or Connect
AppleInsider › Forums › Mac Hardware › Current Mac Hardware › Software-based iPad Wi-Fi problems detailed by Princeton
New Posts  All Forums:Forum Nav:

Software-based iPad Wi-Fi problems detailed by Princeton

post #1 of 51
Thread Starter 
A software glitch in Apple's iPad causes the device to continue using a network-assigned IP address after its lease has expired, according to data collected by Princeton University.

The Internet Protocol (IP) address of a device is a unique identifier that allows communication between devices. An IP address is assigned through a method known as Dynamic Host Configuration Protocol (DHCP), and these addresses can dynamically change over time on a private network as devices disconnect and reconnect.

A DHCP client, such as an iPad, will "lease" a unique IP address to a device on a network for a set period of time. Once that time has elapsed, the DHCP client asks the server to renew that lease. Apple's iPad, however, does not go through that renewal, according to information from the Office of Information Technology at Princeton University.

Instead, the office has found that the iPad will incorrectly continue using an IP address without renewing the lease, usually for hours. The issue is resolved when the iPad asks for a new DHCP lease, or the iPad disconnects from the network.

The information suggests it is a software issue within iPhone OS 3.2, and is likely something that Apple could repair through an update to its mobile operating system.

The problem was found to be common on Princeton's campus, with more than half -- 25 of 41 -- of all iPads demonstrating the malfunction as of April 18. If that ratio bears out on a larger scale, there are likely a huge number of iPad owners experiencing similar issues, as Apple sold more than 500,000 devices in its first week.

The report noted that some devices have demonstrated the malfunction several times, and eight devices were outright blocked from the university's network.

"When a DHCP client malfunctions this way repeatedly, Princeton blocks the device from using those campus network services which rely on the device's DHCP client respecting lease times," the IT office said. "These include our wireless services. We do this to protect other customers of those services from the disruptions caused by the malfunctioning devices."

The school has published a temporary workaround for the issue, but is working with Apple and providing them with technical data in hopes of solving the problem. That data was sent to Apple on April 7.

The IT office also noted that it has been incorrectly attributed with diagnosing the cause of Wi-Fi signal issues or connectivity issues that have been widely reported by iPad users. The university noted that information is inaccurate, as its documentation relates only to a DHCP client issue, not Wi-Fi signal or connectivity issues.

Princeton has been able to replicate the problem regularly by allowing the iPad to lock its screen before the DHCP lease renewal time, and allowing it to remain locked, with the screen off, until the DHCP lease has expired.

Princeton's workaround has users reconfigure the iPad's settings so that the screen never locks. This can be done by going to Settings, General, Auto-Lock and choosing "Never." Users must also turn off Wi-Fi before they manually lock the screen, turn the iPad completely off (rather than just locking it), or simply leave the iPad on without locking the screen.

"We recognize that this workaround represents some inconvenience to the iPad customer," the documentation reads. "We view the workaround as a temporary workaround to allow these customers the opportunity to use their devices on the campus network until there is a fix from Apple."

Soon after the iPad was introduced earlier this month, users began to report connectivity issues with the device. One of the most common problems experienced relates to rejoining a Wi-Fi network after the iPad has been restored from sleep. Apple responded by issuing a support document.

Wi-Fi is currently the only method by which users can access the Internet on their iPad. Apple revealed on Tuesday that its 3G capable model, which offers high-speed wireless data over cellular networks, will be released in the U.S. on April 30.
post #2 of 51
I was at Princeton for an internship last summer, and I have to say that their network is ridiculous. My MacBook Pro got banned for a couple days because of some detail of how VMWare Fusion shared the connection. Had to hack the os to fix it. Bottom line, OIT at princeton is a bunch of (oh shit, here comes godwin's law!) Nazis, and have the network locked down to a ridiculous degree. This nonsense doesn't reflect poorly on Apple, but on Princeton for having a bizarrely non-standard wifi network.

C
A Conclusion is the place where you get tired of thinking. - Lesicus Stupidicus
Reply
A Conclusion is the place where you get tired of thinking. - Lesicus Stupidicus
Reply
post #3 of 51
Rather nasty bug. Guess I'll just be assigning a static address at home. Anywhere else, 3G. That should suffice until the bug fix.
post #4 of 51
Quote:
Originally Posted by Celemourn View Post

I was at Princeton for an internship last summer, and I have to say that their network is ridiculous. My MacBook Pro got banned for a couple days because of some detail of how VMWare Fusion shared the connection. Had to hack the os to fix it. Bottom line, OIT at princeton is a bunch of (oh shit, here comes godwin's law!) Nazis, and have the network locked down to a ridiculous degree. This nonsense doesn't reflect poorly on Apple, but on Princeton for having a bizarrely non-standard wifi network.

C

You have to wonder why there aren't tons of universities reporting similar problems.
post #5 of 51
I thought the issue was that it holds an IP even while in sleep? It seems more like it is by design then anything. If you haven't left the network, why give up the IP? Maybe they should expand their subnet. This would make sense as to why it isn't a problem at other universities, they thought ahead and have a higher class subnet or supernet.
post #6 of 51
Quote:
Originally Posted by AppleInsider View Post

A software glitch in Apple's iPad causes the device to continue using a network-assigned IP address after its lease has expired, according to data collected by Princeton University....

While it's definitely a bug in Apple's software that's at the root of it, if you read the report carefully, you can see that this is a "problem" of Princeton's own making.

20 devices (campus wide over a multi-week period), *sometimes* don't release their IP??? Seriously? And the only effect if they did absolutely *nothing* about it, would be that once in a while a user would get an "IP in use" dialogue box before they were instantly handed a valid IP?

Wow, that's totally a good reason to ban the devices campus wide.
post #7 of 51
Requires simple software fix.
post #8 of 51
Sure, it's not Apple's fault There are a number of universities that have already banned it. AI should just probably google for more articles, they are all over the web.
The great things for the great, the abysses for the profound, the thrills for the refined, and, to sum up shortly, everything rare for the rare.
Reply
The great things for the great, the abysses for the profound, the thrills for the refined, and, to sum up shortly, everything rare for the rare.
Reply
post #9 of 51
This is only one of the problems users face with wifi on the iPad it seems. I have a feeling it is not the majority of problems too. If you read all the forums(especially the official apple forums) you will see there are multiple symptoms. My disconnects happen even with reserved IPs on the DHCP server so there is zero chance that 2 devices will get the same IP address. A friend has the same problem but he gets disconnected even on wakeups. I only get disconnected every 30 minutes or hour. Signal strength is full when it happens too. If I immediately go into settings and enable wifi again it works. The iPad doesn't work at all with a second router I have(older Buffalo nfiniti wireless-N). My HP laptop, macbook air, PSP, DSi, etc all work fine with that same router. Even if I restrict it to wireless N only the macbook air works with it but the iPad doesn't.
post #10 of 51
According to the princeton site, they lock down, log and monitor their network more thoroughly than one would expect the average IT administrator.
post #11 of 51
This is really a user error not a bug in the product.
post #12 of 51
Quote:
The IT office also noted that it has been incorrectly attributed with diagnosing the cause of Wi-Fi signal issues or connectivity issues that have been widely reported by iPad users. The university noted that information is inaccurate, as its documentation relates only to a DHCP client issue, not Wi-Fi signal or connectivity issues.

And yet the title of your article is "Software-based iPad Wi-Fi problems detailed by Princeton"
post #13 of 51
Quote:
Originally Posted by Cam'ron View Post

I thought the issue was that it holds an IP even while in sleep? It seems more like it is by design then anything. If you haven't left the network, why give up the IP? Maybe they should expand their subnet. This would make sense as to why it isn't a problem at other universities, they thought ahead and have a higher class subnet or supernet.


except when it wakes up and still has the same IP it will cause a conflict with another device that was assigned that IP
post #14 of 51
Amazing how quick AI readers are to suggest Princeton are in the wrong here. This is clearly a bug in the iPad software, and I would guess Apple will be happy with the methodical way Princeton have isolated the bug so they will be able to re-create it and fix it easily.

The company I work for locks down it's network to this sort of level also. Networks within companies are pretty much mission critical, and anything that misbehaves is treated as a potential threat.

People here need to remember that just because we like Apple products does not mean Apple are infallable.
post #15 of 51
Quote:
Originally Posted by nasdarq View Post

Sure, it's not Apple's fault There are a number of universities that have already banned it. AI should just probably google for more articles, they are all over the web.

I saw your comment before you changed it !

I'm not saying it's not Apple's fault. But these things go both ways: This is only getting attention because it's Apple. How many other wifi enabled devices cause similar headaches for organizations and we never hear about it?

I used to have a Netgear router that had problems with all my Apple gear: It rarely handed out the reserved address. Without wasting time digging into it I couldn't tell who was to blame. The situation at Princeton involves more layers of equipment and software that could be to blame, but maybe they've done enough research to say without question that it's Apple's fault. But maybe they haven't.
post #16 of 51
Quote:
Originally Posted by nasdarq View Post

Sure, it's not Apple's fault There are a number of universities that have already banned it. AI should just probably google for more articles, they are all over the web.

I think I have found 5 schools that are anti-ipad. and actually none of them have banned it. just they won't be giving them away to students. And 4 of them admitted that their campus networks can't handle the potential added load which is why they won't give them away or carry them in the campus store at this time.

Quote:
Originally Posted by al_bundy View Post

except when it wakes up and still has the same IP it will cause a conflict with another device that was assigned that IP

wouldn't keeping track of assigned IPs and not duplicating assignments be a part of the network not the device software.

A non tech's thoughts on Apple stuff 

(She's family so I'm a little biased)

Reply

A non tech's thoughts on Apple stuff 

(She's family so I'm a little biased)

Reply
post #17 of 51
Quote:
Originally Posted by Prof. Peabody View Post

20 devices (campus wide over a multi-week period), *sometimes* don't release their IP??? Seriously? And the only effect if they did absolutely *nothing* about it, would be that once in a while a user would get an "IP in use" dialogue box before they were instantly handed a valid IP?

The problem comes as the devices get more popular. Not giving up a DHCP lease is a big deal, but I understand why it would be done (especially on networks with ridiculously short lease periods) to conserve battery life.

I think we are coming into the dawn of the era where WiFi just stops working effectively. A single access point can't support that many users-- 50-60 on the good equipment, 20-25 on the cheap stuff. It really requires a lot more effort than it did 3-5 years ago to make things work effectively.

If your access point is doing the DHCP (bad design), your network will come to its knees quickly. If you centralize management of the access points and have reasonable address space for the DHCP it shouldn't be a major issue though.
post #18 of 51
Quote:
Originally Posted by AppleInsider View Post

A software glitch in Apple's iPad causes the device to continue using a network-assigned IP address after its lease has expired, according to data collected by Princeton University.

I have heard that an increasing number of universities and hospitals are banning the iPad for various reasons.

I hope that Apple or someone else can fix that.

Without university support, the dream of textbooks widely issued on the iPad becomes more distant.

And medical use? That too was thought to be a promising market.
post #19 of 51
Quote:
Originally Posted by Celemourn View Post

My MacBook Pro got banned for a couple days because of some detail of how VMWare Fusion shared the connection. Had to hack the os to fix it.

You changed your MAC address, didn't you?
Dick Applebaum on whether the iPad is a personal computer: "BTW, I am posting this from my iPad pc while sitting on the throne... personal enough for you?"
Reply
Dick Applebaum on whether the iPad is a personal computer: "BTW, I am posting this from my iPad pc while sitting on the throne... personal enough for you?"
Reply
post #20 of 51
Quote:
Originally Posted by charlituna View Post

I think I have found 5 schools that are anti-ipad. and actually none of them have banned it..

Princeton.

GWU

Cornell.

None of them lightweights. My guess is that textbook publishers are taking notice of this trend.
post #21 of 51
Quote:
Originally Posted by charlituna View Post

wouldn't keeping track of assigned IPs and not duplicating assignments be a part of the network not the device software.

The device has responsibility NOT use an dynamic address without an active and valid lease. Duplicate IP addresses can be a nasty problem.
post #22 of 51
Quote:
Originally Posted by MJ Web View Post

Requires simple software fix.

Yep. Not that big of a deal. They probably use a really short lease like an hour. I set mine to 7 days on the server. The iPad is using the network even though the screen is locked to listen for IM and other background WIFI functions. But if there is an IP address shortage at Princeton that might explain a short lease.

Apple needs to fix it but for most people other than Princeton or other really crowded networks it would be a non-issue.

Life is too short to drink bad coffee.

Reply

Life is too short to drink bad coffee.

Reply
post #23 of 51
It sounds as if a programmer used

if (leaseend == current_time)

rather than

if (leaseend <= current_time)

Oops!
post #24 of 51
I haven't had this problem with my iPad, but my Mac does something similar. When it wakes from sleep it often assumes it can continue using the same IP address it had when it went to sleep. When it discovers another device is using that address it pops up an error on screen and I have to manually re-obtain an address, rather than obtaining a new address on its own. It's very irritating.
post #25 of 51
Quote:
Originally Posted by Naboozle View Post

The device has responsibility NOT use an dynamic address without an active and valid lease. Duplicate IP addresses can be a nasty problem.

I think it is definitely a problem not sure how nasty, one or both devices should receive the error "The IP address 192.168.1.XXX is being used by another client on the network". Then you have to close the network and relaunch. I wonder if Princeton is IPV6?

Life is too short to drink bad coffee.

Reply

Life is too short to drink bad coffee.

Reply
post #26 of 51
Quote:
Originally Posted by PaulMJohnson View Post

Amazing how quick AI readers are to suggest Princeton are in the wrong here. This is clearly a bug in the iPad software, and I would guess Apple will be happy with the methodical way Princeton have isolated the bug so they will be able to re-create it and fix it easily.

The company I work for locks down it's network to this sort of level also. Networks within companies are pretty much mission critical, and anything that misbehaves is treated as a potential threat.

People here need to remember that just because we like Apple products does not mean Apple are infallable.

You are clearly a paid M$ shill. Apple rulz, lololol. Expect to begin receiving hate mail for suggesting that Apple is anything less than perfect.
post #27 of 51
Quote:
Originally Posted by charlituna View Post

I think I have found 5 schools that are anti-ipad. and actually none of them have banned it. just they won't be giving them away to students. And 4 of them admitted that their campus networks can't handle the potential added load which is why they won't give them away or carry them in the campus store at this time.



wouldn't keeping track of assigned IPs and not duplicating assignments be a part of the network not the device software.


that's not how DHCP works. the client is responsible for not using an IP after the lease expires.

and in this case the iPad is keeping an address that it's not supposed to have. what happens then is that when the wifi router asks which machine has so and so IP, two devices answer. on copper it's called an ARP request or something like that
post #28 of 51
Quote:
Originally Posted by aaarrrgggh View Post

The problem comes as the devices get more popular. Not giving up a DHCP lease is a big deal, but I understand why it would be done (especially on networks with ridiculously short lease periods) to conserve battery life.

I think we are coming into the dawn of the era where WiFi just stops working effectively. A single access point can't support that many users-- 50-60 on the good equipment, 20-25 on the cheap stuff. It really requires a lot more effort than it did 3-5 years ago to make things work effectively.

If your access point is doing the DHCP (bad design), your network will come to its knees quickly. If you centralize management of the access points and have reasonable address space for the DHCP it shouldn't be a major issue though.

I'm not sure that I agree (that this is a big deal). On our network, all that would happen is the second user who was given an IP that is still in use by a sleeping iPad somewhere, would get a conflict dialogue and then immediately get a new IP. It's not like it's a static IP. Other than the transient dialogue box (on a tiny amount of computers a tiny amount of the time), there would be no real disruption to service.

Princeton admits in their statement to having a network where they monitor these things rather more closely than is typical in the industry. I'm thinking the other institutions didn't discover it because of the fact that Princeton is the only one doing the deep IP analysis and therefore the only one noticing it.

What I dislike, is Princeton making this big announcement, and screwing over their users at the same time as they are disparaging the iPad when it all just seems unnecessary.
post #29 of 51
Quote:
Originally Posted by nasdarq View Post

Sure, it's not Apple's fault There are a number of universities that have already banned it. AI should just probably google for more articles, they are all over the web.

In those articles, the bans are about the wireless networks simply being overwhelmed by the amount of data the influx of new devices has demanded.
This particular problem - the release of IP addresses - is new as far as I can find.

Maybe, if you had some technical understanding of the difference between the two problems, you'd be less snide?
post #30 of 51
What I don't understand is how the iPhone is not susceptible to these same failures, because aren't they based off the same chipset for wireless? So the code for the iPad and iPhone should be the same, right? I'm guessing the 3GS would have these problems too...? Because they're using essentially the same OS and same hardware.
post #31 of 51
Quote:
Originally Posted by Prof. Peabody View Post

I'm not sure that I agree (that this is a big deal). On our network, all that would happen is the second user who was given an IP that is still in use by a sleeping iPad somewhere, would get a conflict dialogue and then immediately get a new IP. It's not like it's a static IP. Other than the transient dialogue box (on a tiny amount of computers a tiny amount of the time), there would be no real disruption to service.

Princeton admits in their statement to having a network where they monitor these things rather more closely than is typical in the industry. I'm thinking the other institutions didn't discover it because of the fact that Princeton is the only one doing the deep IP analysis and therefore the only one noticing it.

What I dislike, is Princeton making this big announcement, and screwing over their users at the same time as they are disparaging the iPad when it all just seems unnecessary.

How do you expect them to get headlines? This is a non-story but Princeton needs a bit of the old spotlight. This will pass. Apple will fix the wifi drivers and la de da.
post #32 of 51
Quote:
Originally Posted by solipsism View Post

You changed your MAC address, didn't you?

College years used to get banned every bloody day...
MacBook 5,1, MacPro 4,1
Reply
MacBook 5,1, MacPro 4,1
Reply
post #33 of 51
Quote:
Originally Posted by bartfat View Post

What I don't understand is how the iPhone is not susceptible to these same failures, because aren't they based off the same chipset for wireless? So the code for the iPad and iPhone should be the same, right? I'm guessing the 3GS would have these problems too...? Because they're using essentially the same OS and same hardware.

Actually iPhone is 3.1.3 and iPad is 3.2. iPhone OS 3.2 probably is closer to iPhone 4.0 beta which I believe has had a few WIFI bugs reported so far.

Life is too short to drink bad coffee.

Reply

Life is too short to drink bad coffee.

Reply
post #34 of 51
Quote:
Originally Posted by WilliamG View Post

Princeton.

GWU

Cornell.

None of them lightweights. My guess is that textbook publishers are taking notice of this trend.

False. Princeton says definitively that they have not banned the iPad. GWU have blocked all iPhones and iPads from joining their wireless networks because of specific issues with their networks which they are working on correcting. Likewise you are wrong about Cornell. The iPad is not blocked from their networks at all. None of these campuses have "banned" the iPad.
Please don't be insane.
Reply
Please don't be insane.
Reply
post #35 of 51
The wifi problem with the iPad reminds me of this

http://www.dukenews.duke.edu/2007/07/cisco_apple.html
post #36 of 51
Quote:
Originally Posted by WilliamG View Post

I have heard that an increasing number of universities and hospitals are banning the iPad for various reasons.

I hope that Apple or someone else can fix that.

Without university support, the dream of textbooks widely issued on the iPad becomes more distant.

And medical use? That too was thought to be a promising market.


You and your use of the passive voice are wrong

http://www.cit.cornell.edu/news/stor...ID_1942=162201
post #37 of 51
Quote:
Originally Posted by djdj View Post

I haven't had this problem with my iPad, but my Mac does something similar. When it wakes from sleep it often assumes it can continue using the same IP address it had when it went to sleep. When it discovers another device is using that address it pops up an error on screen and I have to manually re-obtain an address, rather than obtaining a new address on its own. It's very irritating.


Okay, I know this 'issue' is NOT the first time I've heard about this... Why do I remember this exact? similar? problem happening to Mac laptops with airport cards...

Maybe I'm wrong, but this type of problem sounds _very_ familiar however it was QUITE some time back... perhaps in the very early days of Apple Airports becoming mainstream. This issues I'm reminded of might go back to 1999, 2000, 2001... Tried searching google but searching out 'ancient problems' isn't so easy.

Oh and to those who try and suggest that the iPod isn't looked kindly on at college campuses... might wanna read this.

Quote:
"Seton Hill University recently promised free iPads for every student arriving in the fall, as part of the school's technology program. The devices will be used for reading digital textbooks, communication, file sharing, note taking, and other tasks."

and other tasks... *cough* CRAZY-NET-PORN *cough*

Apologies for the childish comment but it was screaming at me, so I had to do it....
Apple Fanboy: Anyone who started liking Apple before I did!
Reply
Apple Fanboy: Anyone who started liking Apple before I did!
Reply
post #38 of 51
Quote:
Originally Posted by al_bundy View Post

except when it wakes up and still has the same IP it will cause a conflict with another device that was assigned that IP

Ah, I was under the impression that it was in an on state. I guess I misunderstood. Makes more sense now. Thank you.
post #39 of 51
Quote:
Originally Posted by walshbj View Post

You have to wonder why there aren't tons of universities reporting similar problems.


The issue is specifically noticable there because Princeton's own docs on the problem say they use a very short lease interval because Princeton wants their client IP addresses to be reachable globally and not hidden behind NAT (address translation)

Those are in short supply and that's what's driving the three hour lease cycle.

Maybe a Mod can merge my thread here. Has link to Princeton's docs
Paul

http://forums.appleinsider.com/showthread.php?t=108860
post #40 of 51
Quote:
Originally Posted by WilliamG View Post

Princeton.

GWU

Cornell.

None of them lightweights. My guess is that textbook publishers are taking notice of this trend.

Nonsense.

Textbook publishers would be stupid to base long-term decisions based on short-term hiccups like these which will disappear in weeks. I recall people complaining about the same thing when the iPhone originally came out.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Current Mac Hardware
AppleInsider › Forums › Mac Hardware › Current Mac Hardware › Software-based iPad Wi-Fi problems detailed by Princeton