montreal Service broker Back-end service execution ASP Home, office, shop, etc. 3/18/2001 Y, Pervasive Service Infrastructure 19
Loading Overheads So. D Loading (s) Loaded URLs Java Class Local Load (s) Intranet Wave Dialup Cnt Size (B) calculator 0. 1 ~0 +0. 1 +4. 7 2 9520 calendar 0. 2 ~0 +0. 4 +4. 7 4 12952 editor 0. 2 ~0 +0. 3 +5. 2 7 15885 game 0. 2 ~0 +0. 3 +6. 1 9 18739 agenda 0. 2 +0. 1 +0. 7 +8. 4 1 35360 12 59846 ftp 0. 6 +2. 1 +3. 2 +20. 9 3 107725 22 142155 mail 2. 5 +2. 7 +15. 1 +129. 4 1 675046 138 365574 Service 3/18/2001 Y, Pervasive Service Infrastructure 20
Services on Demand, Cont. q Service Composition/Deployment on Zero-Installed Devices v v v service characterization (naming, keyword, extension, requirements) service discovery (brokers, resource matching, scalability) deployment (transparent downloading & caching, sandboxing for service intercommunication q De/re-Coupling of Service Context Between Devices v v v remote storage (clients are volatile, access any time anywhere, user/service transparency, stackable client RFS, synchronization/encryption) context information (per service state, preferences, expressed as files) Adaptation to different client (resource re-evaluation, select service) q Tolerance to Disconnection of Services v v client side cache (static content: URLs, classes; user files/data; user/service network interactions reconnection (pluggable reconciliation per service/file, client-initiated 3/18/2001 Y, Pervasive Service Infrastructure 21
Architectural Principles q q Minimal changes (extensions) to the underlying client sys SW Minimal changes to services (mainly splitting) Use existing mid-point servers in the infrastructure Psi will rely on standard communication protocols (IP, WAP, etc) however, it will have the ability to support extension to new ones q Heterogeneity of underlying OSes and implementation of JVMs q The primary metric will be enabling new services and better use of resource-constrained devices 3/18/2001 Y, Pervasive Service Infrastructure 22
Implementation Details q Prototype environment v v Jornada Pocket PC Java VM, Chai VM wireless 802. 11 Linux PCs for infrastructure servers q Services v v v tools: editors (text/image) visualization PIM: calendars, mail, calculators, stock analysis interactive games etc. 3/18/2001 Y, Pervasive Service Infrastructure 23
Technical Approach q Develop a Y architecture & prototype implementation q Experimental, quick prototyping, 3 -month increments q Spiral approach v v Incrementally demonstrate new functionality Uncover new key questions q Engage partners at incremental 3 -month phases v HPL, HP, university, industry (external) q Patent/publish as we progress 3/18/2001 Y, Pervasive Service Infrastructure 24
Showcases q Adaptive offloaded services initial mobile app splitting, evaluate split advantage (scale/perf/power) v v v app which is too resource intensive for existing devices app which is interactive → mobile editing and interactive gaming app which is data intensive → mobile multimedia and visualization q Service-on-Demand service broker lookup, automatic download, disconnected mode (cache/remote) v app that can work disconnected (runs part of the service locally) v app that accesses user data from a server v app in two different flavors, to adapt to Palm constrained resources 3/18/2001 Y, Pervasive Service Infrastructure 25
Current Insights q q q Opportunity for memory offloading Complex interactions in Java objects Transparent remote storage: class interposition, bytecode editing Disconnection: caching of some class client/service interactions Reconciliation through plugins 3/18/2001 Y, Pervasive Service Infrastructure 26
Team Members q Resource-Constrained Devices, team lead OS, JVM, consumer products (Sony) q Team member HA, distributed systems (SRI, Informix, Oracle) q Services-on-Demand, team lead OS, JVM (Bull, OSF RI) q Dejan Milojicic, (myself) Psi PM OS, distributed systems & agents (Belgrade RI, OSF RI) 3/18/2001 Y, Pervasive Service Infrastructure 27
Competition q Telcos v AT&T, Motorola, Ericsson q Consumer electronics companies v Sony, Philips, etc. (HAVI) q Traditional computer systems companies v Sun (Jini, embedded/real-time Java); Microsoft (. NET, Windows CE) IBM (pervasive computing, embedded tools) q Data base companies v Oracle, Sybase q Many startups v Stream. Theory, Omni. Shift, Transvirtual, M 2 Verticom, Wireless. Knowledge Many universities, government q Numerous competition, but huge business & innovation opportunity 3/18/2001 Y, Pervasive Service Infrastructure 28
Related Work q Adaptability & Offloading v v v UW (Portolano) Active Fabrics + service infrastructure Berkeley’s (Endeavor) Ninja + general service offloading Palm, Win. CE + offloading to shared mid-points OSGi + general purpose + distributed; UW Kimera + dynamic service split CMU Odyssey, UIUC Active Spaces + service splitting MIT Oxygen SW environment + resource awareness q Data storage & resource management v v v CMU Coda & Odyssey + reconciliation framework UCLA File mobility + trust for services; Web. FS + more than web browsing OSGI + mobility; Java OS JDI & JPI + API signatures q Java platform v v Java. OS + service infrastructure; Pocket. Linux + disconnected + other OSes JNT, but not just a terminal; Web. OS + service brokers & trust 3/18/2001 Y, Pervasive Service Infrastructure 29
What Ψ is Not About q q q q q Client HW HW (re)configuration Service design & implementation New protocols New programming languages General purpose wireless Device location technologies General purpose OS development Fault-tolerance Hard real-time 3/18/2001 Y, Pervasive Service Infrastructure 30
Summary q Intersection of services, the Internet, and wireless: v v v adapt services to existing client resources automate service deployment reason about new, disruptive technologies in pervasive computing q Potential long-term area of research, ties in v v v servers (IDCs) & client pervasive infrastructures academia, government, big & small companies different markets q A lot of opportunities for contribution; risk-aware, leverage potential q It is fun to do these things! Immediate benefit obvious 3/18/2001 Y, Pervasive Service Infrastructure 31
Pervasive Computing Landscape application model (information access & presentation) user/agent model (intelligence) nomadic task mobility global data placement location transparency information architecture context management Y user studies internet com (IP) device connectivity XML UDDI VML … mobile com. (non IP) service connectivity 3/18/2001 Y, Pervasive Service Infrastructure security mgmt pricing/ charging development adaptivity Y 32
- Slides: 32