Slashdot Mirror
Why Blacklisting Spammers Is A Bad Idea
Roland Piquepaille writes "For the last two months, an eternity in Internet time, I was unable to reach -- and to contribute to -- Smart Mobs, the collective blogging effort around the next social revolution initiated by Howard Rheingold. Why that? Because an unknown customer of Verio decided it was a spamming site and asked the company to blacklist the site. Verio complied -- probably without even checking it -- and my problems started. It took me dozens of e-mails and phone calls and two visits to the headquarters of my french ISP, Noos, to fix the situation. More about this horror story is available here."
And other RBLs require usually multiple reports from multiple sources. And you have fairly straightforward way of getting de-listed, too.
What's with the current boo-hoo over blacklists? Do we have some kind of spammer astroturf going here?
This article should have been called...
"Why it's important to have good policies and procedures in place when blacklisting spammers"
I use blacklists to mark probable spam, but still generally see it. Recently, some people had reported an email from GoDaddy (domain registrar) that was only sent to customers, and it was asking them to very information. If, say, my ISP was blocking email from them based on this, I'd never see it. ISP's should err on the side of caution, let users take more risks if they personally desire.
"blacklisting" in this article refers to completely block an ip address. This is not a "bad idea", but complete nonsense. First time I've heard of something like that. This is not to be mistaken for using an open relay blacklist or similar, which only blocks mail from a certain address. I bet those "network administrators" clicked on some fancy "block site" button, not knowing what they were doing...
The fact that a strategy (such as blacklisting) can be mismanaged and that it is not invulnerable to abuse does not necessarily make it a "Bad Idea". It just means it needs to be managed more carefully, and better secured from abuse.
Why is the blacklist being done on a domain level. Spam is usually email....so block the email address. That is simple enough to do with intrusion detection systems, some application level firewalls, and if your really bored....an access list on a router. Whoever decided to block ftp or http to stop spam was not all there. They should have stopped smtp traffic from there instead and been done with it.
Black listing of spammers is a good idea, we just have to make sure we are only blocking them and not innocent bystandards.
Stop signs are only Suggestions
From the article: My ISP has a partnership with Verio to handle its traffic in the U.S. When Verio blacklisted Smart Mobs, any request from Noos went unanswered -- sorry, there was the (in)famous 404 error.
I want to be sure I understand this correctly. Verio wasn't (only) discarding mail from Smart Mobs, because they thought it was spamming site, they were refusing to pass through http (or other) connections to it?
Discarding mail is one thing, but blocking an IP address is quite another. What's the justification for this? To prevent the (supossed) spammer from profitting from the spam, by preventing anyone from connecting to it to (presumably) buy the product touted in the spam?
Discarding mail from a spammer can be justified, by, among other things, the argument that spam mass-mailings strain system resources. But connecting to sites happens all the time -- an ISP should should be set up to handle that traffic, and can traffic to sites touted in spam really increase the volume that much?
To me, this seems like a dubious policy on Verio's part -- even without the problem of mis-identifying sites as in the case of Smart Mobs.
Opinions on the Twiddler2 hand-held keyboard?
I left an HTTP proxy on on an open port - on the same machine that does SMTP. I didn't even know that spammers could relay via an http proxy using a PUT to the local SMTP server. mea culpa.
I fixed it in 3 days (too long, I know).
I contacted mail-abuse.org and submitted a removal request. It took them 2 weeks to take me off the list.
It frustrates me that their site is so unresponsive to removal requests, and that they fail much of their process. They were supposed to send email at several stages, which they did not do. The email they did send was badly formatted (broken urls, urs that weren't relevent).
I won't ever use an RBL because they just don't seem responsible.
Yeah, I know - pot kettle black. But I'm not supplying a service to thousands of users.
So the question presented by this article would be "WHY is blacklisting spammers a bad idea?" Unfortunately, it doesn't answer the question.
The blurb mentioned by the article submitter is the entire coverage of any such activity. The rest of the piece then goes on to complain about the user's ISP. Those who haven't RTFA'd can feel comfortable in skipping this one.
I'm sure this submission will provide nice fodder for expressing annoyance over spamming and horror stories of "collateral damage". But then - we've had plenty of those before. It would have been nice if an article had provided some framework around this kind of conversation.
This article doesn't.
The good it does is far outweighed by the bad. Just like everything else in life, mistakes will be made. You can have a problem with the process to correct mistakes, but advocating RDNS blacklisting should go away doesn't make sense.
How about making use of micropayments so that sender's account is charged some nominal amount that goes into receiver's account?
How about not? Of all of the proposed solutions to the spam problem, micropayments are the worst.
In all fairness, some of this does make sense. 50% of calls are "When I click on the E thing I can't see the Internet.", the "Internet" being your ISP's home page. The first questions from technical support is usually your number, your name, what kind of cable modem you have, and how many lights are on. Fixing basic connectivity solves over half the problems.
40% of calls deal with email issues, of which half are actually connectivity problems, the rest are customers with a new computer that need to have their settings switched over to their new machine. This can be checked by having the customer email themself.
This leaves us with 10% viruses, spam, malware, browser settings, router settings, bricking, QOS/server issues, and the occasional kook that insists the ISP is blocking access to some obscure site.
Based on this story, it seems Verio decided to block the presumed source of spam by means of the routers. That's a rather extreme measure. Doing such things in routers, whether by access list, or by blackhole routing table entry, is not nearly as easy, and does not scale as well, as blocking at the receiving mail server. But they may have wanted to do so because so many mail servers are run by clueless people that can't configure their way out of a paper bag.
I block spam source at mail servers, not routers (except in very extreme cases, but there are current none blocked at routers). That gives me the option to whitelist specific senders and/or specific recipients. So I'd say the real issue he is not that blocking/blacklisting spammers is bad, but that blocking them in stupid ways that lose control is what is bad.
Blocking spam and spam sources should be an end-point decision. There are risks in blocking, and different people have different needs and different sensitivities to that risk. Even your own ISP shouldn't block spam for you unless you agree to it with the understanding of how they are doing it. The best solution is for you to have total control if you wish, particularly in the ability to whitelist, and even blacklist, specific exceptions you want. Those who don't know the details of how this is done would have to delegate that to someone (such as their ISP).
Even content based spam filtering can be broken. What if my girlfriend sends me mail telling me what she's going to do with certain parts when she comes over tonight. I sure would not want that to bounce. Of course I can whitelist her email address (and hope her computer doesn't get infected by some spamming virus).
Blacklisting spammers is good ... when done right. Verio didn't do it right.
now we need to go OSS in diesel cars
Why when this happened did he not instantly start shopping around and then demand to speak with a manager and tell them that unless they got a clue about the diffrences between protocals that he was leaving?
Actually you are right. The real problem is people willing to put up with shitty customer service. If enough people stopped putting up with it, and did switch, we might actually see some corporate changes.
When my last ISP gave me crap about a similar problem, I immediately started looking for a new ISP and eventually switched to Speakeasy. It was the best move I could have made. I have been with Speakeasy for over 3 years now and I honestly could not be happier. I have customer service that listens to me and technical support that actually helps me.
-sirket
First, it's obviously a bad idea to block all IP traffic for an entire netblock (except under extreme circumstances -- attacks, for instance).
Spam is a huge problem, and there are some very effective DNSBL's (DNS blocklists) out there that can let a mailserver reject mail coming from a certain IP address. There are many different DNSBL's out there, and each has their own policies on what IPs they will list, how they will de-list, etc.
I don't like DNSBL's that list IPs based on non-spam related criteria. Examples include: country/continent of origin and service class (consumer vs. commercial). Blocks based on such criteria just divide the Internet, and don't even take into account where spam is coming from. I think it's a slap in the face of the Internet for a company to say, "I'm going to block all traffic from dynamic IPs, because they are not commercial connections".
Then there are the blocklists that block IPs that send spam. I like this approach because the lists are designed to block what I don't want; spam. sbl.spamhaus.org blocks regions of the Internet that perpetually send spam. blackholes.easynet.nl similarly list established spam sources. relays.ordb.org and list.dsbl.org block open relays and proxies that were found to be points of abuse.
Over the past 6 months, some 65% of spam (and spam attempts) that my ISP received came from less than 0.16% of the assigned IPv4 address space.
Almost 2/3's of the spam we saw was sent over SMTP connections from one of 77 CIDR blocks (ranging from /16 to /30 in size).
These 77 CIDR blocks represent less than 1/6 of
1 percent of the assigned IPv4 address space.
BTW: The CIDR list growth factor is not much when you move from the 65% level to the 90% level.
Spam is truly a world wide problem. Those 77 blocks, by national/region, break down as follows:
"Yes, Virginia", a few IP address blocks do transmit most of the spam.
chongo (was here)
6 months later, its proponents were telling people the same thing - "every entry was verified an open relay" (here)
Of course, these lists can be workable when combined with a system such as spamassassin, which uses them to weight whether or not a message might be spam, thus taking into account the too often power tripping and overreacting operators.
It must be frustrating playing whack-a-mole with spammers, but, slandering entire network service providers is wrong, too.
Remember the old adage: "be careful when you fight monsters lest you become one yourself"?
Or, how about "100 guilty men go free than for one innocent man to be put to death"? Just like with censorware, when people see legitimate sites and users suffering at the hands of the "protectors", it leads to wariness of placing much trust in these "protectors".