From sullivan at cs.virginia.edu  Fri Jan  7 17:12:38 2005
From: sullivan at cs.virginia.edu (Kevin Sullivan)
Date: Wed Mar 22 16:00:38 2006
Subject: [Hog] Bandwidth requirements
Message-ID: <200501072212.j07MCaXg019787@ares.cs.Virginia.EDU>

Guys,
 
I spent a little time thinking about our bandwidth requirements. Suppose we
want to send H messages per second to each of K devices, where each message
comprises a device identifier and a data packet of D bits. The minimum
individual message length B is thus log_2(K) + D, and the minimum bandwidth
required is B * H * K. Let's assume that message lengths come in increments
of 8 bits. The numbers say for 64K devices and 1 byte data per message at
one cycle per second we'd need a bandwidth to each device of 1.6 MBit/sec.
That's basically 802.11b. I don't know how large, expensive, or power-hungry
802.11b receivers are. I'll look into it.
 
Kevin
 
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.Virginia.EDU/pipermail/hog/attachments/20050107/5b54ab1e/attachment.htm
From sullivan at cs.virginia.edu  Fri Jan  7 17:29:46 2005
From: sullivan at cs.virginia.edu (Kevin Sullivan)
Date: Wed Mar 22 16:00:38 2006
Subject: [Hog] CSR unveils 'lowest cost' Wi-Fi chip | The Register
Message-ID: <200501072229.j07MTjXg021556@ares.cs.Virginia.EDU>

It looks like one-chip WiFi cost is in the vicinity of $8. A small 8-bit
processor is available for about $0.50. We'd also need a circuit board,
battery power supply subsystem, I/O devices (minimum LED and a few buttons),
and packaging. I don't yet have any real intuition about whether it can all
be made to work. The bandwidth requirement seems to be the main cost driver.

http://www.theregister.co.uk/2004/11/09/csr_wi-fi_chip/ 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.Virginia.EDU/pipermail/hog/attachments/20050107/4911a793/attachment.htm
From sullivan at cs.virginia.edu  Fri Jan  7 13:07:57 2005
From: sullivan at cs.virginia.edu (Kevin Sullivan)
Date: Wed Mar 22 16:00:38 2006
Subject: [Hog] Technical question for you
Message-ID: <200501071807.j07I7uXg009059@ares.cs.Virginia.EDU>

Hi Ron,
 
Happy New Year. Are you willing to field a quick technical question?
 
I'm trying to find out what kind of technology is available (or potentially
realizable today or very soon) meeting the following requirements:
 
* cost, at least in production, somewhere in the range from under $1 to no
more than $5, or perhaps $10 at the very most
* device/board form factor small enough to be suitable for incorporation
into single-hand-held device
* supports wireless download of small amounts of information to the device,
according to an address that it recognizes
* minimal computing ability -- we can fit in 1K byte; minimal processing
except for the need to continually tell if "this" device is being "talked
to" by wireless
* can run on a very small battery for at least a few hours -- e.g., a
hearing aid battery
* minimal I/O capability -- e.g., output by illumination of say 3 LED's;
similarly minimal input, e.g., something like 3 mouse buttons
 
I don't know anything really about RFID tags, but the ideal would be some
kind of RFID tag on steriods -- minimally programmable, minimal I/O,
wireless, and at the right price point. I don't need the "passive (no
battery)" aspect at the moment, and, at least in the initial work, I don't
need capabilities to upload any data from the device. I just want to
communicate from a central point out to a set of individually addressable
devices. Any ideas on what kind of devices I should be looking at (or
thinking about designing)? Any thoughts on feasibility (if not currently
available)? Thanks
 
Thanks!
Kevin
 
P.S. Both Skadron and Davidson suggested independently that I talk to you.
That's how I got here, to your doorstep. Hope you're having a good rest
before classes start.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.Virginia.EDU/pipermail/hog/attachments/20050107/f6f78cd4/attachment.htm