We’ve been getting rained on a lot this week, in more ways than one. Carol’s garden is going gangbusters, and I’ve never seen an explosion of wildflowers along my accustomed hiking paths as I’m seeing right now. There’s a bee shortage somewhere in the country, I’ve heard, but the little buggers are thronging the wildflowers here. Temps are deliciously cool, for June, which seems to be a trend this year.
On the flipside, some dorks broke into my hosting directory a few days ago and inserted porn spam links into all my static HTML. They tried to modify the PHP in my instances of the Gallery photo manager for purposes unclear, but Gallery stopped working and I had to delete both instances. (I have backups of all the photos and captions and will reinstall as time permits.)
That whole adventure happened while I was on deadline reading copyedits on the first five chapters, and it did not endear me to cloud computing. I’ve had some time to think about the whole sorry mess, and some larger questions arise:
- How do we keep crap like that from happening? (This is a mostly rhetorical question; I’m not sure that we can.)
- Apart from portability (i.e., accessing your data on the road) what’s the real value-add in cloud computing? Remember to figure in the cost-benefit of having to find and sometimes pay dearly for a broadband connection to use it.
- And if portability is the only value-add, why screw with something as inherently pricey and dicey as the Cloud?
Why not put the Cloud in your pocket?
I just ordered my very first 32GB thumb drive. I skipped the 16 GB size entirely, because my trusty and much-missed 2001 Thinkpad X21 had a 32 GB hard drive, and I never filled it up. It contained all my major apps (Word, Excel, PowerPoint, MapPoint, InDesign) and all my Internet apps, plus a scattering of smaller utilities. It also contained a great deal of data, including everything I had written electronically since 1979, though in truth much of the bulk lay in MP3s. Document files are remarkably small.
One of the most significant trends of the last two years is the explosion in “portable” apps, meaning software that does not require a formal installation process beyond unzipping it into a directory. Nothing goes into the Registry, nor into \windows\system32. The whole app lives in a single folder. What a brilliant idea! (Wait…all software used to be like that…)
There have always been portable apps, but for the last fifteen years or so it’s been seen as declasse to produce them. Why? Think for a second: Once installed, a conventional Windows app can’t simply be lifted out of its folder and copied to another machine. It was one of the earliest forms of stealth DRM, invented, I suspect, specifically to keep MS Office from wandering.
No more. There are now lots of portable packages, many or most of them completely free. See PortableApps and 100 Portable Apps for Your USB Stick. You can get OpenOffice, the Gimp, Thunderbird, Firefox, Kompozer, and just about anything else you might need in portable versions. You can unzip them into directories on a thumb drive, and execute them from File Manager. (There are also portable app managers like CodySafe that give you a separate UI for your portables and stick data.) Portables run like conventional Windows apps, except that they don’t crap in your machine.
I got into portable apps while thinking about degunking for Windows. The Registry is Gunk Central, and much havoc is caused by duelling and mis-versioned DLLs dropped like softball-sized hailstones into the system32 directory. When I got my new Core 2 Quad last summer, I resolved to install only what conventional apps I absolutely needed, and use portable apps for everything that I could. The results? I have a cleaner-running machine that boots fast and has a remarkable lack of line items in Task Manager’s Processes tab. I’ve tried to stick with FOSS apps, because commercial apps are always down there in your taskbar popping up nag balloons, trying to upsell you or force updates down your throat.
It’s worked very well. What I want to try is having a single largish thumb drive containing not only data but also the programs used to manage it. Other people have been doing this for years, and it’s time I gave it a try. In the meantime, my view of the Cloud cooks down to this:
Take from the Cloud what can only be had from the Cloud–and keep the rest in your pocket.
Jeff, Just a quick note to offer commiserations on your hosting being hacked (though I’m not sure what it has to do with cloud “computing” per se), and to say my main (paid) use of the cloud is offsite backups. Ever since losing a Money database due to a disk failure (and having to painfully reconstruct said database from an very old CDROM backup), I backup changed files to Amazon S3 every 12 hours. Peace of mind, if nothing else.
I read from a Google report that there’s a 2% chance of a failure per year for modern drives (this increases dramatically if a drive is over 3 years old). For the 8 drives I have attached to my PCs in regular use, that means I have a one in 6 chance of one of my drives failing each year. A dice roll. I don’t know the figures for flash-devices, or for CDROMs for that matter. Anyway, enough to make me want to get stuff offsite.
http://labs.google.com/papers/disk_failures.pdf
(I seem to remember scare stories about some CDROMs becoming unreadable after only a few years. I just checked with my oldest CD backup from 2001. Still readable.)
Cheers, Julian
It’s about cloud computing in that anything reachable through an IP address 24/7 is going to get hacked sooner or later, unless we come up with some sort of miracle.
In the 24 years since I bought my first hard drive, I’ve had exactly three fixed drives fail on me, all conventional magnetic disks. (Floppies are another story, as are the late and unlamented EZ-Drive cartridges, which died like flies.) As bad a rep as they had, I don’t recall any Zip-100 cartridges dying, nor any of the ancient Iomega “cafeteria tray” cartridges, of which I had many in their day.
And so far, flash storage has a perfect record. An EE friend of mine says that they’re probably the most durable digital storage medium we’ve come up with yet, and they certainly get points for compactness.
Apart from the chances of the house burning down, is there any advantage to using S3 over just having a USB hard drive behind the box, doing a copy of important stuff every 12 hours? (I do a monthly offsite backup to our safe deposit box.) Isolation from the Net probably eliminates most chances of getting hacked directly, and if somebody gets root on your machine, you probably have worse problems than just losing a backup.