Energy, WiFi and RAM use by Android messaging apps | Hacker News Reader

Energy, WiFi and RAM use by Android messaging apps

decentim.grafana.net

106 pointsandrey_utkin35 comments

This really interesting, would love to see how Signal stacks up.

Not surprised in the least about Element unfortunately, it really screamed for a rewrite (one of the biggest problems was the horrible choice of database, which unfortunately was very hard to replace)

What was the poor choice of database on Element (Android or Electron), and how does it increase power usage?

Presumably realm in Element for Android. https://realm.io/ "Realm by MongoDB"

Yes

Sounds like the old Element app that was being used was using an always on notification to poll for push notifications instead of using a push service.

Element X only supports FCM cloud push notifications right now I think

I think it used push on play store and polling on f-droid

Correct, the novel thing Element X does different is sliding sync which allows for much faster partial syncs when the app is opened. The classic apps would have to do a much more substantial (and energy intensive) sync each time it was brought up to foreground.

The big drawback are the X versions are still very far from feature complete.

It can use (and I think it's the default) UnifiedPush on F-Droid

Default is pull, but if you have ntfy installed it shows a popup and lets you decide between pull or unifiedpush(ntfy).

Hi! I would like to show you my small analysis of Android messaging apps: power draw, bandwidth and RAM use.

Why did you not include Signal?

Was element (not the X version) using google play services for notifications or did you download it from F-Droid?

Signal?

I scrolled through the document and I don't see any measurements of the number of received messages - that will pretty much dominate the standby power draw of any messaging app (rule of thumb is - 1s of modem/CPU wakeup takes away 1min of standby time).

So the data can be seriously meaningless if you don't control for the amount of messages received and their interval (spread out messages hit the power draw more than batched - phones usually race to idle).

It's a good thing that I have disabled Gmail app on my device and use k9 for all my email accounts. Seems like I should stop using WhatsApp too.

K9 is great and has worked well for me for years, very impressed with its numbers here.

According to Battery Historian [1], WhatsApp was the single biggest cause of battery drain on my phone.

1. https://developer.android.com/topic/performance/power/batter...

It's probably single most used app on your phone too

No, that would be my browser. My browser also accounts for most of the screen-on time, whereas most of WhatsApp's usage is in the background.

Of course, your browser isn't receiving and retrieving messages in the background every time someone sends it to you.

Neither is Whatsapp, thanks to push notifications.

Tangent, I find Instagram to be one of the most egregious battery drainers amongst mainstream apps. A former IG employee made a claim here on HN that the inefficiency was due to ghastly experiments being run on clients devices. Shame, I cannot seem to find that comment.

Seems like measuring the median would be misleading

Can someone summarize the findings? I genuinely can’t read it well. The tables are too wide for the viewport. If I make the browser wider the page becomes two columns and the viewports are still too small to show the full table.

The sections of text also appear to be longer than the boxes they are in, causing scrolling inside which I only discovered after a few minutes and making it hard to read.

Honestly it’s quite frustrating. Which is too bad I really liked that this testing was done.

Yeah this 2-column layout is super weird.

> Conversations (XMPP) seems best for battery life, traffic and RAM.

> Facebook Messenger and Element (Matrix) are the worst from the studied apps, draining up to 300x more power than Conversations when battery optimization is disabled. Keep battery optimization enabled for them, I guess. Need more data to see better.

> Element X (experimental Matrix app) is coming up with a more than 100x reduction in energy consumption (comparing to Element), drawing just 2x more power than Conversations.

Not surprised they've found Element to be so inefficient. It used to drain a LOT of my battery, despite me never actually using it. I find myself disappointed with Element apps in general (it's why I use Nheko Reborn on desktop).

I guess it's time to uninstall Element.

Have been running it alongside Element X so far, but the energy consumption does not justify the few additional features

the conclusion isn't good enough?

Conclusions Conversations (XMPP) seems best for battery life, traffic and RAM.

Facebook Messenger and Element (Matrix) are the worst from the studied apps, draining up to 300x more power than Conversations when battery optimization is disabled. Keep battery optimization enabled for them, I guess. Need more data to see better.

Element X (experimental Matrix app) is coming up with a more than 100x reduction in energy consumption (comparing to Element), drawing just 2x more power than Conversations.

A few more proprietary messengers have been measured and they fall in between the space between the above mentioned extremes. Little can be said about them besides that most of them continuously draw substantial amount of power and bandwidth even when not interacted with.

I got down to about the graphs before I gave up.

If there was a conclusion beyond that, I’m sorry but I missed it.

This site reminds me of the early 90s when people would make frame layouts with dozens of individual little tiny frames and put different content in every single one.

There’s a reason no one does that anymore.

No surprise for Facebook Messenger. Facebook have a Messenger Lite app with exactly the same feature set ,for years. No idea why they still have the non-lite version as default.

they recently removed the lite version.

The tables are too wide for the viewport. If I make the browser wider the page becomes two columns and the viewports are still too small to show the full table.

The sections of text also appear to be longer than the boxes they are in, causing scrolling inside which I only discovered after a few minutes and making it hard to read.

I saw the domain, and was not surprised to see this comment here. Frankly, this is what a ton of "modern" software is like. Flashy, bloated, and overmarketed with buzzwords, but just barely accessible.

A plain HTML page with regular tables and images would be far superior.

I haven’t seen a webpage that looks like this since the early 90s when people would use 12 frames to make their website “cool“ and “different“.

It's a bunch of numbers that don't even seem to be controlled for the amount of messages sent and received. Or usage time. Or run time.

It's... wierd.

Reply on news.ycombinator.com