Is this affecting iMessages? No one’s replying back to me.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Multiple Major Websites Down Amid Akamai Outage
- Thread starter MacRumors
- Start date
- Sort by reaction score
Because you are not in the IT Industry.
A couple months ago my area had a huge credit outage which affected many businesses including the company I work for. I travel between locations with my IT job and was at a location about an hour away at the time. Luckily I had a twenty tucked away for an emergency so I could have enough gas to get back. The outage only lasted a few hours but I make sure I keep a twenty always tucked away now because I never normally carry cash.
Why would or should airlines like Delta or Southwest build out their own CDN when that's not what their business is or their specialty?
These are the same airlines that have had to delay or cancel hundreds of flights almost every year due to repeated network problems of their own. You really think they'd do a better job than the CDN providers like Akamai, CloudFlare, Fastly, Google, Amazon, CacheFly, StackPath, etc.?
News
Discover exciting world events, luxury travel deals, safety tips, and more. View the latest international travel news and information at T+L.
Widespread computer problem at United delays travelers for hours
Travelers flying United late Thursday reported hours of delays after the airline was hit by a widespread computer problem.
money.cnn.com
Multiple airlines report early-morning delays due to technical issue checking passengers in
The first reports started popping up on Twitter at about 2 a.m. Central time Friday
www.khou.com
Delta computers crash, causing delays and cancellations. Experts say it shouldn’t have happened.
The airline blamed the problem on a power outage at its Atlanta base of operations.
www.washingtonpost.com
Last edited:
it’s all good bro. Your mom says hi.
Most likely they just moved their traffic between other CDNs. I worked for Limelight which is another big CDN that delivers traffic for a good chunk of the Internet. We were usually not the only CDN a company had. We did business with just about everyone and the big tech players have 3 CDNs generally. Akamai has the largest network by pushing their edge devices into every single ISP, but their prices are terrible and their management technology is junk. That's why Fastly has blown up in the last 5+ years. Their speeds are impressive, their cost is better, and their API is really nice.
What exactly happens when a whole service like this goes down? I thought it was designed to have backups and redundancy .
Its not decentralyzed. I think there are about 4-8 main data center companies, if they go down basically the whole internet as we know it will go down. Plus I think domains names must be registered with ICANN, one issuer of domain names.
I heard once 40% of the internet works on AWS
Whats the difference between a CDN and a data center? I know a CDN provides a copy of the data at a data center to a closer location to the user. Doesn't that make every CDN a data center too? You can buy a server for whatever use you want?!
Its not decentralyzed. I think there are about 4-8 main data center companies, if they go down basically the whole internet as we know it will go down. Plus I think domains names must be registered with ICANN, one issuer of domain names.
I heard once 40% of the internet works on AWS
Whats the difference between a CDN and a data center? I know a CDN provides a copy of the data at a data center to a closer location to the user. Doesn't that make every CDN a data center too? You can buy a server for whatever use you want?!
Apple has their own servers in several datacenters around the world plus they rely heavily on cloud storage in several different clouds and a good number of CDNs. There may be an advantage to them eventually building out a CDN as they deliver more and more video content. That's what Netflix did when they left Limelight years ago. They built out a really big network slowly, but it's a lot of costly work. It requires a large amount of software development as well as rolling out tens of thousands of servers to ISPs around the world, which you then get to manage. There's a lot to be said for recognizing your core competencies as a business and just paying someone else to do that even if it costs more.
“According to Apple's System Status page, iCloudBackup, iCloud Mail, iCloud Storage Upgrades, and Photos are experiencing issues that could be related to the outage.”
I think Apple’s problem is due to me trying to add my 101st iCloud Hide My Email account in the Beta phase.
Was kind of surprised there is a limit on this. If you can’t give a unique email to each website, it’s gonna punch a hole into the bottom of Apple’s no fingerprinting initiative.
I think Apple’s problem is due to me trying to add my 101st iCloud Hide My Email account in the Beta phase.
Was kind of surprised there is a limit on this. If you can’t give a unique email to each website, it’s gonna punch a hole into the bottom of Apple’s no fingerprinting initiative.
Their co-founder died in the 9/11 terrorist attacks. They never recovered from a technology perspective and basically managed to stay in business with huge contracts they garnered during the initial dot-com boom. After that they just hired sales geniuses and engineers good enough to keep things limping along.
If you look at a cloud provider like Amazon they have regions where they cluster dozens of data centers together to provide services. In the US that would be regions like US-east-1 , US-west-1, or US-west-2 (go Oregon). As a business, you're putting your systems in one or more (hopefully more) of those regions. The problem is those regions aren't right where your customers are. When you stream content on iTunes there's certainly a good amount of API calls that happen there to authorize you, to see what songs are available, to toss ratings data, etc etc and those will come from Apple's systems. There's no need for Apple to actually give you the song though from their Oregon datacenter if you live in say Florida. You want that content to be really close to you to prevent latency and jitter as you stream it. The CDN sources the original content and distributes it to tens of thousands (or hundreds of thousands) of servers globally to get that content super close to you. If you're a Comcast customer every city has Akamai gear and Netflix gear (they run their own CDN) in it. Other CDNs like Cloudflare, Limelight, or Fastly have collections of CDN servers in major metro areas then they peer fiber to the ISPs in those regions. The ISPs love CDNs because it means they don't have to backhaul that traffic over expensive fiber links. Sometimes it comes entirely from their head end. A great example of where something like a CDN really shines is software updates. When I worked at Limelight we distributed a lot of these for gaming consoles and some mobile devices that people here probably use. Say Apple releases iOS 15 tomorrow. They have that in their datacenter, but the various CDNs they use will distribute that around the world and going back to that Florida example 50,000 people in that city will download it from a series of Akamai nodes right in Comcast's equipment room. Never touches the greater Internet which makes it super fast.
Sounds like someone just did reset the Internet...
Be careful what you wish for, you may just get it.
That does make sense for Apple, but Amazon and AWS? They are the biggest cloud service provider on the planet, it makes zero sense for Amazon to be contacting CDN's from a smaller quasi-competitor.
Akamai is a staple in the content delivery industry. I'm sure you've heard of more recent members of this industry, including Amazon CloudFront and Cloudflare, but Akamai goes back years before these newer players and their marketing efforts.
Isn't this Akamai's second big blip this year? They've been known for being extremely stable and transparent. Apple has used them for decades to deliver media.
One of the ways to make the internet, as a whole, more stable is to spread resource loading out over several providers. It makes perfect sense for Amazon to be using other CDNs in their setup, in case their have their own outage. What doesn't make sense is that there was no fallback when this outage occurred.
When you use your phone or watch as a payment method, I don't believe it needs to have an internet or data connection. The device is the payment card.
Spreading across multiple providers is the best way to achieve resilency, not hosting everything in-house. Big companies that fail to spread across providers are the ones that are impacted the most.
You're making my point though. This was designed to be decentralized....or it CAN be decentralized but there's more profit in consolidation and capitalism gonna capitalize.
Akamai moves a very big part of the Internet around and when they go down, we suffer. AWS, Rackspace, Akamai, Limelight, Google, Azure or if ICANN screws up, it goes down.
Yup. Single Point of Failure. Bad design. I can understand smaller shops having this issue, but big companies should definitey have fallbacks in place.