Power Usage Reduction Efficiency Improvements Discussion/Suggestions

hey guys so after a while using qubes on my laptop i’ve noticed that the battery demand is pretty drastic
i mean obviously i never expected a vm system to be on the low end but non the less…
i’d love to discuss different ways to reduce energy use/waste

any suggestions useless services/… to disable
ways to increase efficiency and so on

most solutions will probably not be too useful for most users such as my self with different trade offs in areas such as performance and security…

(and yeah for people such as my self who’re looking to scale and use qubes for it’s compartmentalization and security… well servers are expensive and so is power so obviously small improvements stack up)

but regardless i do think there should be some more discussion around here on the matter
because i was supersized how just how quickly my battery would run out

and i’m sure many feel like me

personally i don’t feel the need to sacrifice security/… for efficiency/power savings/… (at this moment at least it is cheaper or me to use qubes then hire a expensive expert)
but qubes is already fairly secure surely some users would love the trade off
using qubes in a bit lesser of intensity instead of using a different less secure os
and or a bear-bone os with a “normal” vm setup
which is obviously not ideal

anyway i hope you’ll give good advice and find the advice’s posted here useful

(and try to be as generic and wide ranging as possible people…:slight_smile: )

a couple of small advices for a start

i’d have to dig into the matter a bit more btw before i could argue which type of a os is more efficient for example… i assume fedora is a bit better then debian but idk
either way just an example

there are some things you can do like backups for example… which would compromise security but… if you were to just backup everything unencrypted and transfer into lets say an air gabbed encryption setup then you can probably improve efficiency (and security-depending on the situation) drastically…
dedicated hardware is much faster and more efficient
i mean hsm/tmp/… are great
but obviously such things depend on application

lets say… and this is a drastic over simplification but you have rout a network across 3 domains where in each you collect pockets/data/whatever… comands pings… again drastic oversimplification
lets say it’s sorta like an internal vpn that has to be shared between domains/users and encrypt all data
now you might think it’s best for the network to pass 3 different vm’s
for each of which the data is decrypted and encrypted before leaving to later arrive at you’re main server or something where you can decrypt everything

or something else entirely… again an oversimplified example i hope is clear enough
because i’m trying to be as generic as possible

lets say we’re talking about a database which for some reason you keep in 3 different vm’s a b and c…

a database which includes:
email/security question/…

and lets say (but keep in mind it’s a horrible idea to maintain a database this way… but again example)
you need all 3 to compromise an account 2/3 won’t do the trick…

so when you’re communicating with you’re main server/…
which has all decryption keys
you need to encrypt and send all data in an encrypted manner from you’re computer

what’s more efficient to encrypt all 3 databases/1 entry of them …on 3 different vm’s off the network

or is it more efficient to encrypt database a on vm 1 then send the encrypted version to vm 2 add database b to the encrypted database a
and send encrypted databases ab to vm 1 to be encrypted again before being send to you’re main server?

the answer is fairly obvious
(well if done at the same time and not accounting for other factors but u get the point)

and yes this is a poor example
but i’m trying to be as generic as possible
because each scenario application and requirement would be a bit different

surely the less vms one can use and the better in between vm’s networking setup is the better…
much better

Bet you can find useful things here:

1 Like