Skip to content

iCloud

The annual Apple Developers Conference was held in San Francisco 6-10 Jun 11. Steve Jobs announced iCloud, a new free service. The rationale is that most people find it a hassle to synch mobile devices with their main desktop. Instead they will be able – at greater cost, risk and convenience – to synch with the iCloud.

This is not exactly a new idea. I have lost count of the number of internet start-ups that try to lure you into keeping all your files on their servers, in return for putting up with delays and ads, both removable for a modest subscription.

It clearly has some convenience, especially for those who travel extensively and unpredictably. Rush off without any gadgets to Rio and lo, you just log onto iCloud and start working on your files whilst playing your stored music.

I have to admit that I have been fairly impressed with Time Machine. I keep a 1/2 terabyte (1TB = 1000 gigabytes) disk plugged into a firewire port and Time Machine did incremental backups every hour. The retrieval system is unusually neat and allows me to scan quickly backwards and select whichever old version of a file I want. The only snag was that I began to notice an irritating slow-down in app speed, especially if I was playing music on iTunes. I would often be waiting for simple operations – like a new window opening, or the contents of a new email in a list – for ten seconds or more.

Of course, the backup itself slows to a crawl, so the wretched thing is taking an hour instead of a minute or two.

Maybe the difficulty is that the 1/2 TB disk is now 90% full. Maybe I need to switch it for a 2TB disk. But the wretched thing is supposed to dump old backups as necessary to free up space (it is backing up a 1/4 TB drive). It is really just there to increase my peace of mind and now it is becoming a constant aggravation.

I tried to fix it with Time Machine Editor, a free app which restricts backups to set times instead of the standard once an hour for the basic Time Machine. I opted for just a handful a day. But that did not work either, because the wretched app would not skip them if the machine was sleeping at the time. So if I woke it up at noon, it would immediately launch into the 10am backup, even though another was scheduled at 1pm.

Matters were made worse by serious performance degradation in Safari v5.0.5. So bad, that after looking at various forums and seeing that others were hitting the same problems, I dumped it in favour of Chrome.

All this is manageable. But all the comms involved are local. Data is being moved by bus or firewire. Now Jobs wants me to shift all my data to the other end of a broadband link. Last year some clown put a mechanical digger into a cable at the new Olympics site 5 miles north. It turned out that it was the broadband trunk for the local exchanges. I was impressed with BT. They got a temporary microwave link up and running in less than a week. But my broadband was totally down for days.

Also, I am less impressed with BT’s tariff structure. “All you can eat” broadband has not yet arrived. So there is a significant cost to each 1GB of download and I have many tens of GB of sound, video and image files.

Even when broadband is working, it can be surprisingly slow. I am a mile south of Canary Wharf, one of the world’s major financial centres, so you might think I would get broadband speeds of maybe 100MB/s. No. The maximum is about 0.3MB/s. They are still thinking about fibre optic. Meanwhile at the southernmost point of the loop, I am apparently several miles from the exchange as the copper wire wends and weaves.

But interestingly many sites are far slower even than that. Where the bottlenecks are is hard to tell, but presumably their speed would not be affected if my broadband speed went up 10x. Also page loading generally has got much slower because of ads. It is common for webpages to be loading a hundred separate items, many from remote ad-serving sites.

If I had more time I would figure out how to prevent the loading of all the bloat that is adding nothing to the information I actually want from the web, but life is too short.

Finally, upload speeds are typically far slower than download speeds. If I buy a new track from iTunes, then presumably it is wafted at high speed to my iCloud space. No upload problem. But what about the 100 GB of stuff I already have on my desktop?

Also, if I download a 20MB book from books.google.com, does that have to go via my desktop? Presumably yes for the moment. We are a long way from the situation where you just direct books.google.com (or any other site) to direct the download to iCloud.

The third problem (after speed and cost) is reliability. The omens are not good. Jack Schofield wrote an excellent piece about backup recently which recounted some of the cloud disasters:

1. Streamload/MediaMax/The LinkUp

LinkUp shut down on 8 Aug 08. CEO Steve Iverson said

… at least 55% of customers’ data was safe. How much of the remaining 45% was saved is not clear …

2. Amazon’s EC2 cloud services crash

Amazon Web Services crashed on 21 Apr 11, taking down thousands of websites. It took several days to restore services. Here is Amazon’s email explaining the situation to affected customers.

Hello,

A few days ago we sent you an email letting you know that we were working on recovering an inconsistent data snapshot of one or more of your Amazon EBS volumes. We are very sorry, but ultimately our efforts to manually recover your volume were unsuccessful. The hardware failed in such a way that we could not forensically restore the data.

What we were able to recover has been made available via a snapshot, although the data is in such a state that it may have little to no utility…

If you have no need for this snapshot, please delete it to avoid incurring storage charges.

We apologize for this volume loss and any impact to your business.

So much for all the guff about proper backup so this kind of thing could never happen.

3. A Flickr user had his Pro (paying) account deleted in Feb this year. He eventually got his 3,400 photos back and free membership for the next 25 years.

On the other hand, when I was locked out of my free Flickr account on its takeover by Yahoo, I did not manage to get back in. Since I had them on a hard drive, I gave up after wasting several quarter-hours getting nowhere.

4. Similarly there are many accounts of people being locked out of Picasa (try googling).

Schofield recommends multiple backups for anything important, including CDs/DVDs and memory sticks. The latter are not to be sneezed at. I remember watching an amusing television programme in which some journalists set out to destroy memory devices. The only one to come through with flying colours (after being run over etc) was the memory stick.

Also, of course, those who rely on things like the Time Machine should remember that same-site back-up has its limitations. Fire, flood, lightning etc can wipe out all devices on the same site simultaneously.

But my main thought is to wonder if OS X is slowly turning into a mess. It is almost impossible to crash unix. But it is increasingly easy to get the stuff that Apple puts on top of it to lock up. In practice that is just as bad as a complete crash. The only remedy is the same: to reboot. Just as bad, is the increasing tendency of the wretched thing to slow to a crawl. Ok, ok so this iMac is 5 years old, but you would think a 2GHz processor, 2.5GM RAM, 1/4 TB drive would be enough to run a little word processing on whilst browsing the web on a 1920 x 1200 screen and listening to opera.

{ 4 } Comments

  1. Tom Welsh | 14 June 2011 at 8:16 am | Permalink

    Simple arithmetic suggests that huge clouds like those offered by Amazon and Apple face insuperable challenges in managing downtime. Ordinary servers or server farms, meeting the needs of ordinary organisations with fewer than, say, 10,000 people typically aim for reasonably demanding up-time goals such as “not more than half a day of unscheduled downtime per year”. Even half a day can be critical if it happens at the wrong time (which Murphy’s Law assures us it will). Then there is the scheduled downtime: some users will not be aware of it, there may be a sudden unexpected need for data right in the middle of the scheduled maintenance period, etc.

    Now consider the situation of, say, Amazon – which is offering server time and space to the entire world. To accommodate its immense number of clients, it needs a correspondingly huge number of processors, hard drives, memory modules, etc. In an ideal world, the system could be designed in such a clever way that everything “fails safe” so that, should a processor or a hard drive go bad, everything is smoothly shifted to other identical elements. But from time to time there are likely to be unforeseen “system” problems, often software-related, which affect all clients (or at least a large number of them). That is when the ambitious nature of the project bites viciously, as the damage is multiplied by by the number of clients who are affected.

    More simply, it’s unwise to put all one’s eggs in one basket.

  2. John Scholes | 14 June 2011 at 8:58 am | Permalink

    I am not really clear why Jobs wants to tangle with the Cloud. Synching is not that big an issue for most individuals, so is there some hidden agenda? What new gadgets would it make possible?

  3. Tom Welsh | 15 June 2011 at 11:07 am | Permalink

    Remember that Jobs is, au fond, a superb salesman. So he instinctively looks at the world and sees opportunities to make immense piles of money. The size of a given pile of money depends largely on the size of the customer base, the revenue per year, and the ability to maintain the revenue stream more or less indefinitely.

    Microsoft more or less started the ball rolling with Windows and Office. The customer base ramped up to hundreds of millions, each unit sale was not unadjacent to $100, most of the sales were invisibly transacted by hardware vendors with no cost to Microsoft, and new versions could easily be sold every few years. Mobile phones took this model further forward: the customer base grew beyond 1 billion, and the rate of churn was very much faster than for Windows and Office – essentially because most customers were individual consumers, who were intensely susceptible to fashion.

    The cloud holds out some potential for even further growth. Imagine if everyone in the world had one or more portable gadgets (smartphones or whatever) that communicated by wireless; and all the server-side functions were owned by one operator! Without the cloud (“server”), none of the gadgets would be worth their weight in rocks. And the potential sales revenue becomes simply unlimited.

  4. John Scholes | 23 June 2011 at 11:19 am | Permalink

    Another tweeted link from Jack Schofield http://www.smh.com.au/technology/security/4800-aussie-sites-evaporate-after-hack-20110621-1gd1h.html .

    More basic this time. An webhosting site simply failed to have adequate security or even adequate backup. It got hit by hackers leaving many of its customers with totally destroyed websites (unless they had the sense to backup elsewhere).

Post a Comment

Your email is never published nor shared. Required fields are marked *