tech

Major banks, airlines in Australia, U.S. report online outages

11 Comments

The requested article has expired, and is no longer available. Any related articles, and user comments are shown below.

© 2021 AFP

©2024 GPlusMedia Inc.

11 Comments
Login to comment

Either Russians laughing at Ol' Joe or Chinese looking to inflict more pain on Oz.

4 ( +6 / -2 )

Damn, hijacked and black mailed and paying ransoms ..............................not good

1 ( +1 / -0 )

Russians. They're no fan of Australia.

1 ( +3 / -2 )

El Reg, citing an Australian IT site, is suggesting that an Akamai component called Prolexic is to blame, although this has not been confirmed.

With services like this, one tiny issue can cause enormous amounts of havoc.

It is a myth that the internet is difficult to break. It is actually very vulnerable. It is dependent upon a large number of complex bits of code. All dependencies are vulnerabilities. The smallest bug or incompatibility, often after an upgrade, can shut down huge amounts of it.

A distributed system would be more fault tolerant by design, but all systems can break (or be broken) and the more complex they are, the more likely they are to fall over.

Banks are particularly difficult to maintain as they typically link multiple disparate systems, many running on legacy code. And when they go down, the user impact crater is larger than for a retail website or app. Simple solution: Always carry some cash.

1 ( +1 / -0 )

Hackers Russians, Thieves Chinese, Thieves plus hacker NK

0 ( +1 / -1 )

In all likelihood the outage was not caused by hacking. All of the affected firms used a Content Delivery Network (CDN) called Akamai. They are an "edge platform", cybersecurity and webhosting service. Rather than hoset sites on a single centralized server, edge platforms have many, sometimes dozens of servers around the world. Content is accessed from the geographically closest server so response time is minimized. The problem with these is the software to accomplish this is complex and a simple configuration change from a normal update or fixit build can crash the system. That is pretty much what caused a similar edge platform Fastley to crash last week.

0 ( +0 / -0 )

The problem with these is the software to accomplish this is complex and a simple configuration change from a normal update or fixit build can crash the system.

That's not a problem. The strength of these systems is that they can control wide ranges of configuration with a single fix/push. Without that, the complexity would become unmanageable.

Their goal is to have fail safes to ensure that no is able to push anything that could crash the system, as happened this time. This was a bug that no one had thought of ahead of time, which has now been fixed. They will have built in tests to ensure it cannot happen again (unless they are entirely incompetent). Someone may say "yeah, but these bugs can happen again", and they'd be right. But that would also be ignoring that these CDN's run 24 hours a day, serving 10s of thousands of sites, with a set of fail safes that enable their service to run almost never having disruptions like this. This is an anomaly, not the standard. We shouldn't define something by its outliers.

0 ( +0 / -0 )

This is an anomaly, not the standard. We shouldn't define something by its outliers.

Fastely's analogous problem last week has been traced to poorly written configuration change that misrouted traffic leading to a crash. The cause of yesterdays problem at Akamai hasn't been publicly announced. The point I was making is that in all likelihood what happened yesterday was not due to hacking but due to a software problem. Incidentally CloudFlare had a similar problem a year ago. Many who post here are predisposed to assume any internet problem is the result of hacking and overlook the more prosaic causes of these problems.

0 ( +0 / -0 )

Many who post here are predisposed to assume any internet problem is the result of hacking and overlook the more prosaic causes of these problems.

I fully agree.

0 ( +0 / -0 )

A distributed system would be more fault tolerant by design

CDNs like Akamai are a type of distributed system. They cache static HTML pages within various servers and respond to requests with the fastest server for the user.

0 ( +0 / -0 )

Login to leave a comment

Facebook users

Use your Facebook account to login or register with JapanToday. By doing so, you will also receive an email inviting you to receive our news alerts.

Facebook Connect

Login with your JapanToday account

User registration

Articles, Offers & Useful Resources

A mix of what's trending on our other sites