Before coming to Monash, I had been attempting to set up a Jini demonstration laboratory, with the cooperation of DSTC and possibly other parties such as Telstra Research Labs and Sun Microsystems. This was put on hold until I finally came here, but now is probably a good time to revive the idea.
Jini is not just about putting toasters on the internet, and the majority of Jini applications are primarily about making software services available. Nevertheless, it is more convincing to demonstrate the value of Jini by using actual physical devices than just pieces of software. In addition, there are a large number of efforts to build devices that may use technologies other than Jini to make themselves into internet devices.
It makes sense to combine the Jini-specific technology and the device technologies into one laboratory, which demonstrates a number of internet devices, with a focus on Jini but using other technologies where appropriate.
A demonstration laboratory must be accessible to be of any use. One way of doing this is to make every component portable so that it can be taken from place to place. One of the early Sun demos of Jini did exactly this, carrying a collection of devices around. Observing the demonstrators, this was a fairly nerve-wracking task, as there were often last minute difficulties and hiccups in setting up the demos.
An alternative is to have a fixed laboratory, but make its visibility portable. This can be done by means of a camera connected to the internet, so that observations can be made remotely of what is happening in the lab. This is the method proposed for this lab.
The School of Network Computing is proposing a new subject "Internet Devices and Services" to be a subject in the Masters of Network Computing. It will also be an option in the Master of Information Technology. This subject would benefit by having a demonstration lab of internet devices available. The proposed laboratory could meet this need.
Where possible, devices that connect directly to the internet would be preferable to devices that require a PC to perform the connection. Currently, devices that connect directly tend to be more expensive than those requiring a PC, because they need their own operating system, network cards and TCP/IP stack. For example, an internet camera is available, which runs Linux with 16M RAM.
Many devices cannot connect directly to the internet as yet, and require a PC to do this. The Lego Mindstorms robots are examples of this. Usually the PC is not required to be anything near state of the art, and an old PC with 32M RAM will often suffice.
The last category of device is not even computer aware, and may need some custom circuitry built. An example would be a lamp, where the current to the lamp could pass through a switch controlled from the serial port of a PC.
The proposed lab would have an initial setup of
The camera would be permanently running and connected to the internet. Currently the camera is not a Jini-device, so can be accessed directly (it runs it own http server) or can be proxied to make it appear as a Jini device. This will allow both Jini and Web clients to access the camera and observe what is going on in the lab.
The lamp and the radio run as Jini services which can be accessed by clients on the local network. I enter the lab and switch on my laptop with a Bluetooth card. When I walk into a room, I like to be able to see and to listen to the radio. So once my laptop joins the local Bluetooth network it is able to find the lamp and radio services. It asks for the lamp to be switched on, and the radio to present a list of radio stations that can be played.
Note that Bluetooth rather than an 802.11 Wavelan (WiFi) wireless connection is proposed precisely because of distance limitations of Bluetooth: I only want "close" rooms to switch on their lights, not every room within two floors. Preferably, a "room centric" network would be even better.
There is also a Lego Mindstorms robot in the room, just because they are fun to play with, but demonstrate a serious capability. This is available as a Jini service that can be discovered by my laptop so I can start driving the robot around.
To demonstrate further home use capabilities, a set of grocery items can have RFID tags attached. They are scanned by an RFID scanner and placed in a "pantry" - which is both physical and a Jini service. A "recipe" service can then interrogate the pantry service to figure out what you can actually cook, given the ingredients in the pantry.
| Item | Cost |
|---|---|
| Axis 2100 Internet camera | 1100 |
| Internet radio | ? |
| Lego Mindstorms | 200 |
| lampstand | 100 |
| Intersoft RFID demo kit | 400 |
| PC Bluetooth card | 500 |
| Laptop with Bluetooth card | 5000 |
| two low-end PCs | 500 |
| serial controlled switches | 50 |
| Total |
To be a resource funded by the School of Network Computing, the lab should be located at Peninsula. However, it would be used by MNC students in the subject "Internet Devices and Services". With the relocation of the MNC to Caulfield, it would be better to locate it at Caulfield. DSTC would also take a greater interest in the laboratory if it is located at Caulfield.