Amazon Cloud Goes Down Again

H

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
The forecast calls for a no clouds anywhere. Sucks to be Foursquare, Turntable.fm, Heroku, Flipboard and everyone else that uses Amazon's cloud computing platform.

Amazon reported errors in eight of its web services stemming from outages in northern Virginia. The failures began just before 11 am on Monday. By 2 p.m., many of the problems had been fixed, but Amazon was still reporting issues with its Elastic Compute Cloud, its database service, and Elastic Beanstock, a service that helps developers speed up the rollout of new software.
Click to expand...
 
One of our clients lost connectivity to their Office365 exchange server from two seperate locations today. You don't see that reported. Crap happens... doesn't mean hosting your own would be more reliable. Granted Microsoft's outage appeared to last maybe 5 minutes, but it still occurred.
 
THIS IS WHY WE SHOULDN'T DO THIS

Why is no one listening! It's Happening! It's Happening! You could have stopped this! I warned you!
 
What isn't reported is that someone there was downloading so much pron that he overloaded the system.

Five outages in a year and a half? Anybody know how many hours that is in total?
 
Problem is when it does happen you're f'd cause you have no access vs having the stuff on your own computer.
 
Yet cloud crap is still touted as the "new" and "great thing" of the future.
 
lilbabycat said:
THIS IS WHY WE SHOULDN'T DO THIS

Why is no one listening! It's Happening! It's Happening! You could have stopped this! I warned you!
Click to expand...
Because in this day and age common sense isn't considered hip & trendy and is therefore frowned upon. :(
 
I'm all for people having sobering lessons about the "wonders of the cloud".

Though it does suck for people who are trying to host websites on amazon that have need the scalability to handle traffic surges in connection with special events. I guess hosting on Amazon isn't quite as reliable as I initially thought.
 
c4v3man said:
One of our clients lost connectivity to their Office365 exchange server from two seperate locations today. You don't see that reported. Crap happens... doesn't mean hosting your own would be more reliable. Granted Microsoft's outage appeared to last maybe 5 minutes, but it still occurred.
Click to expand...

Funny how much chaos 5 minutes of outage can cause.
 
We have 1000's of VM servers running at our company on vmware... knock on wood, at least 5 years without an outage so far. So cloud computing, but cloud computing in house.

Works better.
 
It's not like Cloud Computing is all bad.

But like everything else, it shouldn't be the only thing to be relied on.
 
Apart from the insanity which is relying on somebody elses computer to access your own critical data, this must be what has been going wrong with Netflix for the past 4 days. (Not available in your location)
 
This is precisely the reason I've been saying for the last couple of years that cloud computing is a terrible idea.

It's like putting locks on your house that only work when it isn't storming outside. Sometimes, that will happen, and if you're not in your house when the storm starts, you're going to end up left out in the rain.
 
Cloud computing is good when you backup your information you need to save on the cloud but you keep a version of that info on your local computer. So in case your local copy is gone due to hard drive failure or something, then all would not be lost.

I don't know why anyone would use it as a single point of failure.
 
Stiletto said:
This is precisely the reason I've been saying for the last couple of years that cloud computing is a terrible idea.
Click to expand...
There isn't anything wrong with cloud computing — it's literally just the moving of resources from one context to another. There is something wrong with relying on one single cloud infrastructure to run your business, however. Any good online business has some sort of contingency plan.

The cloud isn't a magic bullet. There is no magic bullet.

Sly said:
Five outages in a year and a half? Anybody know how many hours that is in total?
Click to expand...
Less than Amazon's SLA allows for.
 
I thought the point of cloud storage type stuff was multiple redundant copies online so that if anything happened to one copy/one data center, it wouldn't matter as any traffic could just be redirected to the next copy/center?

Maybe I'm confused on what the real point of putting things in the cloud is :confused:
 
It comes down to money.

You can move your internal IT out of your location, and pay for the service your opex shifts over to capex.

Big wig company execs love that shit because it makes the bottom line better.

Most company execs could care less about technology, so when most of those slimy cloud sales guys some sneaking around, the execs buy into the hype. They’ll circumvent the infrastructure folks, and then all data and services is now hosted externally. Bottom line will look awesome up front, then you start running into other issues. Performance, reliability, and security. Lost productivity will mount and then general over all work slows to a crawl.

If you are an infrastructure guy, and your company is seriously considering moving to a paid private cloud. Start writing your resume because you'll become nothing more than a contract manager and your skill set(if you have any) will goto complete crap. Ontop of that.. if you work for a company that already has a fairly large IT infrastructure staff, but the leadership wants to farm out your infrastructure.. The leadership has zero faith in your skill set, and thinks the admin teams are a bunch of clowns regardless or not how good you may or may not be. It might be time to move on to a different company anyway.
 
ShinseiTom said:
I thought the point of cloud storage type stuff was multiple redundant copies online so that if anything happened to one copy/one data center, it wouldn't matter as any traffic could just be redirected to the next copy/center?

Maybe I'm confused on what the real point of putting things in the cloud is :confused:
Click to expand...

Moving things to the Cloud isn't a backup scheme or something done for redundancy. It's getting rid of the server room (at least for the most part) and paying someone else to host everything you need. They do the backups, they host your apps and your data and it's all done on a monthly billing cycle. If you need to allocate more resources, you can do so by hitting a site, or making a call and it's done pretty much right away. If you need a new app you do the same and it's ready for users to log into. There are quite a few good things about the cloud, in that regard, but there are also issues of uptime and security that have to be kept in mind as well. You know longer have the same level of control you did when the servers and all your data was in your building, or at least in your company.

kdh said:
It comes down to money.

You can move your internal IT out of your location, and pay for the service your opex shifts over to capex.

Big wig company execs love that shit because it makes the bottom line better.

Most company execs could care less about technology, so when most of those slimy cloud sales guys some sneaking around, the execs buy into the hype. They’ll circumvent the infrastructure folks, and then all data and services is now hosted externally. Bottom line will look awesome up front, then you start running into other issues. Performance, reliability, and security. Lost productivity will mount and then general over all work slows to a crawl.

If you are an infrastructure guy, and your company is seriously considering moving to a paid private cloud. Start writing your resume because you'll become nothing more than a contract manager and your skill set(if you have any) will goto complete crap. Ontop of that.. if you work for a company that already has a fairly large IT infrastructure staff, but the leadership wants to farm out your infrastructure.. The leadership has zero faith in your skill set, and thinks the admin teams are a bunch of clowns regardless or not how good you may or may not be. It might be time to move on to a different company anyway.
Click to expand...

Pretty much. Most execs have never understood IT. They've always seen it as a cost center and not valued how much money the IT department has saved or what all they bring to the table. They just see the expenses and the high salaries and think farming that out will save tons of money and be better for the company. In some cases they're right and in some cases they're not. It's a lot like outsourcing was. Everyone rushed to do it and some companies found it was more costly and they lost a lot of control, while they and others lost customers due to shoddy support or shoddy programming that was a result. Now, the Cloud and Bring Your Own Device are the current fads and IT departments are scrambling to try to get these things implemented without the company throwing caution to the wind to totally screwing the pooch.
 
wtburnette said:
Moving things to the Cloud isn't a backup scheme or something done for redundancy. It's getting rid of the server room (at least for the most part) and paying someone else to host everything you need. They do the backups, they host your apps and your data and it's all done on a monthly billing cycle. If you need to allocate more resources, you can do so by hitting a site, or making a call and it's done pretty much right away. If you need a new app you do the same and it's ready for users to log into. There are quite a few good things about the cloud, in that regard, but there are also issues of uptime and security that have to be kept in mind as well. You know longer have the same level of control you did when the servers and all your data was in your building, or at least in your company.
Click to expand...

Then there's no point if you already have a good setup for all that, other than supposed money gains?

And I didn't mean it was used for simply making a backup of local, but that by putting it on a "Cloud" service (and getting rid of your local server) your system was mirrored to different physical servers/locations so if one goes down it's not such a big deal. After all, I thought that was how they got their supposedly near-perfect uptimes as anything other than catastrophic failure in multiple locations wouldn't change availability. Is that not true in the least?
 
I suppose a hybrid-cloud would meet your definition. You store locally and in the cloud, so you have a higher uptime possibility then with either on their own. I think the biggest selling points for cloud services are SaaS and PaaS (software as a service and platform as a service). You can certainly also use cloud for storage as well, or a mixture of all the above. My biggest problem with the cloud has always been twofold, first uptime and second, security. I just have a hard time trusting company data (which for most also includes customer data) to a third party. At least with an in house infrastructure, you trust your IT team and your backup storage company (like Iron Mtn or whoever you use). I guess there is risk either way, but I'm leery of trusting cloud vendors over much.
 
You must log in or register to reply here.
Back
Top