Slashdot Mirror
Amazon S3 Adds Option To Make Data Accessors Pay
CWmike writes "Amazon.com has rolled out a new option for its Simple Storage Service (S3) that lets data owners shift the cost of accessing their information to users. Until now, individuals or businesses with information stored on S3 had to pay data-transfer costs to Amazon when others made use of the information. Amazon said the new Requester Pays option relieves data providers of that burden, leaving them to pay only the basic storage fees for the cloud computing service. The bigger question with the cloud is, who really pays? Mark Everett Hall argues that IT workers do."
With advertising revenues dropping we could see this as a new trend for accessing content. Of course, many sites are popular because they are free so this would likely reduce traffic. I could see how this would be useful for a site like Fark, though, who already has a paying crowd.
Of course, the big users I can see are porn sites.
...I'm not sure how accurate that is. In my experience S3 and EC2 enable small companies to do things they might not otherwise hassle with.
The article also says "The glory days of the UNIX system administrator and the Java programmer are dead and buried". Really? From what I've seen, good Unix sysadmins are in high demand - whether the servers are in your colo rack or in a RackSpace facility, you still need someone to mind the farm and twiddle the Puppet manifests. Not sure about Java programmers, but demand for Ruby (especially Rails) programmers is quite high.
The Army reading list
The problem of making web businesses profitable has been with us for a long time. Micropayments, internet dollars, memberships, the list of attempts is long - with some successes and a heck of a lot of failures. The number of sites saying "free for the first 3 months" is ridiculous. Then they try to charge and all their members go away. Nasty. Bad for business.
S3 is - basically - a tax on bytes. Maybe that's a way to go. But it would end up encouraging sites that move large amounts of data, instead of being useful and efficient. Not so good.
It's for sure we need some sort of reward mechanism to allow innovation to survive. At the moment all we have is advertising. This not enough - Google not withstanding. Heck, I turn them off .. so where is the revenue?
Any ideas?
"Cats like plain crisps"
Depending on the popularity of a piece of data, I could easily see the cost of transfer being much higher over time than the cost of storage. After all, once you have stored your 1GB file on their servers, with a fixed asset cost for that 1GB likely 1$, and then it gets accessed by a couple people a day for a year or so. I'd be willing to bet that the cost of moving the data around would be signifigantly larger than the cost of keeping it stored.
Mark Everett Hall really looks at things from a biased IT perspective. Yes, in his analysis, this will hurt US as IT professionals, but we do not constitute the entire economy. In fact, when businesses get leaner the economy gets better on a macro scale. Yeah, we could be hurting, but that doesn't mean the economy will be.
And that, of course, is assuming you buy the thesis that people will move into this and fire their IT staff all in 2009 - here's a clue: they won't.
And IF they did, what do you think us highly skilled laborers would be working on? My guess is cloud computing - those things don't code, administer or test themselves. Truly, looking on the bright side, one can visualize all IT effort being concerted into one specific area, hastening the arrival of new innovations - but I'm of the opinion that that analysis goes too far the other way.
My bet, the pendulum will continue on its slow course back to dumber clients and smarter servers, but it's not going to change anything in major way overnight, or over 2009.
If you can read this... 01110101 01110010 00100000 01100001 00100000 01100111 01100101 01100101 01101011
It is not accurate. I think that laws of conservation of energy apply here.. If the companies won't develop and/or deploy in-house some software and use SaaS, the resources they don't need were used by others to provide the service.
"Intel, Hewlett-Packard, IBM and their supply-chain partners in China need fewer engineers and assembly-line workers to design and build machines to run the packaged or custom apps"
WTF? Does Amazon et al run their services on abacuses?
And if companies want to get all they can from the service, they need IT people coz the sales&mgmt ppl won't do anything.
Nobody wants to look at the data I've stored. Not even myself.
Have gnu, will travel.
Rather, when IT people are writing about coal miners losing their jobs, it's progress; while when IT people are writing about IT people losing their jobs, it's pernicious.
S3 (a developer platform) coupled with a consumer front end (such as JungleDisk which works on Win/OS X/Lin) makes a great storage/backup solution. With JungleDisk's ability to keep old versions and deleted versions for a specified interval, you don't even have to worry so much that your only backup solution is mirroring (as you'll be able to recover point in time with a little effort). It's great for personal use in this sense.
JungleDisk can even encrypt your files before uploading them, and transparently decrypt them when you retrieve them (3des I believe, with a user-provided private key), so there's no worry about their contents being snooped on by Amazon or overzealous LEO's.
I use this as one of my tiers for backing up my photography and music library (about 43 gig of photography and 30 gig of music). Monthly billing from Amazon is per gb, and it's something really reasonable like $0.16/gb/mo. My billing comes out to under $17 (I've got some other stuff in there too, but the photos and the music would be the hardest to replace).
Slay a dragon... over lunch!
As we consume more IT resources the number of workers per resource unit has to fall - or we're going to wind up spending our entire budget on IT. The question for IT workers is whether the amount of IT workers has peaked or not. I don't think it has yet.
The first computer I worked with was a PDP 11/70. Less than 1 MIP and we had a dedicated operator. By that measure my laptop needs several thousand support personnel.
However, we spent close to $500,000 (in 1981 dollars) for that system. It supported 32 terminals. Today, I could put together 32 desktops plus a server system for less than $100,000 but would probably still want to have a dedicated IT support person for that many desktops (given that it's a small company and that's all of our IT infrastructure - larger companies get by with fewer desktop support people due to economies of scale).
My wife worked at Oracle here in Japan for a while. The director of the Oracle certificate program once set a long term goal of, I think, 5 million certified Oracle DBA's in Japan. Now, Japan has a total population of about 128 million so he was setting a goal of 4 out of every 100 people to be Oracle DBA's. Absolutely ludicrous.
Personnel are now the largest cost in IT. Anything that reduces IT costs will be reducing personnel costs. The real question is whether the IT budget overall is shrinking or growing.
The interesting long term question is whether IT will mature like power or plumbing to the point where an average company does not keep IT specialists on staff but just calls them in as needed. I would argue that it is different since IT done properly is a strategic asset customized to your company somehow but time will tell.
Right. Who do you think is developing the programs that run on EC2? And who do you think is needed to manage the instances (VMs), or write software to manage them?
Adapt or die.
Don't thank God, thank a doctor!
Here is a post from the futute...
Local IT is old technology and something new has come along to potentially replace it in its industry. There's no new technology to replace today's SaaS, therefor these workers who are being laid off are at the forefront and they laying off cannot be called progress.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
No, the IT guys who lose their jobs will most likely be those supporting commodity solutions that the average secretary is too lazy to figure out for herself. There is still going to be a huge need for local support and IT staff no matter what happens, because some things are just too damn important to trust outside of the organization, or too expensive to not do yourself.
Few, if any, companies are going to use an outside provider to hold their critical and/or proprietary data. There is no way in Hell, for example, that my Subversion repository will be stored anywhere but on a machine in my company, under my direct control, physically and logically - regardless of SLAs, encryption, or whatever else.
And I am most certainly not putting our accounting database anywhere that could possibly require a "rent payment" or external connection - if I lost access for 1 minute, we're out of business entirely. Or, what if it leaked out and competitors had access?
There is no possible way to make a cloud of untrusted machines substitute for locally owned, managed, and controlled systems.
For the love of God, managers don't even trust the geek in the cubes, what makes anyone think that they will trust a "cloud" with anything important?
Some managers have proprietary information that gives them an edge over the competition. Toss that out there? Some may even be Bernie Madoff, and those fucks aren't giving up any information.
A whole lot of IT will stay local just for security and paranoia reasons alone.
In short, clouds may expand what we do nowadays but they won't supplant anything.
Oh, and who is going to troubleshoot the loss of the Internet connection?
Convenience of access from anywhere (only needing to install the client if I really needed access to a file from a remote location), plus the convenience of effort-free off-site backup.
$16/mo to safely and effortlessly back up 8 years of photography seems like an incredibly good deal to me. I pay way the heck more than this for my homeowner's insurance, and this is just another form of personal property insurance. The only monthly bill that regularly falls under this dollar amount, in fact, is my credit cards, mostly because they're an emergency tool only for me. Everything else costs more. Land line, cell phone, TV service, electric, gas, security system, etc - all much more than this, and I'd be way less devastated if any of these services were lost.
GP said he was shopping for storage, and I recommended a reliable, fast, easy, convenient, and inexpensive option =). A common lament from people who have lost their house and all their possessions in a fire is that they cannot get back their photographs. Everything else they can eventually (assuming no deaths), but the photos are gone forever. And the point is you pay per usage. $0.15 (rates have gone down since I signed up for this apparently) per gb per month is a good rate IMO.
I have a local NAS box w/ redundant disks too. That NAS also backs itself up to external USB disks. I work locally, and back up to the NAS automatically at night. It backs itself up automatically the following night (it runs a few hours before my desktop->NAS backup) giving me 2 days local recovery time, plus 7 more days (how I have it configured in JungleDisk) online should I accidentally delete or corrupt something. Plus in the event of a catastrophic failure (house fire, massive power surge, etc), I have the online backup I can fall back on.
FWIW, I've not found inexpensive external drives to be even remotely reliable as a backup. $100 1TB drives (or similarly discounted drives) have a ridiculously high failure rate. Good if you're transferring some files to a buddy, but a really bad idea as a backup solution. Those things tend to fail as a brick. The whole device just quits some day leaving your data trapped, maybe there, maybe not. Maybe reported writing successfully when it didn't. Maybe it wrote successfully, just can't be read again. That's the sort of drive my NAS backs itself up to, and by this point I've replaced enough of those that I really should have bought better ones out the gate, it would have been cheaper.
Slay a dragon... over lunch!
I love that word pernicious... I've started naming all my servers after nasty p-words. pernicious, persnickety, pugnacious...