What is Digg?
35 Comments
- chongli, on 07/22/2008, -1/+6999% Uptime equates to 3 days and 15 hours of downtime per year. 99% Uptime doesn't sound very good in that light, does it?
- teh_spazz, on 07/21/2008, -2/+556hrs? That's nothing. MobileMe hasn't been working for a week now.
- poprocksandsoda, on 07/21/2008, -1/+42If your business depends on any IT you need to have an adequate Business Continuity Plan and Disaster Recovery procedures in place which don't shift all the risk to a third-party provider. People love to mitigate risk by handing it off, but the bottom line is anyone taken down completely because of this just didn't do sufficient due diligence when planning out their technology infrastructure.
- dsenman, on 07/22/2008, -1/+25As a S3 customer with a business who uses it for a mission-critical role, it certainly sucked. But like other owners, I took a calculated risk when deciding to outsource to Amazon. Although I risk downtime like we had yesterday:
1) Everything crashes at some point. ***** happens.
2) The cost savings are enormous. We sped 1/4 what we used to when we ran our own SAN. I'm talking over $50k/yr in savings.
3) Amazon knows their stuff. I have a great deal of faith in their technology and team, even if there is the occasional downtime. Their constant updates yesterday were excellent.
4) Less stress. I don't have to worry about running my own equipment - I know that in this case someone else is doing the work for me.
Yes, 6hrs is a long time. But I'm willing to risk that occasional possibility for the rewards listed above. It's definitely not for everyone, but for many of us S3 is an incredibly valuable tool - and my confidence in it hasn't budged the least. - netgeek06, on 07/21/2008, -3/+166 hrs is a lot of downtime.
- fsuarez2005, on 07/21/2008, -4/+16Ha! Reddit had problems because S3 went down.
- iota, on 07/22/2008, -0/+11hate to have to digg you up but its true
- mmcgrath, on 07/22/2008, -0/+11I'm not saying Amazon should have this downtime, but if you're using S3 for something in your critical path (IE: it makes you money) then you're what we call an early adopter. New technologies mean risk, it's your job as the early adopter to identify and mitigate risk.
- skinrock, on 07/22/2008, -0/+10A lot of people are paying for it...
- SpacePirate, on 07/22/2008, -0/+9That's why you should really look for _at least_ 4 nines. 99% just isn't that good.
- bradleyland, on 07/22/2008, -0/+9The thing is, S3 is one of the few products of its kind that are within the reach of small businesses and startups. Yes, it's good practice to have redundancy at every step of the way, but what would be the *real* costs associated with having a roll-over replacement for S3? And who would you go to in order to purchase this service? Akamai? Akamai doesn't even want to talk to my customers, because they're not nearly big enough. There are a number of market specific CDN services out there, but most require a "bucket" purchase model, where you buy what you anticipate you will use, rather than pay only for what you use. This is a significant fiscal advantage for startups, and when you're competing in a market with other companies that use S3, you can hardly afford to double or triple your costs.
S3 has been a boon to "long tail" sites and services. It brings large-scale infrastructure within the reach of your average development shop. The flawed assumption by the community at large was that Amazon would be able to deliver the same uptime through S3 as they do their own website. This assumption was not entirely illogical. Amazon touts S3 as being based on lessons learned from the operation of the largest online retail store in the US. - mindracer, on 07/22/2008, -0/+7I think its reasonable to expect some downtime during the year, and agree with poprocksandsoda you should have a backup server/provider and code your site/servers to switch over to that backup temporarily, if your business depends so much on the internet that you cannot afford a few hours of downtime. my 2 canadian cents
- mrsteveman1, on 07/22/2008, -1/+67 8 all of them
- jemka, on 07/22/2008, -0/+57 did, and scared the piss out of 6.
- AndreasTh, on 07/22/2008, -0/+4INSUFFICIENT DATA FOR MEANINGFUL ANSWER
- caltheos, on 07/22/2008, -0/+4More like what happened to 5 9's.
- bradleyland, on 07/22/2008, -0/+4Can I get the name of that "absolutely worst shared hosting" provider? I'd like to go ahead and move all my websites over to them.
- karlyguy, on 07/22/2008, -0/+3The article expects them to have 99% uptime - so over 1 year that allows for 3.6 days of downtime. If their service was out for 6 hrs recently and 6 hrs in february. I think they are meeting the SLA of 99%
unless the article actually meant to say 99.99% that would be different. But welcome to information system technology - 99.99 is not gonna happen. - skoles, on 07/22/2008, -0/+3Let us ask Multivac how to achieve that extra 1%.
- wwwdeveloper, on 07/22/2008, -1/+3About the Author:
"ReadWriteWeb was founded and is edited by Richard MacManus, from Wellington New Zealand. Prior to RWW Richard did research, analysis and product development work for companies in Silicon Valley and beyond. He has also worked for some of New Zealand's top organizations as a Web Manager."
I wouldn't hire him to do analysis if he can't even figure out that 6 hours a year is > 99% uptime. - Drizzit, on 07/22/2008, -0/+25 9's is a marketing term. It does not exist. There are still outages but they're scheduled and if they're scheduled then they're exempt from the 5 9's accounting.
Find me someone who says they have less than 5 minutes of downtime yearly, and I'll show you a liar. - t0nic, on 07/22/2008, -0/+1I am amazed every day it seems about the lack of processes and procedures of companies with regards to disaster recovery and continuity plans. I guess its expected in my line of work (consulting) to see the worst of the companies out there but its just downright jaw dropping how fortune 500 companies operate sometimes. If I had a nickle each time I've seen horrible plans I'd be able to retire a happy man right now.
- kalikkalik, on 07/22/2008, -0/+1I thought 7 8 9?
- anthonywr, on 07/22/2008, -0/+1Well stuff goes down, thats life.
But what we call cloud computing today is a misrepresentation of simple client-server computing. We have a lot more work ahead of us before we are truly cloud computing rather than pie in the sky computing - booyahbitch, on 07/22/2008, -0/+1I tend to agree with you on this...most places are looking to accept risk instead of mitigating it...the worse risk model is the one where you accept the risk always.
- epheterson, on 07/22/2008, -0/+1This is a good thing.
S3 is relatively young and surprisingly got a lot of clients rather quickly. They couldn't handle it, I presume they didn't forcast such a success.
After a great catastrophic failure, things turn around. They know that after this they must earn the respect, and business, of their clients and not take it for granted -- that is they will likely be the most reliable provider. This is a very embarrassing situation for Amazon and they do not want to be in the spotlight, under these circumstances, again.
They will put money, effort and manpower into ensuring this doesn't happen again. Watch for it. - MacHarborGuy, on 05/20/2009, -0/+1no host is perfect. hell, i had a web host that installed a new update on their servers, and over half of the servers were down for the better part of 2 days. needless to say I am no longer with them
- mardraum, on 07/22/2008, -3/+2it came as a complete surprise when Ocelot discovered the S3 data from GW. not a bad idea though - using fire to fight fire, creating the perfect assassin to retire Solid Snake’s brother.
S3 stands for Solid Snake Simulation. it’s a development program to artificially reproduce Solid Snake, the perfect warrior. the result is a FOXHOUND commando when FOXHOUND no longer exists, a simulated Solid Snake shaped by VR regimen... sound like someone you know, Jack? - tenio, on 07/22/2008, -4/+3if 6 hours is too much...get a new (probably more expensive) host
- Braxo, on 07/22/2008, -4/+2Why doesn't Amazon use S3 for their site?
- scratt, on 07/22/2008, -6/+2Were Amazon's own sites up and running during that downtime? I'd wager money they were.
Even the absolutely worst shared hosting I have had the misfortune to use has never left me out of business for that length of time.
Amazon should be ashamed of themselves. And they should do some work on their contextual advertising also. It's so random it's almost a joke! - taseedorf, on 07/22/2008, -6/+0Agreed with poprocksandsoda. People who depend on third party services often are left with a disaster.
Simple solution is hire an intelligent staff who can back up and store everything on site, with alternative power sources available. While this might not be viable for a lot of small businesses, is is a necessity.
http://trillabyte.com - arcticsoft, on 07/22/2008, -9/+2but its not comparable. you dont really run a business on mobile me do you?
- tenio, on 07/22/2008, -8/+1if 6 hours is too much...get a new (probably more expensive) host
Browsing Digg on your phone just got easier with our enhancements to the 
© Digg Inc. 2009
— Content created and posted by Digg users is