iOS multitasking - 3 minute background limit?!

[mi7chy]

Figured I'd give something for my iPad to do. IRC client is pretty simple and doesn't really need the power and energy consumption of a PC so an ARM device is ideal. Tried a couple of iOS IRC client apps and to my surprise discovered they all have a limitation of running in the background for only 3 minutes. Even when there's only one task running, the IRC client, and nothing else if you switch to the home screen IRC dies after 3 minutes.

Apparently this is a confirmed iOS limitation and vSSH also confirms "Background work (up to 3 minutes for iOS 7/8, up to 10 minutes for iOS 5/6)". This pretty much eliminates the ability to run any persistent background connection such as SSH, telnet, ftp, ftps, sftp, etc. like any other true multitasking OS can do.

Since background tasks with persistent connection on other mobile OS like Android work correctly and have no significant impact on battery life (measured 2.7 days with $25 Moto G running background IRC) it was probably an Apple design decision to forgo true multitasking so that they can ship devices with minimal DRAM to save on BOM cost since most iOS devices have only 1GB DRAM while the industry standard is 2 to 3GB DRAM.

Ended up going with a solution that does do persistent background IRC correctly with a Moto G for $25 (better than iPhone 6) and the best IRC client on any mobile platform. Going to try running an IRC server next.

Unfortunately, iPad goes back to collecting dust until Apple fixes this limitation.

 

[Bone_Enterprise]

Okay?

 

[CEpeep]

InfinityTask will do exactly what you're looking for. Grab it off Cydia and tell it to run on TurboIRC.

 

[Mchart]

I like how Kyle deletes posts because they are 'off topic' when the OP's own post goes off topic. Did you read the original post Kyle? Just delete the thread.

He begins with iOS multitasking issues. OK, whatever. (Should be under the Apple section)

Then he for whatever reasons brings up a moto G and a iPhone 6, both of which have nothing to do with ipad multitasking.

The OP as is typical baits people and then reports posts.

Anyways OP, I'm sure all 5 people that care about running IRC servers on their cell phone will take your warning seriously.

 

[Johnny78]

He begins with iOS multitasking issues. OK, whatever. (Should be under the Apple section)

Then he for whatever reasons brings up a moto G and a iPhone 6, both of which have nothing to do with ipad multitasking.

That's weird, I don't recall the Hardforum rule that states you can only discuss one subject and one subject alone per post. Please point that rule out for me because I missed it.

Fact is, Michy is discussing phones and phone related operating systems within a subsection called, "smart phones and devices". That sounds pretty legit and perfectly within his rights. Although I didn't read what Kyle deleted, I would bet my left nut he deleted posts that just attacked Michy and did not have any bearing on the subject matter itself.

I get it, he annoys and rubs some people the wrong way but the people who act so butt hurt are even more annoying. It's just phones people. If you don't like him, use the ignore button.

 

[Ocellaris]

I haven't read mi7chy's post (he's on Ignore, though I suspect he kept up his Moto G trolling) however iOS is very aggressive at prevent background tasks from killing a user's battery. Android just lets apps run wild and this is often why you see users complaining about super short Android battery life. Background task manager apps are very popular on Android since apps are essentially unregulated when left on their on in the background. There is a white list of stuff iOS apps can do in the background, maintaining a constant connection to anything isn't available.

As one example, we work on an app that can sync data from the phone to a website, however this only happens when the app is in the foreground for iOS. On Android we can set it up so that it syncs huge blocks of data hundreds of times a day and Android doesn't care.

Fact is, Michy is discussing phones and phone related operating systems within a subsection called, "smart phones and devices". That sounds pretty legit and perfectly within his rights.

There is a fine line between "discussing" and trolling.

 

[mi7chy]

Apparently, this 3 minute background limitation affects any persistent TCP socket connection so other protocols are affected such as telnet, SSH, ftp, sftp, ftps, etc.

vSSH confirms the following under description.

https://itunes.apple.com/us/app/vssh/id527244258

Background work (up to 3 minutes for iOS 7, up to 10 minutes for iOS 5/6, alert on timeout)

Already got semi-banned and redirected by Freenode to a #check-network-connection channel for persistent dropped connections/reconnections. With SSH client you can potentially lock yourself out of a remote network device since it has a finite vty limit.

InfinityTask will do exactly what you're looking for. Grab it off Cydia and tell it to run on TurboIRC.

Thanks. I'll keep it in mind after evaluating the client apps. Hopefully, iOS 9 fixes this.

 

[Mchart]

That's weird, I don't recall the Hardforum rule that states you can only discuss one subject and one subject alone per post. Please point that rule out for me because I missed it.

Fact is, Michy is discussing phones and phone related operating systems within a subsection called, "smart phones and devices". That sounds pretty legit and perfectly within his rights. Although I didn't read what Kyle deleted, I would bet my left nut he deleted posts that just attacked Michy and did not have any bearing on the subject matter itself.

I get it, he annoys and rubs some people the wrong way but the people who act so butt hurt are even more annoying. It's just phones people. If you don't like him, use the ignore button.

Oh? Then the posts that were discussing the tired Moto G / iphone 6 comment he posts in every thread in this section shouldn't have been deleted.

 

[CEpeep]

Thanks. I'll keep it in mind after evaluating the client apps. Hopefully, iOS 9 fixes this.

There's no change in this behavior in iOS 9, at least at this time. It's not a bug. Apple recommends that apps that require persistent TCP connections use a server to maintain the connection and then issue push alerts to the client device. For your particular use case, Android is a much better choice, since it does not force apps to behave in ways that preserve the device's battery life.

 

[Mchart]

There's no change in this behavior in iOS 9, at least at this time. It's not a bug. Apple recommends that apps that require persistent TCP connections use a server to maintain the connection and then issue push alerts to the client device. For your particular use case, Android is a much better choice, since it does not force apps to behave in ways that preserve the device's battery life.

I was going to say that I'm surprised that an IRC client for iOS doesn't have this push behavior back end added in. Regardless of the client being iOS or Android it would make a lot more sense from a battery life perspective to build an IRC client with push support due to the battery life fact. I realize this would require way more work on the application developers part (They'd have to develop a server back end that would do the actual interfacing with IRC and then develop the actual APP to interface with their go-between server) but I'd find this more useful then always having an application burning in the background just to keep a stupid IRC connection up. (I don't even use IRC anymore so meh)

 

[CEpeep]

I was going to say that I'm surprised that an IRC client for iOS doesn't have this push behavior back end added in.

Many do, but not TurboIRC.

 

[mi7chy]

I wonder if it's a legacy thing that's carried over without relevance because background tasks on the $25 Moto G doesn't seem to affect battery life. Just fired up background IRC on Moto G and will check back in about 8 hours which is how long it lasts with screen-on-time under normal use.

May have to jailbreak and try InfinityTask to satisfy curiosity of battery life delta with and without background task.

 

[mi7chy]

Many do, but not TurboIRC.

I saw IRCCloud while checking out the available apps but the drawbacks are possible monthly service fee, the app has low 2 star rating for current version and I'm not a fan of middle man service in regards to security. One of the reviewers commented:

After creating an account with this app, for the first time ever, somebody tried to login into my facebook account using the same password used with this app. Not sure if this is usually possible, but it happened a couple of hours after registeration. Location was OH, USA, using IE from Windows 7.

 

[Mchart]

I wonder if it's a legacy thing that's carried over without relevance because background tasks on the $25 Moto G doesn't seem to affect battery life. Just fired up background IRC on Moto G and will check back in about 8 hours which is how long it lasts with screen-on-time under normal use.

May have to jailbreak and try InfinityTask to satisfy curiosity of battery life delta with and without background task.

You really think running an IRC client in an active chat room won't effect a devices battery life? Every single message in the chat room will have to use the cell radio. If you use push you can set it to just push you a single time for an update, and then no further pushes until you open the client up again.

 

[sethiano]

Figured I'd give something for my iPad to do. IRC client is pretty simple and doesn't really need the power and energy consumption of a PC so an ARM device is ideal. Tried a couple of iOS IRC apps and to my surprise discovered they all have a limitation of running in the background for only 3 minutes. Even when there's only one task running, the IRC client, and nothing else if you switch to the home screen then after three minutes IRC dies. Apparently this is a confirmed iOS limitation. This pretty much eliminates the ability to run any persistent background client or services on iOS.

http://i990.photobucket.com/albums/af21/mi7chy/TurboIRCTimeout_zpsjqdtf7sy.jpg

Ended up going with a solution that does do persistent background IRC correctly with another new Moto G for $25 (better than iPhone 6) and the best IRC client on any mobile platform. Going to try running an IRC server next.

http://i990.photobucket.com/albums/af21/mi7chy/Screenshot_2015-06-22-13-12-25a_zpsn0ox8o0v.png

iPad goes back to collecting dust.

Do you know if this is the same in iOS 8 or only 7?

 

[Ocellaris]

Do you know if this is the same in iOS 8 or only 7?

It is the same in all version of iOS since they allowed any level of background processing. Apple wants developers to make efficient apps. Developers should use push notifications to alert the user when something happened.

 

[AltTabbins]

Do you know if this is the same in iOS 8 or only 7?

Its been the same through IOS history. It more than likely wont change because it is not considered a bug or a negative thing. It's a part of the app policy and Apple feels its a good thing because of the positive affect it has on IOS device battery life.

 

[mi7chy]

Do you know if this is the same in iOS 8 or only 7?

10 minutes for iOS 6 and 5.
3 minutes for iOS 8 and 7.
Haven't tried iOS 9 beta but supposedly 3 minutes.

 

[AltTabbins]

Edit: Nevermind.

 

[mi7chy]

Curious as to why Apple decided not to support true multitasking in iOS and hearing an unofficial reason being related to battery consumption I set out to test this theory. Using a $25 Moto G since it's based on the same ARM architecture and running background IRC, at 21 hour 47 minute the battery was down to 66% so it should be able to go about 2.7 days on a full charge and IRC didn't die. So, Apple's decision is unlikely due to battery life and probably more to do with restricting background tasks so they can ship with minimal DRAM to save on BOM cost since most iOS devices have 1GB DRAM while industry standard is 2 to 3GB DRAM.

 

[CEpeep]

If you fire up InfinityTask and actually run a benchmark on the device in question (instead of running a benchmark on a different device with a completely different operating system and trying to extrapolate the results) I'm curious what your findings are.

 

[Mchart]

Apple's decision making to push apps towards supporting push is because most people use their cell phone as a cell phone and constant TCP usage over the cell radio for running an IRC client would destroy battery life on ANY device. I'm glad you like your moto G and likely never leave the house with it and have constant WIFI access. Although I'd question why battery life even matters to you at that point. It's laughable that you are using this as a reason to bash on the iPhone when you should be wanting an IRC client that has proper push support regardless of the OS choice / hardware choice.

Run your own test with WIFI turned off on your moto G running that IRC client and get back to us. I want to see the test NOT running the client and normal usage for you w/ the cell radio, and a test with the IRC client running w/ cell radio. Finally, as stated, you don't even have results from an iPhone 6 to compare to, although I will say your moto G is hardly impressive considering I can leave my iPhone 6 unplugged running all data through wifi and have stuff that runs constant push updates in the background anyways. I usually end up with about as much battery life left.

 

[Deleted whining member 223597]

Battery life with Airplane mode...that's really useful.

 

[Mchart]

Battery life with Airplane mode...that's really useful.

Well, to him it is I guess. He likely buys all these devices but never actually uses them like most people do. They're just wireless devices that get used in his house or when he needs to go out and buy more fritos and soda at the Jewel-Osco.

Wow, I didn't catch the airplane symbol. If that's the case then those battery results are actually pretty horrid if it's down to 60% even w/ the cell radio completely turned off.

 

[Deleted whining member 223597]

Well, to him it is I guess. He likely buys all these devices but never actually uses them like most people do. They're just wireless devices that get used in his house or when he needs to go out and buy more fritos and soda at the Jewel-Osco.

That's what has always bothered me about a lot of the battery life screenshots people post up and where they recommend you to shut off all the background stuff to increase battery life. I have emails coming in every 15 minutes and every app that updates does so very often, that's just how I want my phone to be.

 

[Mchart]

That's what has always bothered me about a lot of the battery life screenshots people post up and where they recommend you to shut off all the background stuff to increase battery life. I have emails coming in every 15 minutes and every app that updates does so very often, that's just how I want my phone to be.

Off topic but I own almost every current and last gen big Android phone release and with my daily routine all but my Z3 Compact burned through more battery then my iPhone 6 does. Sure, my Android phones -are- running a number of apps/widgets that poll every 15-30 minutes or so, but I have apps/widgets on my iPhone 6 that do the same anyways but because they are built to only update when the device comes out of sleep (or push ONCE. I only need a SINGLE push just to know that a specific app has an update! Not 50!) I think that contributes to why iOS devices have better 'real world' battery life. Many Android apps are just built lazily and don't have proper push support to really conserve on battery like you find with iOS apps. (because they have to)

 

[mi7chy]

If you fire up InfinityTask and actually run a benchmark on the device in question (instead of running a benchmark on a different device with a completely different operating system and trying to extrapolate the results) I'm curious what your findings are.

To be honest I'm not too anxious about installing Chinese software on my PC and iPad for jailbreaking and circumventing the security of stock ROM. Also, prefer getting my apps from the official store. Maybe I can borrow one that is already jailbroken or if time permits set up a temporary PC/VM that I can blow away afterwards.

 

[Mchart]

With iOS 9 / xcode 7 you'll be able to run unsigned apps/code without having to jailbreak anymore.

 

[mi7chy]

Battery life with Airplane mode...that's really useful.

Go spend $25 so you're not cross platform ignorant. Since I use it on WIFI only of course it's going to be in airplane mode with always-on WIFI since cellular radio isn't needed. My iPad is WIFI only for comparison. Why would I pay for a bunch of plans when I can do everything including VOIP on WIFI or WIFI tether to my primary phone.

 

[CEpeep]

To be honest I'm not too anxious about installing Chinese software on my PC and iPad for jailbreaking and circumventing the security of stock ROM. Also, prefer getting my apps from the official store. Maybe I can borrow one that is already jailbroken or if time permits set up a temporary PC/VM that I can blow away afterwards.

I guess? Previous version of the jailbreak are considered safe and have been vetted by security professionals like Stefan Esser, but I understand the concern with the very latest version that was just released. Either way, it doesn't seem that iOS devices have much utility for you in their stock state. Many of your complaints about the platform are easily resolved by jailbreaking the device, so I'm more than a little surprised you haven't already taken the plunge. After the new 8.3 JB is vetted, definitely go for it. You'll be glad you did.

With iOS 9 / xcode 7 you'll be able to run unsigned apps/code without having to jailbreak anymore.

Only if you have the source for them and can compile them yourself. Also wouldn't let you install something like InfinityTask anyway since it needs to escape from the app sandbox in order to work, and that requires jailbreaking/root access.

 

[Mchart]

Go spend $25 so you're not cross platform ignorant. Since I use it on WIFI only of course it's going to be in airplane mode with always-on WIFI since cellular radio isn't needed. My iPad is WIFI only for comparison. Why would I pay for a bunch of plans when I can do everything including VOIP on WIFI or WIFI tether to my primary phone.

Cool.

At least you've admitted that battery tests and usage scenarios for you are basically irrelevant for 99% of other cell phone users.

 

[Deleted whining member 223597]

Go spend $25 so you're not cross platform ignorant. Since I use it on WIFI only of course it's going to be in airplane mode with always-on WIFI since cellular radio isn't needed. My iPad is WIFI only for comparison. Why would I pay for a bunch of plans when I can do everything including VOIP on WIFI or WIFI tether to my primary phone.

That doesn't even have anything to do with either platform...I have an old iPhone 4S that I just run on WiFi, should I go compare it's battery life to an S6 that is dealing with a bunch of notifications and searching for signal?

Sorry, I don't have to carry around 2 phones, my iPhone 6 on Verizon allows me to do everything with 1.

 

[mi7chy]

At least you've admitted that battery tests and usage scenarios for you are basically irrelevant for 99% of other cell phone users.

What are you mumbling on and on about cellular like an apologist? Both devices are running on WIFI. Even on LTE the battery life is equivalent to 2.4GHz WIFI. If anything, 5GHz WIFI probably consumes more power than LTE.

 

[Deleted whining member 223597]

What are you mumbling on and on about cellular like an apologist? Both devices are running on WIFI. Even on LTE the battery life is equivalent to 2.4GHz WIFI. If anything, 5GHz WIFI probably consumes more power than LTE.

I guess this is rocket science....LTE = generally searching for signal much more often than WiFi.

 

[Mchart]

What are you mumbling on and on about cellular like an apologist? Both devices are running on WIFI. Even on LTE the battery life is equivalent to 2.4GHz WIFI. If anything, 5GHz WIFI probably consumes more power than LTE.

http://www.cs.columbia.edu/~lierranli/coms6998-7Spring2014/papers/rrclte_mobisys2012.pdf

WIFI uses more power while idle, but the cell radio uses substantially more when actually doing work. Considering you were testing battery usage of an IRC client that constantly runs in the background constantly TX/RX packets this is kind of a big deal.

So yes, my comment is completely relevant. Most people are using their cell phone as a cell phone and don't have constant WIFI access.

Either do your test using the cell radio or go sit in the corner.

 

[mi7chy]

I guess this is rocket science....LTE = generally searching for signal much more often than WiFi.

Stick to facts and not BS. LTE is running a persistent Groove IP VOIP app in the background so what Mchart is saying about persistent TCP connection over cellular killing battery is also BS.

LTE

WIFI

 

[Mchart]

Stick to facts and not BS. LTE is running a persistent Groove IP VOIP app in the background so what Mchart is saying about persistent TCP connection over cellular killing battery is also BS.

"Despite several new power saving
improvements, we find that LTE is as much as 23 times less power
efficient compared with WiFi, and even less power efficient than
3G, based on the user traces and the long high power tail is found
to be a key contributor."

That is fact. No amount of pseudo-science BS you post will change that.

Relevancy -

This is why apps that don't utilize push are stupid for a mobile environment for most users. You know, people that actually use their cell phone as a cell phone. This is why Apple smartly dick slaps developers who don't follow this model.

 

[Mchart]

Also; connection state / "persistence" doesn't mean a packet is being sent every second. So again, your VOIP client cannot be compared to your IRC client that is receiving new messages in the chat room every time someone sends one.

 

[mi7chy]

This is why apps that don't utilize push are stupid for a mobile environment for most users.

So, your solution is to use push for ssh, telnet, ftps, sftp, etc. even on a VPN?

 

[Mchart]

Wrong layer try again.