Multicom Hijacks: Do you peer with these turkeys (AS35916)?
Well, it took less than a day for my last missive here to get the hijacks associated with AS202746 (Nexus Webhosting) taken down. I guess that somebody must have smacked Telia upside the head with a clue-by-four at long last. So, with that out of the way, let's see what else I can accomplish this week. As I understand it, the theory is that the thing that keeps the entire Internet from descending into the final stages of a totally broken "tragedy of the commons" is peer pressure. As everyone knows, there is no "Internet Police", so the whole system relies on the ability and willingness of networks to de-peer from other networks when those other networks are demonstratably behaving badly. Let's find out if that actually works, in practice, shall we? According to bgp.he.net, the top three peers of AS35916 (Multacom) are as follows: AS2914 NTT America, Inc. AS3223 Voxility S.R.L. AS209 Qwest Communications Company, LLC I'd like help from any and all subscribers to this mailing list who might have contacts in these companies. I'd like you to call their attention to Multacom's routing of the following block specifically: 163.198.0.0/16 This is a long-abandoned Afrinic block belonging to a semi-defunct company called "Agrihold". In fact, this block was a part of the massive number of hijacked legacy Afrinic /16 blocks that I pointed out, right here on this maling list, way back last November: https://mailman.nanog.org/pipermail/nanog/2016-November/089164.html After that posting, whoever was responsible for all those blatant hijackings got cold feet, apparently, and stopped passing all of those bogus route announcements out through their pals at AS260, Xconnect24 Inc. And so, for a brief time at least, the wanton pillaging of legacy Afrinic /16 blocks, and the reselling of those stolen blocks to various snowshoe spammers stopped... for awhile. But it appears that on or about January 6th of this year, Mulutacom lept into the breach and re-hijacked both the 163.198.0.0/16 block and also the additional Afrinic legacy block, 160.115.0.0/16. (They apparently stopped routing this latter block some time ago, for reasons unknown. But that fact that Multacom was indeed routing this second purloined legacy Afrinic /16 block also is in the historical records now, and cannot be denied. Multicom's routing of both blocks began around January 6th or so of this year, 2017.) Just as a courtesy, I sent the block absconders at Multacom a short email, earlier today, asking them if they had an LOA which demonstrates that they have rights/permission to be routing the 163.198.0.0/16 block. Of course, the mystery person (noc@) who emailed me back claimed that they did, but unfortunately, he was not under oath at the time. I asked if he could show me a copy of this purported LOA, and I haven't heard back from anybody at Mulatcom ever since. I don't really think there is any big mystery here, nor do I think that Multacom has or had, at any time, any rights to be routing these two legacy Afrinic /16 blocks. But they have done so, and continue to do so, in the case of the 163.198.0.0/16 block at least, quite obviously because -somebody- is paying them to do it, even in the total absence of a legitimate LOA. And as it turns out, it is quite easy to figure out who Multacom has been routing these two hijacked legacy Afrinic /16 blocks both for and to. It's trivially easy to run a traceroute to any arbitrary IP address within the 163.198.0.0/16 block. No matter which one you pick, the traceroute always passes through a particular IP address, 178.250.191.162, before the remainder of the traceroute gets deliberately blocked. That IP address is registered *not* to some long lost African concern, but rather to a Romanian networking company called Architecture Iq Data S.R.L. That company itself is apparently owned by a fellow by the name of Alexandru ("Andrei") Stanciu who hails from the city of Suceava, Romania. (Note that this is apparently *not* the same Alexandru Stanciu who the FBI arrested on bank and wire fraud charges in 2014. That one apparently hailed from Bucharest.) Anyway, "networking" seems to be only one of our Mr. Stanciu's many and varied business interest. His networking company, Architecture Iq Data S.R.L. has a web site (http://architekiq.ro/) but it is "shallow" to say the least. Many, and perhaps evenmost of the links on the home page of that company's web site seem to lead nowhere. In cotrast, Mr. Stanciu has the following other well-developed web sites and companies: ads.com.ro promoart.ro largeformatprinting.ro Promoart S.R.L. Advertising Distribution Supplies S.R.L. Mostly, he seems to be in the advertising business, as evidenced by the above web sites, and also by his membership in the "Email Marketing Gurus" special interest group over on LinkedIn: https://ro.linkedin.com/in/alexandru-stanciu-8846aa12a Given Mr. Stanciu's apparent professonal interests, it is not really all that surprising that the two hijacked legacy Afrinic /16 blocks that Multacom has been kind enough to route... both for him and to him... do in fact seem to be associated with numerous domain names that obviously consist of just two random dictionary words smashed together, followed by either .com or .net. This exact motif is quite commonly used by and among many of the Internet's most prolific snowshoe spammers. And of course, Mr. Stanciu's snowshoe spamming domains would not be maximally productive unless they each had SPF TXT records attached... ones that would pass muster with the recipients of Mr. Stanciu's spams. Those SPF TXT records are listed here, along the relevant domain names: https://pastebin.com/raw/BbK2YGe6 (Whenever possible snowshoe spammers also like to be able to send out their spams from from IP addresses where they have already set up nicely mattching reverse DNS, because a lot of recipient mail servers these days just won't accept inbound email anymore from no-reverse-dns IP addreses. But unfortunately for Mr. Stanciu, and for Multacom, the fact that they both just sort of walked off with the 163.198.0.0/16 block means that although they can -route- that space, they can't get the authority to control the reverse DNS for this block delegated to them. In order to do that, they'd have to get permission to do reverse DNS for the block FROM THE REAL AND LEGITIMATE BLOCK OWNER. And since that ain't them, nor even anybody who even knows what these clever fellows are up to, they can't. So Mr. Stanciu is stuck sending out his spams in a sub-optimal way, without either matching reverse DNS or even *any* reverse DNS for the entire /16 block he's stolen. Sorry Mr. Stanciu! Sorry Multacom!) As anybody who understand this stuff will by now be utterly convinced, the legacy Afrinic address block, 163.198.0.0/16, has been hijacked, stolen, or whatever you prefer to call it, by Mr. Alexandru ("Andrei") Stanciu of Suceava, Romania, specifically for "snowshode" spamming purposes, and with the significant help and assistance of AS35916, aka Multacom Corporation of 16654 Soledad Canyon Rd #150, Canyon Country, Calfornia, 91387, which is actually the entity announcing the routes to this clearly illicitly "liberated" IP block. So now, would one or more of you kind folks on this list who are more fortunate than me, and who have connections please be so kind as to let the following entities know about what Multacom is acctually up to here? AS2914 NTT America, Inc. AS3223 Voxility S.R.L. AS209 Qwest Communications Company, LLC Maybe they won't care, but they should. Maybe we can find out if the notion of peer pressure... or perhaps even de-peer pressure... works as well in practice as it allegedly does in theory. Thanks for listening. Regards, rfg
Dear Ronald, Thanks for your report, we'll investigate. Kind regards, Job
RIPE or one of dem dere responsible RIRs should hire him. I got a sales call in a few weeks with NTT, let's see if Job is successful and then I can be duly impressed and even more interested in their products. This shit actually matters, sometimes. /kc On Thu, Aug 03, 2017 at 12:23:51PM +0200, Job Snijders said:
Dear Ronald,
Thanks for your report, we'll investigate.
Kind regards,
Job
-- Ken Chase - math@sizone.org Guelph Canada
A few years back, Ronald named-and-shamed my work's new carrier for facilitating a prefix hijacker on this very list. As luck would have it, I had a fresh, crisp business card from our sales rep, so I passed the (quite legitimate) grievance along, and a short time later, the hijacked prefixes had one less upstream. Years later now, I have a different job, and a circuit with AS209. I'll see if I can't scare someone up (if it's still active by the time I get into the office). Thanks Ronald. Rest assured that many of us remember. :-) Jima On 2017-08-03 05:21, Ken Chase wrote:
RIPE or one of dem dere responsible RIRs should hire him.
I got a sales call in a few weeks with NTT, let's see if Job is successful and then I can be duly impressed and even more interested in their products.
This shit actually matters, sometimes.
Oops, following up a bit late. I was told yesterday that AS209 blocked their acceptance of 163.198.0.0/16 from AS35916 based on a number of complaints, which unfortunately left a path via their peering with NTT (which I presume they can't filter for $reasons). But then a short time later AS35916 withdrew the announcement entirely, possibly because of the traffic engineering implications of that filtering (not sure). Short-term, it's a win, but long-term we may not have seen the last of this prefix. Jima On 2017-08-03 06:24, Jima wrote:
A few years back, Ronald named-and-shamed my work's new carrier for facilitating a prefix hijacker on this very list. As luck would have it, I had a fresh, crisp business card from our sales rep, so I passed the (quite legitimate) grievance along, and a short time later, the hijacked prefixes had one less upstream.
Years later now, I have a different job, and a circuit with AS209. I'll see if I can't scare someone up (if it's still active by the time I get into the office).
Thanks Ronald. Rest assured that many of us remember. :-)
On Thu, Aug 03, 2017 at 02:52:43AM -0700, Ronald F. Guilmette wrote:
And of course, Mr. Stanciu's snowshoe spamming domains would not be maximally productive unless they each had SPF TXT records attached [...]
FYI, 85 of the 101 domains listed here are have been picked up by various spammer domain detection methods in place here. I have no doubt that the other 16 either will be in due course or simply reflect an inadequacy in my methods. ---rsk
Also AS57166 (single upstream AS29632 NetAssist) is likely hijacking 10 ASNs, and AS43659 (currently inactive). Both with mnt-by: D2INVEST-MNT. http://bgp.he.net/AS57166#_peers DATASTAR-MNT created 14 autnum and 31 route dummy objects in RIPE, on resources that looks abandoned (2 of them confirmed hijacking) https://apps.db.ripe.net/search/query.html?searchtext=DATASTAR-MNT&inverse=mnt-by;mnt-domains;mnt-irt;mnt-lower;mnt-nfy;mnt-ref;mnt-routes&bflag=true&source=RIPE#resultsAnchor Someone actually mentioned these back in Oct https://mailman.nanog.org/pipermail/nanog/2016-October/088487.html
participants (6)
-
Jima
-
Job Snijders
-
Ken Chase
-
Rich Kulawiec
-
Ronald F. Guilmette
-
Yang Yu