PHOTOTROPISM

[corylus]

Hi... I'm new to these forums, but first of all, good job!

 

I am interested in looking at how Mother Nature works for inspiration in ICT developments.

 

Currently I am studying Phototropism - in nature plants tend to grow toward the light (a phenomenon called Phototropism). Although some light is transmitted through the protector wall, there is more light higher up, especially light directly entering the opening at the top of the shelter. Plants will try to reach that light.

 

Applying the same principals to computing capacity and power / performance requirements is an interesting approach to resource usage optimisation - where part of the inherent function of 'the processor' is to 'reach' for the greatest light, or resource, output or availability. This negates the software, be it OS or application, having to do just that and in theory frees up the resources wasted in them doing so.

 

I'd be very interested in your thinking on this area.

Edited October 15, 2008 by corylus
Grammar

[dhicks]

Although some light is transmitted through the protector wall

 

I assume you mean those brown plastic widgets that are tied around saplings to stop deer eating their bark?

 

Applying the same principals to computing capacity and power / performance requirements is an interesting approach to resource usage optimisation - where part of the inherent function of 'the processor' is to 'reach' for the greatest light, or resource, output or availability. This negates the software, be it OS or application, having to do just that and in theory frees up the resources wasted in them doing so.

 

Isn't this basically just moving the process/thread management code from the OS to the processor, i.e. implementing process/thread schedualing in hardware? I should have thought modern processors would have supported that by now, or would at least be on the way. Or are you proposing a new scheduling algorithm of some kind?

 

--

David Hicks

[corylus]

Unfortunately, silicon protectors are unable to transmit efficiently and cause the process to divert its energy to vertical 'growth' at the expense of performance growth. This makes the process 'top-heavy' and vulnerable to stalling. It is also a major cause of dieback and increased mortality.

 

Pro protectors transmit exactly the right amount of information to stimulate proper expansion without producing an excessive amount of heat.

 

And this is where we get to the 'Vented Vs Unvented' debate... I'll post more on this when I have completed some research on the side.

Edited October 15, 2008 by corylus

[dhicks]

Unfortunately, silicon protectors are unable to transmit efficiently and cause the process to divert its energy to vertical 'growth' at the expense of performance growth. This makes the process 'top-heavy' and vulnerable to stalling. It is also a major cause of dieback and increased mortality. Pro protectors transmit exactly the right amount of information to stimulate proper expansion without producing an excessive amount of heat. And this is where we get to the 'Vented Vs Unvented' debate... I'll post more on this when I have completed some research on the side.

 

I'm sorry, I have no real idea what you are talking about. Are you talking about trees here, or silicon chips, or other sorts of chips, or what? You'll need to give us some more background here.

 

This forum is mainly for people who run computer systems for schools - we're used to the day-to-day stuff of fixing grotty Windows PCs, sorting out servers, that kind of thing. If you want to discuss more abstract programming tasks you might be better off at somewhere like stackoverflow.com, although I'd suggest making sure you give a good outline of whatever it is you are investigating.

 

--

David Hicks

[Guest Rufus_Ruffcut]

Hi Corylus,

 

This is my first post to a great forum for like minded people. It was great to find your post first post and such a passion for ICT in the environment! I'm based in a rural part of Eastern England and find phototropism is being being ignored in ICT by a great deal of ICT support staff in education. I have been lucky to move my server room to a more intense light source with excellence light to humidity ratios. There has been a marked increase in performance of my 'splitting technique' speed within the raid array and following information from some industrial sources I have got my data drying cycle down to under 12 months! I look forward to hearing about your further research and I will keep you updated with any new developments I have. I am about to embark on a light to bandwidth project across the site, to see if data is effected by the shorter days via the response of the phototropism characteristics.

Rufus

[webman]

Are you suggesting we should install large sun lamps in our server rooms?

[richard.thomas]

Haha, Love it!

Guys, we need to get rid of the A/C and install sprinklers! Let our servers grow!!

 

Either I'm being really stupid or this is a joke...

[Guest Rufus_Ruffcut]

Hey Richard,

 

This is no joke, have you read the latest BSF documents and other info from Partnership for Schools? The carbon footprint and efficiencies are high on the agenda. I have been oxygenating my server room for a couple of years now and I have noticed a marked response to reliability due to a stable atmosphere and better ying and yang. My site can't afford air con for the server room so light to encourage plants is effective. Obviously we don't use sprinklers! On the roof I have planted some Sedum which increases insulation and constant temperature. Phototropism is the new Nevana!

 

Rufus

[localzuk]

I'm sorry but you are blurring science with personal beliefs and myth.

 

A higher oxygen level in a server room should not affect anything - most components in a server are sealed - ie. the chips themselves.

 

If too much oxygen is available, would this not simply lead to a higher chance of oxidation of components, and therfore lower the life expectancy? For example, hard disks use iron oxide as their medium, so therefore increasing oxygen can damage that surface can it not?

 

Increasing airflow would increase reliability much more efficiently - ie. having intake and exhaust fans at different points in a server room. This lowers the temperature, increases airflow and should also prevent the need for air conditioning, if the room is not over filled with servers.

 

Also, can someone explain what 'Data Drying' is? Seems like a made up concept to me.

 

To be honest, this entire thread reads like techno-babble, making little to no sense what so ever.

[maniac]

Must admit I've been a tad confuzzled by this thread, the initial idea sounded a bit odd, but interesting none the less, but googling reviels very little about this idea.

 

The only thing I can find a lot of data on is the idea of ditching air conditioning in favour of better air flow through the server room, and getting the optimal temperature for your servers to work in can in principle have positive affects, as every matter has a temperature zone where it's physical properties change, so if you can find the optimal temperature to run your servers at, in theory you can extract better performance from them. However I'm skeptical that this has any major benefits in the long term.

 

Mike.

[DMcCoy]

This thread is most amusing.

 

Nonsense, but still amusing

 

On a slightly more serious note airflow also stops dehumidification, as too low humidity increases static build up.

 

Edit: Airflow being from outside to inside rather than recirculation.

Edited October 16, 2008 by DMcCoy
Um, to edit?

[dhicks]

This thread is getting rather like a game of Mornington Crescent - I just need to figure out the rules...

 

The only thing I can find a lot of data on is the idea of ditching air conditioning in favour of better air flow through the server room

 

Ah, the old "passive ventilation" trick. We had a building at university that was built like that - all the hot air was supposed to rise up the building, drawing in cool air from the bottom and creating a cooling effect and making sure all the rooms were ventilated without having to power fans. We damn near asphyxiated in the robotics lab.

 

Seems silly, though, not to do something with all that waste heat that our servers / workstations / children spew out the whole time - I've sometimes thought a ventilation shaft running from the server room to a polytunnel-style greenhouse would be an idea.

 

--

David Hicks

[srochford]

Nah, you don't want air-con or oxygen in your server room. The way you do these things properly is with liquid carbon dioxide cooling. check this for the way things are done in the real world :-)

[GrumbleDook]

no, no, no, no, no ... erm ... yes.

 

The secret of using plants in your server room is not down to oxygenisation at increased levels allowing for faster flow of electricity at higher levels in the processing core ... it is down to the metaphysical benefit of absorption of negative vocalisation whilst dealing with certain products. The larger the leaf size the better as this gives a greater chance for the the sound waves to be absorbed rather than reflection from the walls and floor. Where possible you should place them in from of glass fronted cabs as there is a great risk here due to refraction of these negative vocalisations, because when this occurs you get a breaking down of the benign sighs, the mild expletives, the harsh words and the snarling diatribe ... and the degrading feature of these snarling diatribes mean that the general performance of your network is lowered.

 

I hope this helps to clear things up.

[conehead]

Ohh..snarling diatribes are lovely, I've got a greenhouse full, and like the venus flytrap they remove unwanted bugs when kept in a Server room.

[webman]

Ohh..snarling diatribes are lovely, I've got a greenhouse full, and like the venus flytrap they remove unwanted bugs when kept in a Server room.

 

Is that cross-platform or does it just work on Windows?

[Little-Miss]

[maniac]

Is that cross-platform or does it just work on Windows?

 

:D:D

[conehead]

Is that cross-platform or does it just work on Windows?

 

You could build any sturdy platform on which it will thrive, Windows, shelves, anywhere.

 

Though I believe it is prone to choking on bloatware worms.

Edited October 17, 2008 by conehead

[webman]

Hmm... that's planted a (random) seed in my head!