 |
|||
 |
Speakers Welcome
and Opening Remarks (Prof.
Madjid Merabti, IWNA5 Chair)
Keynote Speech (Prof. David Marples, OSGi) |
 |
|
 |
|||
 |
|||
 |
|||
  |
Download a PDF version of this presentation. What is Home Networking… Networks installed in the domestic environment for the purposes of shared wide area access and local area command, control and content distribution I will use Home Networking for the purpose of my examples today, but the principles apply equally well to most next generation ‘appliance’ devices… |
||
 |
|||
 |
|||
 |
|||
|
Who is it for? Challenges for home networking…
• Less Knowledgeable users • Low cost points • Data type versatility • Ease of installation • Extensibility • Automatic Security
The Claimed Successes of the PC
• Just over 50% of households have a PC, 75% of these will have home networks by 2005 – Still just under 50% that don’t, and the market is slowing • Improvement in dissemination of data – Information overload – important stuff is missed in amongst the dross • More and more sophisticated capabilities – Less and less users exploiting the full capabilities of the devices – 80% of users use 20% of facilities – Users getting frustrated at what they know they should be able to do! • Huge ease of use improvements from the early DOS based interfaces– Have you ever tried putting your granny in front of a PC???
We need to be careful…
We must not repeat the problems of the PC in the Networked Appliance space! We need simple, intuitive, reliable and easy to use devices, not feature laden monsters.
We’re already heading the wrong way…
HomePNA, HomePlug, 802.11x, Bluetooth, 802.3 etc…. UPnP, JINI, VHN, SOAP, Web Services, ZeroConf etc…. OSGi, .net etc….
Comparison..
• More computers than a jetfighter* • No (user perceived) reboots • Highly unfriendly environment (heat/cold/moisture/vibration/electrical noise) • Local area networking • Service integration • Self diagnosis • Intuitive UI (where applicable)
Specific Example…
Title 13 California Code 1968 "Malfunction and Diagnostic System for 1988 and Subsequent Model Year Passenger Cars, Light-Duty Trucks, and Medium-Duty Vehicles with Three-Way Catalyst Systems and Feedback Control." filed on 11-15-85.
Requires cars sold in California to have a on-board computer processor for on-board self diagnostics of computer sensed emission related components, fuel metering device and EGR (exhaust gas recirculation system). A partial or total malfunction that exceeded exhaust emission standard would illuminate a MIL (malfunction indicator light) and provide on-board identification of the malfunction location. To provide malfunction location information, codes are stored in on-board computer memory. To read codes manufacturers use methods such as flashing MIL light or various serial data protocols.
Why is this possible?
• Single integration agency (auto manufacturer) • Well respected and accepted standards • Safety critical environment demands through testing • Single identifiable product with strong market presence – no opportunity for buck-passing • Cost reduction imperative • Good understanding of the capabilities of the technology built up over years • User interfaces that are logical developments of existing approaches to product
Comparison with Home Networking…
• Single integration agency (auto manufacturer) • No defined integration path yet (individual users at best)
• Well respected and accepted standards • Emerging and fluid standards
• Safety critical environment demands through testing • Nope
• Single identifiable product with strong market presence – no opportunity for buck-passing • Lots of fragmented pieces – plenty of ‘it’s their problem’
• Cost reduction imperative • Good understanding of the capabilities of the technology built up over years • No clear identification of the application yet, much less the technology to address it
• User interfaces that are logical developments of existing approaches to product • By and large poor UIs based on ‘traditional’ computing technology
Things we need
• Appropriate Physical Infrastructures • Integrators (Network Gardeners?) • Simple, pervasive, pluggable standards • Simplification of product in the extreme • Innovative User Interfaces • A Market/Killer App/Whatever (central locking?) • Low cost point devices to plug into this infrastructure • Someone to take the lead
It doesn’t have to be like this
• Playstation/2 has USB, firewire and expansion ports – configuration is virtually zero • Firewire for video – very limited configuration – plug and go • USB – heading towards plug and go (driver issues) • Consumer electronics; how much configuration does your car need? • Can we make appliances according to the same ideas?
What I really want
• Plug a device in – it finds resources, configures to my network, and just works • When it needs upgrades it goes and gets them automatically • It Interworks correctly with other devices on my network and co-ordinates to offer joint services • Is intrinsically secure • Doesn’t require (heavy) networking skills to configure it – I don’t need to be an electrician to use electrical sockets!
Implications
• We need an agreed infrastructure at physical and application environment levels because these are the bits that interface to users… • We can get away without standard protocols _if_ the environments we create are flexible enough to be able to adapt and interwork • Political infighting will lead to the emergence of de-facto standards which will (indeed already has) limit market acceptance of appliance technology
Reality Check
• Home networking is complex – it will be complex for the foreseeable future because there are so many variables…sorry, but that’s how it is • We can compare this with telephony – similar levels of complexity, but it’s offloaded to an operator who takes responsibility off our shoulders for it • Complexity is distilled out and hidden behind simple questions such as ‘Would you like voice mail?’ • Question; Can a similar deployment model work for Networked Appliances? Answer; Yes, through OSGi, and potentially via .net
A quick pitch for OSGi
OSGi enables the delivery and management of services that can be accessed by devices that may be remote and/or have intermittent network connectivity. It defines a framework providing the capabilities required for these dynamic environments including a simple deployment model, remote management, and lifecycle management, amongst others.
To summarise OSGi
• OSGi represents a set of extensions to base JAVA to allow it to be used for large scale deployments where considerations of Reliability, Portability, Dynamicism, Security and Scalability are of paramount importance. • It consists of a core Framework and a number of optional Services. Many services are standardised by the OSGi. • The capabilities of the OSGi make working in long lifetime dynamic systems much less painful as much of the co-ordination responsibility is taken away from the programmer. • Most importantly – independent organisations can write services for deployment in an OSGi environment; A bundle market is enabled.
It’s not all doom and gloom…
• R&D in many large companies has been starved of funding for the last 18 months in an attempt to maintain bottom lines • In another 18 months to 3 years these companies are going to find their development cupboards bare and are going to have to buy in novel technology • Some Venture Capitalists have got money – they battened down the hatches long before the problems in the industry became acute • Do you have that bright idea??
My plea….
Help me to create Networked Appliances in general, and Home Networking in particular, that my dad can use, without having to get a Ph.D. in Communications and Networking. Thank you
|
|||