New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
GMail has 425,000,000 active users. Assuming a more than reasonable average of 5 sent e-mails per day per user, that would come down to approximately 2.5 billion e-mails sent per day (the real amount is most likely much lower).
The total amount of spam e-mails sent per day, not including undeliverable e-mail or non-filtered spam, is about 150 billion. Note that there's only about 24 billion of legitimate e-mails every day!
Using the (sadly) limited Spamhaus statistics at http://www.spamhaus.org/statistics/countries/ as a base, assuming they have listed innocent IPs and missed some offending IPs, I estimate that there about 15,000 individual listings on Spamhaus, which I'll assume for the sake of simplicity represent 25,000 active individual servers. That brings us to 6 million spam e-mails being sent per day per server, on average. If we consider that GMail is sending 2.5 billion e-mails a day, they would only need 416 servers with equal resources to cover this kind of volume - and they are effectively the largest e-mail sender.
Taking into account that the average Google server is likely more powerful than the average spammer server, that GMail is likely represented by far more than 416 servers, that a measure like this would pretty much instantly kill spam altogether, that the disappearance of 85% of e-mail traffic would greatly lessen the load on their servers, and that even if Google had to run more servers it would be a minor expense... I would say that it is very feasible to implement such a proof-of-work system at a difficulty that deters spammers, but is still possible to operate with for a provider like Google.
EDIT: If anyone has better statistics about the amount of individual spam servers, I'm very open to those. This is the best I could dig up. The main point I'm trying to illustrate, is that spam sending differs from legitimate e-mail sending in that it operates on a very small margin, on commodity hardware, and in large volumes. A proof-of-work scheme would make this unsustainable.
@Ruchirablog there is a cable cut in mid east,Thats why slow.
http://www.webhostingtalk.com/showthread.php?t=1251115
The client sending the email could be the one to do the proof.
I doubt gmail, hotmail etc IPs even get blocked.
Chances are they've 'fiddled' the statistics. If you think about it, most of the zombies will be on multiple botnets as they will be the type of user that installs that kind of thing by accident. Some of the zombies they are counting are probably doubled and trippled in terms of statistics and they are probably counting maxed out upload speeds of lets say 100kbps. Most botnets you see on the news that are 'large' are typically 200k or so meaning realistically a 200k botnet could push a true 2gb per second. I seriously doubt anyone could push that kind of traffic.
These news reports hugely dramatize. They probably fudged it even more by taking the upload speed of the fastest possible connection found and giving it to all suspected zombies.
In terms of spam if again each user sent say 100 messages and you had a 200k botnet thats a lot of emails.....
Also keep in mind that some people use VPS to DDoS people so if they used multiple servers at the same time they could peak extremely fast speeds but even then they would need like 300 VPS and they would need to be maxing 1gb lines which would be basically impossible.
It's just another story designed to scare people.
I did indeed intend for the client to complete a proof of work. However, Hotmail and GMail are largely used as a web-based e-mail client, so it would be a lot trickier (but not impossible) to outsource the proof of work to the client. For the sake of simplicity, I've assumed the worst case where everyone solely uses GMail/Hotmail as a web client, and the browser is not reasonably capable of completing a proof of work.
I've experienced yesterday total communication blackout to (as far as i noticed at the moment) networks in the USA, i thought something big is going on, then it got better just a little bit, and started to barely open sites.
Luckily i am using distributed authorative DNS infrastructure for my most important services, but some of the services (like smaller sites) that aren't distributed i had to move them manually to another geo location.
PS: I've tried to access both websites of SpamHaus and CyberBunker right now, but they both seems to be down.
SpamHaus displays service down message from CloudFlare, and CyberBunker i just get HTTP connection timed out.
They aren't. The DDoSers are using DNS amplification to create a much larger attack than their own pipes can handle.
People have been suggesting these kind of improvements since forever, and there are already tons of solutions out there. This is like asking "The IPv6 RFC standard was published in 1998 - why aren't we using it yet?"
It's not that anyone prefers blacklists. It's the Email and SMTP, like IPv4 (but to a ridiculously higher degree) suffers from technological inertia.
E-mail was never really designed to exist outside of closed networks like ARPANET. However, ARPANET expanded into the internet and the email spec had to be hacked onto to work with newer and older clients, and over time things like attachments were added as a feature and the spec had to be expanded to work with both old and new clients, and this goes on and on until email is such a mess that it really should be thrown out as a whole because things like encryption or proof-of-work would only truly work if every provider supported it.
So, instead, we make do with what we have and used blacklists and bayesian filters and to make a best guess at what is and isn't a valid e-mail. It's not a sure thing like proof of work would be, but it's a lot easier than the alternatives like getting worldwide cooperation of adoption of a new standard or a technology or fragmenting communications into networks that do and don't work with each standard.
Also, is being blacklisted on spamhaus really that big a deal? are some organizations so lazy that they say
If IP XXX.XXX.XXX.XXX appears on just spamhaus's blacklist, automatically score it at 100 and do not deliver?
Nr. 1. on HN at the moment:
A Note from one of Cloudflare's upstream providers (cluepon.net)
http://cluepon.net/ras/gizmodo
of us are trying to run for-profit commercial networks, so when 300 Gbps
of NEW traffic suddenly shows up and all wants to go to ONE location,
someone is going to have a bad day.
Yeah, like I said elsewhere, these attacks get noticed upstream since a flood to one client and/or destination. Upstream mitigated big time, not per se Cloudflare.
Richard Steenbergen
Chief Technology Officer
GTT
Public Company; 51-200 employees; GTLT; Telecommunications industry
May 2012 – Present (11 months)
Leading the strategic direction and operation of GTT's global network, spanning over 80 countries worldwide.
I asked the guy from Cyberbunker and the answer is simply: no
Spamhaus takes down websites and blacklists IPs and ISPs to blackmail people...
As someone that has in the past ended up without a functional e-mail address because Spamhaus listed the entire webmail provider I was using... yes, I would say that it's really that big of a deal.