Bing dies (briefly) after Microsoft hits wrong button Bing was hit by a widespread outage Thursday evening after Microsoft accidentally updated the live site with changes intended for a test environment. While the search site was offline for only a half hour, it embarrassingly was ankled the same day Microsoft was parading about new Bing features and updates to woo market share …
google has had outages too. gmail has a history of being offline for hours at a time, multiple times per year, and it's not beta anymore.
this "cloud computing" thing has a severe single-point-of-failure problem. that the biggest players in the business are having problems does not portend well for smaller companies that want to hop on the bandwagon.
Yep, and El Reg bit that hand when they strayed too close as well. As they should, damnit.
As for "the cloud", until it is possible for my clients to have TRUE multi-point internet access where I am (we don't, we have 2 providers, and they share the same trunk downline), I'm telling them to stay away from the single-point-of-failure solution.
OK, the same mistake could happen to anyone, but the fact it's Microsoft just makes it several orders of magnitude funnier :)
Bing may be the third most popular search engine around, but I can imagine it won't take many years for it to succomb to the same problem that affects most Microsoft products - bloat. And it wouldn't surprise me if at some point in time they roll out a version of Bing that makes heavy use of Silverlight. All the time increasing the eye candy while not concentrating too hard on the algorithms behind the information gathering, collating, categorising and indexing.
"OK, the same mistake could happen to anyone, but the fact it's Microsoft just makes it several orders of magnitude funnier :)"
I quite like that they've been open about it and admitted a mistake rather than just blame it on a hardware failure. It doesn't seem like "the Microsoft way".
Black helicopters, 'cos they're clearly up to something.
Well I thought I'd take a look at the map example.
Really? With first a warning to use a "supported browser," then the map draws (but decides Scotland isn't in Britain.) Then this "Welcome to Bing Maps Beta Sorry, but Microsoft Silverlight doesn’t work with this web browser." Followed by a classic Microsoft catch-22 your options are cancel, cancel, cancel or install silverlight, ie windows.
Oh well, I tried,
Everything else made by Microsoft is unreliable so why should Bing be any different?
Yes, it does happen to Google and everyone else but at least they will admit when they make a mistake.
Microsoft on the other hand either tries to cover it up or plays it off like it's no big deal.
Ok, well it may not be a big deal now but when they try to convince people and companies to trust them with handling their data an files then it becomes a big deal and for a company with a horrible reputation like Microsoft has, a company that can't seem to get anything right, I think it's going to be a tough sell.
It's just my opinion.
This rather suggests they're running a very small group of computers for them all to go down at the same time. If thats the case I guess it means they don't have plans for Bing to exactly takeover from Google anytime soon.
As for looking into what they can do to stop this - erm, make sure the dev's can't press a button and make the code go live? You know, permissions....even if it is Windows.
<blockquote>You know, permissions....even if it is Windows.</blockquote>
Yeah, but, they had to have permissions to C:\My Documents\My Development Site, and we couldn't work out how to do that without also giving them permissions to all of C:\My Documents.
In light of this technical hitch we got Dave from accounts to create a new folder called C:\New Folder, but so far nobody can get into it, apart from Dave, but he's on holiday now until next week.
"This rather suggests they're running a very small group of computers for them all to go down at the same time"
There's this great thing called asynchronous processing, you know. If it takes 2 seconds to push out an update to a server's config file, and you have 1,000 servers, you could kick off 20 threads to update all servers in a couple of minutes.
The only reason people notice is that microsoft puts it as default in IE, and in some of the most recent versions, makes change to google almost impossible (really, I just bought an Acer aspire one preconfigured so horribly that google was not possible to add to IE 8; I switched to ArchLinux 4.2 minutes afterwards).
MS have cunningly ensured that google is always on the 2nd page of the "select a search provider" and considering that most users just press "next", "next", "next", "next"... until the annoying prompts go away and let them get on with downloading porn. They've also rigged it so that google is worse reviewed than the others, even some of the utterly inane and unused search engines.
Conspiracy? where?
This type of thing happens to people who have badly configured test environments and poor change control. It's completely indefensible. The only thing that astounds me more than this incompetency at the most basic level is that people are trying to excuse them, on the basis that "at least they admitted it".
Releasing test code into production is a dismissal offence where I work and damned right too.
...It's a testing problem. Testers should *never* have access rights to the production site, it's that simple.
It's got fuck all to do with the cloud. This could have been running any any kind of infrastructure, and the same problem would have occured.
Steven R
[Fail, because you all do]
I am sure chairs are flying in Redmond as we speak over this issue.
I can see the ad-spinners at work now:
"Bing -- built as solid and reliable as an Xbox 360" (and perhaps, being hosted on a server-farm of these devices to boost their rank in sales of game consoles).
IMHO, Microsoft certainly deserves to be crushed by the Googlenaut. They can throw all the money and advertising dollars they have at trying to beat every other company in the world, without realizing that the central problem is THEM; their archaic and bureaucratic management style, the fact that they are invariably late to the table because they are too big and bloated to react to actual consumer needs.
Remember the October 2009 Microsoft data loss, where Microsoft erased the personal data of about 1 million Sidekick phone users in the United States. Emails, contacts, calendars, photos, erased.
Microsoft managed to recover a bit, but not all the data. I'm surprised the media hasn't followed up on that story. Microsoft has a long history of outages and data loss. I would not trust the Borg with any data.
I love it when cloud operators have a tiny brief period of downtime, which will no doubt be reduced to nothing in the future, commentards here believe that's the death of cloud. As if all their vulnerable stacks of raid servers all placed in one place ready to burn have 100% uptime and perfect security!
The cloud is coming. The national grid used to fail occasionally, but we're not all using our own diesel generators in 2009, and for reliability I know which I'd prefer now a days!! The main reason we use the grid? Cost! I guess this is meaningless to some people!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
