Looks like issues at Verizon... maybe the transition to Frontier (which is supposed to close April 1st) isn't going well...

Edit: Looks like Miami is a focal point, too, and Florida is, of course, one of the regions being acquired by Frontier in just a few days...

Edit 2: Random aside, headline is rather editorialized. These outages aren't "across the internet". They appear to affect a single operator in a couple of major regions.

25% packet loss between Level 3 and Sprint

http://internethealthreport.com/Main.aspx?Metric=PL

... fallen to 12.5% now

10 hours later

Cogent <> Level 3 is the only problem link left with 13.33% packet loss

Savvis <> SBC is now 25% and multiple Verizon is 10%

Row Verizon, Column Verizon is yellow. It looks like they're having 2.9% packet loss sending packets to themselves!

I think Frontier is acquiring just Verizon's local network; no backhaul or anything.

Possibly related: Frontier, at least in my area, went down last night for a few hours.

FWIW, this site is very often inaccurate and out-of-date.

For more reliable and accurate information, the outages list is probably what you want to refer to.

Oh, thought it said "outrages", that's just another day on the Internet.

That "massive" outage happens to affect only one ISP and <0,003%(1, 2) of Internet users, no need to panic.

1. https://en.wikipedia.org/wiki/Internet_in_the_United_States#...

2. http://www.internetlivestats.com/internet-users/

Just called a friend in Baghdad over Skype about an hour ago and I can confirm that there are massive outages of the Internet there too. We had to call each other like 10 times for a 15 minutes call. It was very tedious!

Sounds like Skype is working as normal at least.

Huh, I have Frontier fios, near Redmond, and it's been up and down for the last day or so.

Does packet loss actually translate to unreachable hosts, or does me mean there is just going to be a slowdown due to having to send lost packets again?

Both. It slows down traffic because packets have to be resent, and because of TCP (which assumes that if packets are lost, the connection must be congested, and therefore data should be sent slower).

If enough packets are resent the connection may not complete or it may timeout (if a packet needs to be resent multiple times, for ex).

I'm the founder of NuevoCloud, and this is one of the problems we are working on. Basically, we can route connections around congestion. So for example, if your website is in dallas and the visitor is in germany, but the normal route has an issue.. then we might reroute it germany > virginia > dallas until the issue is resolved. So your website still stays up.

High packet loss makes things with a real-time dependency like gaming, streaming, etc. near-impossible.

WhatsApp just stopped working for me. I don't know if it is a local server thing or some central issue.

Is it possible that you are using whatsapp web? The app updated right now. I had the same issue.

Works fine here

From Italy all is fine ;)

There are countries besides US on the planet.

It's amazing if HN can recognize places outside of California, don't expect too much.

Pretty sure there arent :P