Apps & Design

[Hood, Tom]

Being an ex-communications guy, I can address question 3 anyway...

"jok707s at mail.smsu.edu" wrote:
> 
> Several questions:
<snip>
> 3a. Has anyone successfully assembled a completely wireless Beowulf with all
> of the nodes and the server connected to each other only by radio (or maybe
> infrared)?  Could this be done with cellular technology?

Wireless TCP networks are in use today.  That said, the speeds across
the links are pathetic to say the least.  The best throughput I was able
to achieve was using multiplexed microwave, and we never got more than
10Mbps.  Our experiment was done several years ago, and I imagine
someone has built a better uwave rig by now.  RF (radio) is not very
efficient and requires loads of power.  Cellular (in it's commercial
configuration) has so much overhead that the percentage of data to
carrier is way too low to sustain high data rates.  As I was leaving the
job where I did this work, we were starting to play with lasers and
getting great data rates, but all the nodes had to be fixed in place,
the range was quite short (less than a mile), and there had to be clear
line of sight between transmitter and receiver.  Extremely secure
however...


> 
> 3b. If a wireless Beowulf can be made to work at all, would it be possible for
> its performance to get into the general area of the wired versions using
> technology available in the foreseeable future?

Yes, with lasers or high power uwave.

> 
> 3c. If a wireless Beowulf were constructed entirely from laptops (including
> the server), could one have a "virtual" Beowulf with the individual nodes
> moving around while the system as a whole continued to function?  How much
> would performance degrade as the nodes got farther apart?  What would be the
> distance limits beyond which the system would fail completely?  How much
> difference would be made by the application(s) that were being run?

Yes, using RF (any band) or uwave, you can have the nodes be mobile. 
Performance does not degrade until SNR gets to the point where error
correction begins to require packet re-transmission.  Depending on the
frequency used and the power of the transmitters, this can range from
yards to miles.  For most RF bands (up to Extra SHF), the range is
measured in 10's of miles at best depending on antenna characteristics. 
In the uwave range, best results are achieved by bouncing the signal off
satellites to ground stations, not directly peer to peer.  The downside
to this is the power consumption.  There's also a human physical problem
with being next to a high power uwave transmitter for any length of
time.


Thomas Hood
thood at ifn.com

> 
> This should keep people emailing for a while?  :-)
> 
> Joel
> 
> _______________________________________________
> Beowulf mailing list
> Beowulf at beowulf.org
> http://www.beowulf.org/mailman/listinfo/beowulf