At first, you might dismiss this problem since it only affects a small subset of users, but Balogh brings up an important question: "what should the developers do?" People who have purchased his application are now angry that it doesn't work, yet there's nothing he can do to help them.

A few weeks ago, Apple released their updated mobile OS, the iPhone OS 3.0, which included support for Push notifications among many other things. The two available NotifyMe applications were configured to use the new technology in both the free and paid versions. With these applications, users can receive push messages that remind them of items on their to-do list that need their attention.

Almost immediately after the company released the apps to the iTunes App Store, the support requests began rolling in. Balogh quickly realized there was a problem. Around 80% of the requests were from users who had installed NotifyMe on an unlocked phone. The users were complaining that the app either didn't work reliably or didn't work at all. Unfortunately, there was nothing Balogh or his co-developer Pavel Serbajlo could do to fix the situation.

Says Balogh, the problem involves the Push Notification service:

"...Every Push application has to request the unique token from the Apple's APNS servers to identify the device it's running on. Thanks to that token, APNS servers always know which device is yours. The token can be understood as an IP address -- the server has to know where to send the notification and for which application. APNS can also change your token regularly for higher reliability, so it's critical that the application requests the token again on every start (or when enabling the Push feature) to replace the old one if new token is forced by APNS.

On any unlocked iPhone, the application requesting the token is stuck. APNS does not provide any response at all and the application can either cancel the request completely by automatic timeout or let user wait with the progress bar forever. Either way, the user will never receive any Push message, because APNS has not provided the token."

In other words, if you're running an unlocked phone, you can forget about Push.

What Should Developers Do?

It may be easy for iPhone owners who haven't hacked their device to scoff at this issue: "Well, that's what you get for monkeying around with the firmware!" But the matter is not that simple.

Developers will have to determine how they're going to proceed now that they're aware of this limitation. Should they try to support the hacked phones? Should they just place a warning message in their app's description in the iTunes App Store? Should they ignore the problem (like Apple is doing)? Should they refund the money for the purchases?

Even worse, many of the unsatisfied customers are leaving poor, 1-star reviews when rating the application since they're unhappy it wasn't working for them. That seems incredibly unfair to the developer who has created a perfectly good application that works within the confines put forth by Apple. Yet now, new potential customers - including those content with their unmodified phones - will see these negative reviews and likely choose not to purchase, potentially overlooking great applications that would have worked just fine for them.

Apple's Involvement: Zip, Zero, Nada

Apple has every right to ignore this situation, we suppose, and that's exactly what they're doing. After all, the issue affects only a small community of hackers who have modified their phones. Or does it?

Does Apple have any responsibility to communicate this limitation to the developer community so they're not caught off-guard as Balogh was? After all, it's the developers who have to deal with the fallout - the overwhelming support requests, the unhappy customers, the bad reviews, etc.

At the very least, Apple could configure their APNS (push) servers to return an error message of some sort to let the developers know what caused the connection to fail, suggests Balogh. That way, the developers could at least plan to put a warning message in their app's description to cover themselves against these sorts of complaints.

Does that seem like a fair request? What do you think either Apple or the developer community should do regarding this issue?

According to the company, during a routine upgrade on Friday night, they ran into some problems with photo storage. The issue appears to have stemmed from several drives failing at once. Because simultaneous hardware failure such as this is rare, Facebook says they're still trying to figure out what happened.

In the meantime, though, the photos are being copied to new drives - a process that can take some time due to the large amount of data that was affected. The company asked users not to worry because they store photos in a way that maintains multiple copies of the data in case of hardware failures such as this. By early this week, everything should be back to normal.

The Cloud is Not Perfect

This recent issue with Facebook photos is just one of many cloud-based outages and issues we've seen recently - a trend that moved some to question the level of trust we should have for these online services. Over the past few weeks, we've seen the social bookmarking site Ma.gnolia completely fry and lose all their customers' data when the service's database crashed and took with it a half a terabyte of information. Unfortunately, no reliable backups were to be had.

But while Ma.gnolia may have represented the dangers of trusting a small startup with your data, Google has proved that even companies as large as themselves are not immune from problems. A recent four-hour Gmail outage from routine maintenance "gone wrong," caused a cascading failure in Google's European data centers.

And on the heels of the Gmail outage, Google Groups also went down for some time. All groups were affected for a short while.

These incidents led Google to follow in the footsteps of other cloud storage companies like SalesForce.com and Amazon in the launch of a Google Apps status dashboard so they could better communicate with customers whether their online applications were up and running.

In Google's case, repairs were made and customers were credited as necessary, but even so many businesses and individuals were affected in ways that can't necessarily be quantified so easily. Although some quickly rallied to Google's defense, reminding that Gmail's uptime is often much better than on-site hosted email systems, the point many folks are missing is that unlike in "the old days," not everyone keeps copies of their data on their computers anymore - so when the cloud shuts down, that data is just gone, albeit only temporarily in most cases.

Be Safe, Use More than One Service!

Still, even if you've switched over to cloud services for storing and accessing the majority of your data, you can prevent outages from affecting you. The trick is to store your data in more than one online service or use a hybrid cloud/desktop solution. For example, if your email is mission critical, use an IMAP-enabled desktop or mobile client. Those doing so during Gmail's outage were able to access their inbox to retrieve old emails - they just couldn't send and receive.

For photos, like the ones that recently went missing from Facebook, there are a number of online services where they could have been stored. Today, there's really no reason to only keep your photos in one spot. An easy way to upload photos to multiple sites is to use a tool like Pixelpipe which shoots photos, videos, and audio files to over 60 social networks, photo/video sites, and blogs.

Even documents can be stored in more than one location. Google Docs and Zoho may be the best known of the web office services, but you could also keep critical files saved to your computer then backed up using another third party backup service like Mozy. Or you could upload files to storage sites like Google's own Google Sites or Microsoft's "Sharepoint Lite" Office Live.

No matter what you do, there's always a chance of losing data, even if you only save files on your computer, completely eschewing the cloud altogether. But that's probably not as safe as keeping files in multiple online services. With the number of services available, an outage shouldn't mean we have to lose access to our files. Use hybrid solutions or spread your data across multiple services instead and you'll almost always be okay, outage or not.