Re: Zayo opinions
Zayo owns what used to be Abovenet. In my experience, your experience will vary from market to market, depending on the network you're based on. As of late, we've had repeated capacity issues and packet loss in the San Francisco Bay Area, however other metros have been perfectly stable. On Wed, Nov 12, 2014 at 1:16 PM, james jones ja...@freedomnet.co.nz wrote: I am current going through some vendor selection for tier 1 providers. I was trying get some opinions on Zayo. I have personally never heard of them. Thoughts?
Re: Keeping Track of Data Usage in GB Per Port
On Wed, Oct 15, 2014 at 12:38 PM, Colton Conor colton.co...@gmail.com wrote: So based on the response I have received so far it seems cable was a complicated example with service flows involved. What if we are talking about something simpler like keeping track of how much data flows in and out of a port on a switch in a given month? I know you can use SNMP, but I believe that polls in intervals and takes samples which isn't really accurate right? It depends on what you're talking about. Network devices implementing the SNMP IF-MIB have counters for each interface that when polled, show the number of bytes being transmitted and received. Conventionally, network operators poll these counter values, compute the difference from the last time it was polled, and extrapolate a rate (bit volume over a time unit) from that. Often, this is done over a 5 minute interval. This introduces some averaging error. However, if an operator is just computing cumulative transfer, it's pretty easy. Just continue to sum up the counter value deltas from poll to poll. It could be easy to mess this up if the counter size is too small, or rolls more than once in-between polls. If a large telecom can't get billing correct, they shouldn't be allowed to do business. Easier solution: stop metering customers, and sink more money into expanded infrastructure.
Re: BGP Session
An Anycasting node. For example, as part of a reliable DNS service. A /24 is usually the smallest prefix length that is portably accepted. Also, applications where connections need to appear to be coming from many source IPs. On Saturday, July 19, 2014, Suresh Ramasubramanian ops.li...@gmail.com wrote: A single linux box with a whole /24 on it? What sort of use case is that, BTW? On 19-Jul-2014 10:26 pm, Abuse Contact stopabuseandrep...@gmail.com javascript:; wrote: I know, the DC is going to be giving me a BGP session on their router so I can set it up, I'm not using a Linux server as a router. On Sat, Jul 19, 2014 at 9:04 AM, William Herrin b...@herrin.us javascript:; wrote: On Wed, Jul 16, 2014 at 4:05 AM, Abuse Contact stopabuseandrep...@gmail.com javascript:; wrote: So I just purchased a Dedicated server from this one company and I have a /24 IPv4 block that I bought from a company on WebHostingTalk, but I am clueless on how to setup the /24 IPv4 block using the BGP Session. I want to set it up to run through their network as if it was one of their IPs, etc. I keep seeing things like iBGP (which I think means like a inner routing BGP) and eBGP (what I'm talking about??) but I have no idea how to set those up or which one I would need. Howdy, Unless you have (1) a real router available, not a just a server and (2) an expert available to help you with your first BGP configuration I strongly recommend you simply ask your service provider to announce the /24 to the Internet on your behalf. Server-based BGP software like Quagga for Linux is reasonably good but it should absolutely not be involved in your _first_ attempt to connect with the Internet's default-free zone. Simple mistakes with eBGP can cause tremendous damage to other folks on the Internet. Trial and error is simply not OK. If it isn't worth it to you to buy a BGP-capable router then you also aren't prepared to make the investment in learning it takes to use BGP without causing harm. Regards, Bill Herrin -- William Herrin her...@dirtside.com javascript:; b...@herrin.us javascript:; Owner, Dirtside Systems . Web: http://www.dirtside.com/ Can I solve your unusual networking challenges?
Re: BGP Session
On Sat, Jul 19, 2014 at 10:12 AM, Abuse Contact stopabuseandrep...@gmail.com wrote: Yeah, we're using it for an anycasted node but like, I'm confused on certain parts like, just a really basic question. When doing things like conf t router bgp AS1337 neighbor 208.54.128.0 remote-as AS13335 neighbor 208.54.128.0 description BGP with Upstream neighbor 208.54.128.0 password lolpass address-family ipv4 no synchronization neighbor 208.54.128.0 activate neighbor 208.54.128.0 soft-reconfiguration inboung I'm confused on when doing this, would I need to state like First go to AS13335 then go to TATA then go to my server or would it just automatically do that or would my provider do that? I'm confused on that. how would I state multiple peers.? AS13335 is Cloudflare. How does TATA relate? You have a deicated server connected to TATA and Cloudflare? I'm skeptical. You really ought to do some more reading, learning, and practicing before running public BGP. I would recommend reading this book cover-to-cover: http://www.bgpexpert.com/'BGP'-by-Iljitsch-van-Beijnum/ It's only ~250 small pages. To practice and experiment, emulate some example configurations with GNS3 and Dynamips, or some Linux VMs with Quagga or BIRD. On Sat, Jul 19, 2014 at 10:06 AM, Jonathan Lassoff j...@thejof.com wrote: An Anycasting node. For example, as part of a reliable DNS service. A /24 is usually the smallest prefix length that is portably accepted. Also, applications where connections need to appear to be coming from many source IPs. On Saturday, July 19, 2014, Suresh Ramasubramanian ops.li...@gmail.com wrote: A single linux box with a whole /24 on it? What sort of use case is that, BTW? On 19-Jul-2014 10:26 pm, Abuse Contact stopabuseandrep...@gmail.com wrote: I know, the DC is going to be giving me a BGP session on their router so I can set it up, I'm not using a Linux server as a router. On Sat, Jul 19, 2014 at 9:04 AM, William Herrin b...@herrin.us wrote: On Wed, Jul 16, 2014 at 4:05 AM, Abuse Contact stopabuseandrep...@gmail.com wrote: So I just purchased a Dedicated server from this one company and I have a /24 IPv4 block that I bought from a company on WebHostingTalk, but I am clueless on how to setup the /24 IPv4 block using the BGP Session. I want to set it up to run through their network as if it was one of their IPs, etc. I keep seeing things like iBGP (which I think means like a inner routing BGP) and eBGP (what I'm talking about??) but I have no idea how to set those up or which one I would need. Howdy, Unless you have (1) a real router available, not a just a server and (2) an expert available to help you with your first BGP configuration I strongly recommend you simply ask your service provider to announce the /24 to the Internet on your behalf. Server-based BGP software like Quagga for Linux is reasonably good but it should absolutely not be involved in your _first_ attempt to connect with the Internet's default-free zone. Simple mistakes with eBGP can cause tremendous damage to other folks on the Internet. Trial and error is simply not OK. If it isn't worth it to you to buy a BGP-capable router then you also aren't prepared to make the investment in learning it takes to use BGP without causing harm. Regards, Bill Herrin -- William Herrin her...@dirtside.com b...@herrin.us Owner, Dirtside Systems . Web: http://www.dirtside.com/ Can I solve your unusual networking challenges?
Re: BGP Session
Wow -- be careful playing with public eBGP sessions unless you know what you're doing. It can affect the entire Internet. Since you're just connecting to a single upstream ISP, you wont qualify for a public AS number. So, you'll have to work with your upstream ISP to agree on a private AS number you can use. You will be setting up an eBGP session (which is a session between two different AS numbers, as opposed to iBGP, wherein the AS numbers are the same). As for running BGP on a dedicated server, it'll depend on the OS in use. Assuming Linux, take a look at Quagga, BIRD, and ExaBGP. http://www.nongnu.org/quagga/ http://bird.network.cz/ https://code.google.com/p/exabgp/ It may be a *lot* easier for you to just have your upstream ISP announce your IP space, and route it to your dedicated server, unless you need the ability to turn it off and on over time. Cheers, jof On Wed, Jul 16, 2014 at 1:05 AM, Abuse Contact stopabuseandrep...@gmail.com wrote: Hi, So I just purchased a Dedicated server from this one company and I have a /24 IPv4 block that I bought from a company on WebHostingTalk, but I am clueless on how to setup the /24 IPv4 block using the BGP Session. I want to set it up to run through their network as if it was one of their IPs, etc. I keep seeing things like iBGP (which I think means like a inner routing BGP) and eBGP (what I'm talking about??) but I have no idea how to set those up or which one I would need. Any help would be appreciated. Thanks!
Re: MACsec SFP
On Tue, Jun 24, 2014 at 12:59 AM, Pieter Hulshoff phuls...@aimvalley.nl wrote: On 24-6-2014 8:37, Saku Ytti wrote: On (2014-06-23 11:13 +0200), Pieter Hulshoff wrote: feature and market information for such a device, and I would welcome some feedback from interested people. Discussion about other types of smart SFPs would also be welcome. Feel free to contact me directly using the contact information below. I'd do questionable things for subrate SFP, SFP which I can put to 1GE port and have 10M and 100M rates available. Or to 10GE port and get 1GE, 100M and 10M Use case is network generation upgrade where you still have one or two 100M ports for MGMT ports etc. I've seen this request from others as well. Do you have any proposal/preference to limit the data rate from the switch? Seems like it would be just like emulating a media convertor. Drop any frames in excess of 100 Mbit? Perhaps buffer a little bit? If using the interface for any protocols, configuration might need to be made to adjust link costs.
Re: Odd syslog-ng problem
Peter, it's a bit difficult to tell what's going on without seeing the rest of the syslog-ng configuration and your script's source code. However, a couple possibilities come to mind: - Your script is only reading one line at a time. syslog-ng starts a program() output persistently and expects that it can send multiple messages into its pipe to your script's stdin. - Messages are being buffered inside of syslog-ng. Check out the flush_lines() and flush_timeout() flags to syslog-ng's program() output. Find the right page for your version, but here's v3.3.: http://www.balabit.com/sites/default/files/documents/syslog-ng-ose-3.3-guides/en/syslog-ng-ose-v3.3-guide-admin-en/html/reference_destination_program.html - Messages are being buffered in your shell or script. Maybe try some non-blocking IO with a smallish buffer to see data as it comes in before a whole line or block fills and flushes in. To Anurag's question about open source log management with a WebUI, I agree with Blake: logstash ingesting syslog and inputting it into elasticsearch makes for a great backend for Kibana. The logstash grok filter is great for pulling apart and indexing weird vendor-specific logging formats: http://logstash.net/docs/1.4.1/filters/grok Cheers, jof On Sat, May 10, 2014 at 2:24 AM, Peter Persson web...@webbax.se wrote: Hey, I got a weird problem with my syslog-ng setup, im logging from alot of cisco machines and that works great. The problem is that when i pass this further to a shell program, some lines disapere. My destination looks like this destination hosts { file(/var/log/ciscorouters/$HOST.log owner(root) group(root) perm(0600) dir_perm(0700) create_dirs(yes)); program(/scripts/irc/syslog_wrapper_new.sh template(t_irctempl)); }; The /var/log/ciscorouters/$HOST.log writes correct, but the data thats putted trough to /scripts/irc/syslog_wrapper_new.sh only get the first line, if it gets flooded (like 5 rows per second). Do anyone of you have any idea of what might be the problem? Regards, Peter
Re: Fwd: Serious bug in ubiquitous OpenSSL library: Heartbleed
For testing, I've had good luck with https://github.com/titanous/heartbleeder and https://gist.github.com/takeshixx/10107280 Both are mostly platform-independent, so they should be able to work even if you don't have a modern OpenSSL to test with. Cheers and good luck (you're going to need it), jof On Tue, Apr 8, 2014 at 5:03 PM, Michael Thomas m...@mtcc.com wrote: Just as a data point, I checked the servers I run and it's a good thing I didn't reflexively update them first. On Centos 6.0, the default openssl is 1.0.0 which supposedly doesn't have the vulnerability, but the ones queued up for update do. I assume that redhat will get the patched version soon but be careful! Mike On 04/07/2014 10:06 PM, Paul Ferguson wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 I'm really surprised no one has mentioned this here yet... FYI, - - ferg Begin forwarded message: From: Rich Kulawiec r...@gsp.org Subject: Serious bug in ubiquitous OpenSSL library: Heartbleed Date: April 7, 2014 at 9:27:40 PM EDT This reaches across many versions of Linux and BSD and, I'd presume, into some versions of operating systems based on them. OpenSSL is used in web servers, mail servers, VPNs, and many other places. Writeup: Heartbleed: Serious OpenSSL zero day vulnerability revealed http://www.zdnet.com/heartbleed-serious-openssl-zero-day-vulnerability- revealed-728166/ Technical details: Heartbleed Bug http://heartbleed.com/ OpenSSL versions affected (from link just above): OpenSSL 1.0.1 through 1.0.1f (inclusive) are vulnerable OpenSSL 1.0.1g is NOT vulnerable (released today, April 7, 2014) OpenSSL 1.0.0 branch is NOT vulnerable OpenSSL 0.9.8 branch is NOT vulnerable - -- Paul Ferguson VP Threat Intelligence, IID PGP Public Key ID: 0x54DC85B2 -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.22 (MingW32) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iF4EAREIAAYFAlNDg9gACgkQKJasdVTchbIrAAD9HzKaElH1Tk0oIomAOoSOvfJf 3Dvt4QB54os4/yewQQ8A/0dhFZ/YuEdA81dkNfR9KIf1ZF72CyslSPxPvkDcTz5e =aAzE -END PGP SIGNATURE-
Re: Blocking of domain strings in iptables
This is going to be tricky to do, as DNS packets don't necessarily contain entire query values or FQDNs as complete strings due to packet label compression (remember, original DNS only has 512 bytes to work with). You can use those u32 module matches to find some known-bad packets if they're sufficiently unique, but it simply lacks enough logic to fully parse DNS queries. Here's an interesting example to visualize what's happening: http://dnsamplificationattacks.blogspot.com/p/iptables-block-list.html One quick thing that would work would be to match a single label (e.g. google, but not google.com), but this will end up blocking any frames with that substring in it (e.g. you want to block evil.com, but this also blocks evil.example.com). If you find yourself needing to parse and block DNS packets based on their content in a more flexible way, I would look into either making an iptables module that does the DNS parsing ( http://inai.de/documents/Netfilter_Modules.pdf), or using a userspace library like with NFQUEUE (e.g. https://pypi.python.org/pypi/NetfilterQueue) or l7-filter (http://l7-filter.sourceforge.net/). Best of luck and happy hacking! Cheers, jof On Sat, Feb 8, 2014 at 12:08 AM, Anurag Bhatia m...@anuragbhatia.com wrote: Hello everyone I am trying to figure out the way to drop a domain name DNS resolution before it hits application server. I do not want to do domain to IP mapping and block destination IP (and source IP blocking is also not an option). I can see that a string like this: iptables -A INPUT -p udp -m udp --dport 53 -m string --string domain --algo kmp --to 65535 -j DROP this can block domain which includes domain.com/domain.net and everything in that pattern. I tried using hexadecimal string for value like domaincom (hexa equivalent) and firewall doesn't pics that at all. The only other option which I found to be working nicely is u32 based string as something suggested on DNS amplification blog post here - http://dnsamplificationattacks.blogspot.in/2013/12/domain-dnsamplificationattackscc.html A string like this as suggested on above link works exactly for that domain iptables --insert INPUT -p udp --dport 53 -m u32 --u32 0x280xFFDFDFDF=0x17444e53 0x2c0xDFDFDFDF=0x414d504c 0x300xDFDFDFDF=0x49464943 0x340xDFDFDFDF=0x4154494f 0x380xDFDFDFDF=0x4e415454 0x3c0xDFDFDFDF=0x41434b53 0x400xFFDFDFFF=0x02434300 -j DROP -m comment --comment DROP DNS Q dnsamplificationattacks.cc but here I am not sure how to create such string out and script them for automation. Can someone suggest a way out for this within IPTables or may be some other open source firewall? Thanks. -- Anurag Bhatia anuragbhatia.com Linkedin http://in.linkedin.com/in/anuragbhatia21 | Twitterhttps://twitter.com/anurag_bhatia Skype: anuragbhatia.com PGP Key Fingerprint: 3115 677D 2E94 B696 651B 870C C06D D524 245E 58E2
Re: GEO location issue with google
Here's the FAQ on this topic: https://support.google.com/websearch/answer/873?hl=en It links to a contact form where you can ask for some redress. Cheers, jof On Fri, Feb 7, 2014 at 7:20 AM, Praveen Unnikrishnan p...@pmgroupuk.comwrote: Hi, We are an ISP based in UK. We have got an ip block from RIPE which is 5.250.176.0/20. All the main search engines like yahoo shows we are based in UK. But Google thinks we are from Saudi Arabia and we redirected to www.google.com.sahttp://www.google.com.sa instead of googlw.co.uk. I have sent lot of emails to google but no luck. All the information from google are in Arabic and youtube shows some weird videos as well. Could anyone please help me to sort this out? Would be much appreciated for your time. Praveen Unnikrishnan Network Engineer PMGC Technology Group Ltd T: 020 3542 6401 M: 07827921390 F: 087 1813 1467 E: p...@pmgroupuk.commailto:p...@pmgroupuk.com [cid:image001.png@01CF2418.27F29CA0] [cid:image002.jpg@01CE1663.96B300D0] www.pmgroupuk.comhttp://www.pmgroupuk.com/ | www.pmgchosting.com http://www.pmgchosting.com/ How am I doing? Contact my manager, click heremailto:sha...@pmgroupuk.com ?subject=How's%20Praveen%20doing?. [cid:image003.jpg@01CE1663.96B300D0] PMGC Managed Hosting is now live! After a successful 2012, PMGC continues to innovate and develop by offering tailored IT solutions designed to empower you through intelligent use of technologies. www.pmgchosting.com http://www.pmgchosting.com/. PMGC Technology Group Limited is a company registered in England and Wales. Registered number: 7974624 (3/F Sutherland House, 5-6 Argyll Street, London. W1F 7TE). This message contains confidential (and potentially legally privileged) information solely for its intended recipients. Others may not distribute copy or use it. If you have received this communication in error please contact the sender as soon as possible and delete the email and any attachments without keeping copies. Any views or opinions presented are solely those of the author and do not necessarily represent those of the company or its associated companies unless otherwise specifically stated. All incoming and outgoing e-mails may be monitored in line with current legislation. It is the responsibility of the recipient to ensure that emails are virus free before opening. PMGC(r) is a registered trademark of PMGC Technology Group Ltd.
Re: The state of TACACS+
I don't understand why vendors and operators keep turning to TACACS. It seems like they're often looking to Cisco as some paragon of best security practices. It's a vulnerable protocol, but some times the only thing to choose from. One approach to secure devices that can support only TACACS or RADIUS: Deploy a small embedded *nix machine (Soekris, Raspberry Pi, etc.) that runs a RADSEC (for RADIUS) or stunnel (for TACACS) proxy. Attach it to a short copper with 802.1q, take weak xor'ed requests in on one tag, wrap the requests with TLS, and forward out another tag towards your central AAA box. Kerberos or more certificate-based SSH on routers would be super. SSH with certificates is nice in that it allows authenticators out in the field to verify clients offline, without needing a central AAA server. However, the tradeoff is that you must then make sure all the clocks are correct and in-sync, and root certificates are verified. On Mon, Dec 30, 2013 at 2:06 AM, Robert Drake rdr...@direcpath.com wrote: Ever since first using it I've always liked tacacs+. Having said that I've grown to dislike some things about it recently. I guess, there have always been problems but I've been willing to leave them alone. I don't have time to give the code a real deep inspection, so I'm interested in others thoughts about it. I suspect people have just left it alone because it works. Also I apologize if this is too verbose or technical, or not technical enough, or just hard to read. History: TACACS+ was proposed as a standard to the IETF. They never adopted it and let the standards draft expire in 1998. Since then there have been no official changes to the code. Much has happened between now and then. I specifically was interested in parsing tac_plus logs correctly. After finding idiosyncrasies I decided to look at the source and the RFC to see what was really happening. Logging, or why I got into this mess: In the accounting log, fields are sometimes logged in different order. It appears the client is logging whatever it receives without parsing it or modifying it. That means the remote system is sending them in different orders, so technically the fault lies with them. However, it seems too trusting to take in data and log it without looking at it. This can also cause issues when you send a command like (Cisco) dir /all nvram: on a box with many files. The device expands the command to include everything on the nvram (important because you might want to deny access to that command based on something it expanded), but it gets truncated somewhere (not sure if it's the device buffer that is full, tac_plus, or the logging part. I might tcpdump for a while to see if I can figure out what it looks like on the wire) I'm not sure if there are security implications there. Encryption: The existing security consists of md5 XOR content with the md5 being composed of a running series of 16 byte hashes, taking the previous hash as part of the seed of the next hash. A sequence number is used so simple replay shouldn't be a factor. Depending on how vulnerable iterative md5 is to it, and how much time you had to sniff the traffic, I would think this would be highly vulnerable to chosen plaintext if you already have a user-level login, or at least partial known plaintext (with the assumption they make backups, you can guess that at least some of the packets will have show running-config and other common commands). They also don't pad the encrypted string so you can guess the command (or password) based on the length of the encrypted data. For a better description of the encryption you can read the draft: http://tools.ietf.org/html/draft-grant-tacacs-02 I found an article from May, 2000 which shows that the encryption scheme chosen was insufficient even then. http://www.openwall.com/articles/TACACS+-Protocol-Security For new crypto I would advise multiple cipher support with negotiation so you know what each client and server is capable of. If the client and server supported multiple keys (with a keyid) it would be easier to roll keys frequently, or if it isn't too much overhead they could use public key. Clients: As for clients, Wikipedia lists several that seem to be based on the original open-source tac_plus from Cisco. shrubbery.net has the official version that debian and freebsd use. I looked at some of the others and they all seemed to derive from Cisco's code directly or shrubbery.net code, but they retained the name and started doing their own versioning. All the webpages look like they're from 1995. In some cases I think it's intentional but in some ways it shows a lack of care for the code, like it's been dropped since 2000. Documentation is old: This only applies to shrubbery.net's version. I didn't look at the other ones that closely. While all of it appears valid, one QA in the FAQ was about IOS 10.3/11.0. Performance
Re: WaPo writes about vulnerabilities in Supermicro IPMIs
The primary point of IPMI for most users is to be able to administer and control the box when it's not running. Using the host itself as a firewall is the quickest way to get that BMC online, but it kinda defeats the purpose. On Thu, Aug 15, 2013 at 7:46 PM, Jay Ashworth j...@baylink.com wrote: - Original Message - From: Brandon Martin lists.na...@monmotha.net As to why people wouldn't put them behind dedicated firewalls, imagine something like a single-server colo scenario. Most such providers don't offer any form of lights-out management aside from maybe remote reboot (power-cycle) nor do they offer any form of protected/secondary network to their customers. So, if you want to save yourself from a trip, you chuck the thing raw on a public IP and hope you configured it right. Well, *I* would firewall eth1 from eth0 and cross-over eth1 to the ILO jack; let the box be the firewall. Sure, it's still as breakable as the box proper, but security-by-obscurity isn't *bad*, it's just *not good enough*. It's another layer of tape. Whether it's teflon or Gorilla is up to you. Cheers, -- jra -- Jay R. Ashworth Baylink j...@baylink.com Designer The Things I Think RFC 2100 Ashworth Associates http://baylink.pitas.com 2000 Land Rover DII St Petersburg FL USA #natog +1 727 647 1274
Re: Blocking TCP flows?
Are you trying to block flows from becoming established, knowing what you're looking for ahead of time, or are you looking to examine a stream of flow establishments, and will snipe off some flows once you've determined that they should be blocked? If you know a 5-tuple (src/dst IP, IP protocol, src/dst L4 ports) you want to block ahead of time, just place an ACL. It depends on the platform, but those that implement them in hardware can filter a lot of traffic very quickly. However, they're not a great tool when you want to dynamically reconfigure the rules. For high-touch inspection, I'd recommend a stripe of Linux boxes, with traffic being ECMP-balanced across all of them, sitting in-line on the traffic path. It adds a tiny bit of latency, but can scale up to process large traffic paths and apply complex inspections on the traffic. Cheers, jof On Thu, Jun 13, 2013 at 12:32 PM, Eric Wustrow ew...@umich.edu wrote: Hi all, I'm looking for a way to block individual TCP flows (5-tuple) on a 1-10 gbps link, with new blocked flows being dropped within a millisecond or so of being added. I've been looking into using OpenFlow on an HP Procurve, but I don't know much in this area, so I'm looking for better alternatives. Ideally, such a device would add minimal latency (many/expandable CAM entries?), can handle many programatically added flows (hundreds per second), and would be deployable in a production network (fails in bypass mode). Are there any COTS devices I should be looking at? Or is the market for this all under the table to pro-censorship governments? Thanks, -Eric
Re: Blocking TCP flows?
On Thu, Jun 13, 2013 at 3:38 PM, Phil Fagan philfa...@gmail.com wrote: I would assume something FreeBSD based might be best Meh... personal choice. I prefer Linux, mostly because I know it best and most network application development is taking place there. On Thu, Jun 13, 2013 at 4:37 PM, Phil Fagan philfa...@gmail.com wrote: I really like the idea of a stripe of linux boxes doing the heavy lifting. Any suggestions on platforms, card types, and chip types that might be better purposed at processing this type of data? Personally, I'd use modern-ish Intel Ethernet NICs. They seem to have the best support in the kernel. I assume you could write some fast Perl to ingest and manage the tables? What would the package of choice be for something like this? Heh... fast Perl. As for programming the processing, I would do as much as possible in the kernel, as passing packets off to userland really slows everything down. If you really need to, I'd do something with Go and/or C these days. Using iptables and the string module to match patterns, you can chew through packets pretty efficiently. This comes with the caveat that this can only match against strings contained within a single packet; this doesn't do L4 stream reconstruction. You can do some incredibly-parallel stuff with ntop's PF_RING code, if you blow more traffic through a single core than it can chew through. It all depends on what you're trying to do. --j On Thu, Jun 13, 2013 at 3:11 PM, Jonathan Lassoff j...@thejof.com wrote: Are you trying to block flows from becoming established, knowing what you're looking for ahead of time, or are you looking to examine a stream of flow establishments, and will snipe off some flows once you've determined that they should be blocked? If you know a 5-tuple (src/dst IP, IP protocol, src/dst L4 ports) you want to block ahead of time, just place an ACL. It depends on the platform, but those that implement them in hardware can filter a lot of traffic very quickly. However, they're not a great tool when you want to dynamically reconfigure the rules. For high-touch inspection, I'd recommend a stripe of Linux boxes, with traffic being ECMP-balanced across all of them, sitting in-line on the traffic path. It adds a tiny bit of latency, but can scale up to process large traffic paths and apply complex inspections on the traffic. Cheers, jof On Thu, Jun 13, 2013 at 12:32 PM, Eric Wustrow ew...@umich.edu wrote: Hi all, I'm looking for a way to block individual TCP flows (5-tuple) on a 1-10 gbps link, with new blocked flows being dropped within a millisecond or so of being added. I've been looking into using OpenFlow on an HP Procurve, but I don't know much in this area, so I'm looking for better alternatives. Ideally, such a device would add minimal latency (many/expandable CAM entries?), can handle many programatically added flows (hundreds per second), and would be deployable in a production network (fails in bypass mode). Are there any COTS devices I should be looking at? Or is the market for this all under the table to pro-censorship governments? Thanks, -Eric -- Phil Fagan Denver, CO 970-480-7618 -- Phil Fagan Denver, CO 970-480-7618
Re: Prism continued
Logstash and Splunk are both wonderful, in my experience. What sets them apart from just a plain grep(1) is that they build an index that points keywords to to logging events (lines). What if you're looking for events related to a specific interface or LSP? Not a problem with a modest log volume, as grep can tear through text nearly as quickly as your disk can pass it up. However, once you have a ton of historical logs, or just a large volume, grep becomes way to slow as you have to retrieve tons of unrelated log messages to check if they're what you're looking for. Having an index gives you a way to search for that interface or LSP name, and get a listing of all the locations that contain log events matching what you're looking for. In the PRISM context, I highly doubt their using Splunk for any kind of analysis beyond systems and network management. It's not good at indexing non-texty-things. What if you need to search for events that were geographically proximate to one another? That takes a special kind of index. On Wed, Jun 12, 2013 at 6:13 PM, Chip Marshall c...@2bithacker.net wrote: On 2013-06-12, Phil Fagan philfa...@gmail.com sent: Speaking of Splunk; is that really the tool of choice? I've been hearing a lot of good things about logstash these days too, if you prefer the open source route. http://logstash.net/ -- Chip Marshall c...@2bithacker.net http://2bithacker.net/
Re: PRISM: NSA/FBI Internet data mining project
Agreed. I can already pretty much just assume this widespread surveillance is going on. The Bluffdale, Utah facility isn't being built to store nothing. It's happening whether we like it or not. When I care about my privacy, I know that I have to take matters into my own hands. GnuPG and TLS are mine and your friends. Use them together. Use them in peace. Cheers, jof (0x8F8CAD3D) On Thu, Jun 6, 2013 at 5:07 PM, Alex Rubenstein a...@corp.nac.net wrote: Has fingers directly in servers of top Internet content companies, dates to 2007. Happily, none of the companies listed are transport networks: I've always just assumed that if it's in electronic form, someone else is either reading it now, has already read it, or will read it as soon as I walk away from the screen. So, you are comfortable just giving up your right to privacy? It's just the way it is? I'm sorry, I am not as accepting of that fact as you are. I am disappointed and disgusted that this is, and has been, going on. Our government is failing us.
Re: Cat-5 cables near 200 Paul, SF
I could suggest a few places. Might want to call ahead to make sure they'll have what you need: - Central Computer. Has locations in San Francisco and San Mateo. SF maybe closer, but will take longer with traffic and parking. -- http://www.centralcomputers.com/commerce/misc/sanfrancisco.jsp -- http://www.centralcomputers.com/commerce/misc/sanmateo.jsp - Frys. Much further. Closest shop would be Palo Alto -- http://www.frys.com/template/isp/index/Frys/isp/Middle_Topics/H1%20Store%20Maps/palo%20alto/ - Jameco. Has some limited Ethernet cable selection. Has a will-call pick up at their warehouse in Belmont. -- http://www.jameco.com/ Cheers, jof
Re: Headscratcher of the week
Those are some truly perplexing graphs. Quite strange that it appears linear, as if something is slightly changing over time or growing/shrinking at a constant-ish rate. Do you have throughput or PPS graphs for the intermediate links as well? Any similar correlations in the derivative slope? My only hunch would be some intermediate buffer being increasingly full over time, as some other application riding the path linearly grows in packets/second or bits/second. Cheers, jof On Fri, May 31, 2013 at 3:25 PM, Mike mike-na...@tiedyenetworks.com wrote: Gang, In the interest of sharing 'the weird stuff' which makes the job of being an operator ... uh, fun? is that the right word?..., I would like to present the following two smokeping latency/packetloss plots, which are by far the weirdest I have ever seen. These plots are from our smokeping host out to a customer location. The customer is connected via DSL and they run PPPoE over it to connect with our access concentrator. There is about 5 physical insfastructure hops between the host and customer; The switch, the BRAS, the Switch again, and then directly to the DSLAM and then customer on the end. The 10 day plot: http://picpaste.com/10_Day_graph-YV3IdvRV.png The 30 hour plot: http://picpaste.com/30_hour_graph-DrwzfhYJ.png How can you possibly have consistent increase in latency like that? I'd love to hear theories (or offers of beer, your choice!). Happy friday all! Mike-
Re: need help about free bandwidth graph program
I'm not sure of your specific application, but it sounds to me like netflow/sflow exports would be the most scalable way to do this. For small applications, ntop or bandwidthd can do this. http://www.ntop.org/products/ntop/ http://bandwidthd.sourceforge.net/ Cheers, jof On Mon, Apr 8, 2013 at 12:51 PM, Deric Kwok deric.kwok2...@gmail.com wrote: Hi all Do you know any opensource program bandwidthgraph by ipaddess? Thank you
Re: BGP RIB Collection
Personally, I would just use BGP on a PC to collect this information. Place some import/input policy on your eBGP sessions on your edge routers to add communities to the routes such that you can recognize which peers gave you the route. Then, use an iBGP session to a BIRD or Quagga instance from which you can dump the routes and filter based on the communities. Cheers, jof On Tue, Feb 26, 2013 at 6:24 PM, chip chip.g...@gmail.com wrote: Hello all, I have an application that needs to gather BGP RIB data from the routers that connect to all of our upstream providers. Basically I need to know all the routes available from a particular provider. Currently I'm gathering this data via SNMP. While this works it has its draw backs, it takes approximately 20 minutes per view, its nowhere near real-time, and I'm unable to gather information for IPv6. SNMP, however, is faster than screen scraping. All of the XML based access methods seem to take about the same time as well. I've been watching, with keen interest, the i2rs ietf workings, but the project is still in its infancy. BMP seems to be a good solution but I've not found a working client implementation yet. I see that you can actually configure this on some Juniper gear but I can't seem to locate a client to ingest the data the router produces. The BGP Add Paths implementation seems to be the best choice at the moment and exabgp has a working implementation. Are there any other technologies or methods of accessing this data that I've missed or that you've found useful? Thanks! --chip -- Just my $.02, your mileage may vary, batteries not included, etc
Re: Micro Trenching for Fiber Optic Deployment
I would think that in such a deployment scenario, microtrenching might not be the best bet. Part of the appeal (IMO) of microtrenching in existing pavement is that once filled, the pavement slab provides for some protection and rigidity. If making a small trench into packed dirt, you're much more susceptible to accidental cuts and erosion. I would suggest a Ditch Witch or similar trencher and laying some ABS/PVC conduit could give some protection from small nicks and cuts, and allow for future strands to be pulled along your path. Cheers, jof On Mon, Feb 11, 2013 at 10:34 AM, david peahi davidpe...@gmail.com wrote: Does anyone have experience in running fiber optic cable with micro-trenching techniques in areas where there is no existing asphalt or concrete roadway, just packed earth and rock? Environmental limitations do not allow for constructing an aerial power pole alignment, or underground ductbank. The distance is about 10 kM. David
Re: L3 East cost maint / fiber 05FEB2012 maintenance
My hunch is that this is fallout and repairs from Juniper PR839412. Only fix is an upgrade. Not sure why they're not able to do a hitless upgrade though; that's unfortunate. Specially-crafted TCP packets that can get past RE/loopback filters can crash the box. --j On Tue, Feb 5, 2013 at 7:39 AM, Josh Reynolds ess...@gmail.com wrote: I know a lot of you are out of the office right now, but does anybody have any info on what happened with L3 this morning? They went into a 5 hour maintenance window with expected downtime of about 30 minutes while they upgraded something like *40* of their core routers (their words), but also did this during some fiber work and completely cut off several of their east coast peers for the entirety of the 5 hour window. If anybody has any more info on this, on a NOC contact for them on the East Coast for future issues, you can hit me off off-list if you don't feel comfortable replying with that info here. Thanks, and I hope hope you guys are enjoying Orlando. -- *Josh Reynolds* ess...@gmail.com - (270) 302-3552
Re: L3 East cost maint / fiber 05FEB2012 maintenance
On Tue, Feb 5, 2013 at 9:33 AM, Jason Biel ja...@biel-tech.com wrote: Workaround is proper filtering and other techniques on the RE/Loopback to prevent the issue from happening. Agreed. However, if it only takes one packet, what if an attacker sources the traffic from your management address space? Guarding against this requires either a separate VRF/table for management traffic or transit traffic, RPF checking, or TTL security. If these weren't setup ahead of time, maybe it would be easier to upgrade than lab, test, and deploy a new configuration. This is all speculation about Level3 on my part; I don't know their network from an internal perspective. --j Should an upgrade be performed? Yes, but certainly doesn't have to have right away or without notice to customers. On Tue, Feb 5, 2013 at 11:23 AM, Jonathan Lassoff j...@thejof.com wrote: My hunch is that this is fallout and repairs from Juniper PR839412. Only fix is an upgrade. Not sure why they're not able to do a hitless upgrade though; that's unfortunate. Specially-crafted TCP packets that can get past RE/loopback filters can crash the box. --j On Tue, Feb 5, 2013 at 7:39 AM, Josh Reynolds ess...@gmail.com wrote: I know a lot of you are out of the office right now, but does anybody have any info on what happened with L3 this morning? They went into a 5 hour maintenance window with expected downtime of about 30 minutes while they upgraded something like *40* of their core routers (their words), but also did this during some fiber work and completely cut off several of their east coast peers for the entirety of the 5 hour window. If anybody has any more info on this, on a NOC contact for them on the East Coast for future issues, you can hit me off off-list if you don't feel comfortable replying with that info here. Thanks, and I hope hope you guys are enjoying Orlando. -- *Josh Reynolds* ess...@gmail.com - (270) 302-3552 -- Jason
Re: ATT Uverse/DSL Network Engineer DNS question
These appear to be an anycasted service, as I reach different destinations based on my source address. Hopefully each deployment has unique origin IPs for their recursive queries. I would recommend against looking at RIR registration data to determine IP location. There's often little to no correlation, there. --j On Tue, Feb 5, 2013 at 1:01 PM, Tim Haak thaiti...@hotmail.com wrote: Hi, Can a ATT Uverse/DSL Network Engineer answer a question about the DNS server IPs that are handed out to customers please? I am currently testing from a Florida IP. Can you please let me know if all Uverse and DSL customers across the United States only use these 2 IPs as their primary and secondary DNS servers? 68.94.156.1 68.94.157.1 We provide services based on IP GEO-location. Since the 2 recursive resolvers below are registered in Texas every DNS query for any of our records return results that are intended for IPs in that region. In other words, users on the east coast would actually resolve to a central part of the US or west coast IP. Thanks in advance,Tim
Re: ATT Uverse/DSL Network Engineer DNS question
On Tue, Feb 5, 2013 at 1:10 PM, Jonathan Lassoff j...@thejof.com wrote: These appear to be an anycasted service, as I reach different destinations based on my source address. Hopefully each deployment has unique origin IPs for their recursive queries. Just confirmed this. As these resolvers traverse and query your servers, they'll have different source IPs, depending on the regional resolver. Return differentiated DNS responses, based on that. --j I would recommend against looking at RIR registration data to determine IP location. There's often little to no correlation, there. --j On Tue, Feb 5, 2013 at 1:01 PM, Tim Haak thaiti...@hotmail.com wrote: Hi, Can a ATT Uverse/DSL Network Engineer answer a question about the DNS server IPs that are handed out to customers please? I am currently testing from a Florida IP. Can you please let me know if all Uverse and DSL customers across the United States only use these 2 IPs as their primary and secondary DNS servers? 68.94.156.1 68.94.157.1 We provide services based on IP GEO-location. Since the 2 recursive resolvers below are registered in Texas every DNS query for any of our records return results that are intended for IPs in that region. In other words, users on the east coast would actually resolve to a central part of the US or west coast IP. Thanks in advance,Tim
Re: Whats so difficult about ISSU
On Thu, Nov 8, 2012 at 8:13 PM, Mikael Abrahamsson swm...@swm.pp.se wrote: On Thu, 8 Nov 2012, Phil wrote: The major vendors have figured it out for the most part by moving to stateful synchronization between control plane modules and implementing non-stop routing. NSR isn't ISSU. ISSU contains the wording in service. 6 seconds of outage isn't in service. 0.5 seconds of outage isn't in service. I could accept a few microseconds of outage as being ISSU, but tenths of seconds isn't in service. The main remaining hurdle is updating microcode on linecards, they still need to be rebooted after an upgrade. ... and as long as this is the case, there is no ISSU. There is only shorter outages during upgrade compared to a complete reboot. This. There are some wonderfully reconfigurable router hardwares out in the world, and platforms that can dynamically program their forwarding hardware make this seem possible. It's possible to build things such that portions of a single box can be upgraded at a time. With multiple links, or forwarding-paths out to a remote destination, it seems to me that if the upgrade process could just coordinate things and update each piece of forwarding hardware while letting traffic cut over and waiting for it to come back before moving on. I could envision a Juniper M/TX box, where MPLS FRR or an ae interface across FPCs could take backup traffic while a PFE is upgraded. Of course, every possible path would need to be able to survive an FPC being down, and the process would have to have hooks into protocols to know when everything is switched back.
Re: Detection of Rogue Access Points
On Sun, Oct 14, 2012 at 1:59 PM, Jonathan Rogers quantumf...@gmail.com wrote: Gentlemen, An issue has come up in my organization recently with rogue access points. So far it has manifested itself two ways: 1. A WAP that was set up specifically to be transparent and provided unprotected wireless access to our network. This is actually a really tough problem to solve without either total dictatorial control of your switchports or lots of telemetry and monitoring. At $DAYJOB, we detect the transparent bridge case by having a subset of AP hardware setup as monitors that listen to 802.11 frames on the various channels, keeping a log of the client MAC addresses and the BSSID that they're associated with. Then, by selecting out only those client MAC addresses that are not associated to a known BSSID that we control, we compare that set of unknown client MAC addresses to the Ethernet L2 FIBs on our switches and look for matches. If we see entries, than there is some 802.11 device bridging clients onto our network and we hunt it down from there. I've yet to see a solid methodology for detecting NATing devices, short of requiring 802.1x authentication using expiring keys and one-time passwords. :p Cheers, jof
Re: best way to create entropy?
On Thu, Oct 11, 2012 at 5:01 PM, shawn wilson ag4ve...@gmail.com wrote: in the past, i've done many different things to create entropy - encode videos, watch youtube, tcpdump -vvv /dev/null, compiled a kernel. but, what is best? just whatever gets your cpu to peak or are some tasks better than others? Personally, I've used and recommend this USB stick: http://www.entropykey.co.uk/ Internally, it uses diodes that are reverse-biased just ever so close to the breakdown voltage such that they randomly flip state back and forth. Cheers, jof
Re: best way to create entropy?
On Thu, Oct 11, 2012 at 5:20 PM, Jimmy Hess mysi...@gmail.com wrote: On 10/11/12, shawn wilson ag4ve...@gmail.com wrote: in the past, i've done many different things to create entropy - encode videos, watch youtube, tcpdump -vvv /dev/null, compiled a kernel. but, what is best? just whatever gets your cpu to peak or are You are referring to the entropy pool used for /dev/random and crypto operations ? You could setup a video capture card or radio tuner card, tune it into a good noise source, and arrange for the bit stream to get written to /dev/random Yes, but then you're also introducing a way for an external attacker to transmit data that can be mixed into your entropy pool. While certainly a cool hack, I don't think anything like this would be safe for cryptographic use. /two cents Cheers, jof
Re: dot1q encapsulation overhead?
On Thu, Sep 6, 2012 at 7:55 AM, u...@3.am wrote: A while back we had a customer colocated vpn router (2911) come in and we put it on our main vlan for initial set up and testing. Once that was done, I created a separate VLAN for them and a dot1q subinterface on an older, somewhat overloaded 2811. I set up the IPSec Tunnel, a /30 for each end to have an IP and all the static routes needed to make this work and it did. However, a few days later they were complaining of slow speeds...I don't recall, but maybe something like 5mbs when they needed 20 or so. We had no policing on that port. After a lot of testing, we tried putting them back on the main, native vlan and it worked fine...they got the throughput they needed. So my question is: could the dot1q encapsulation be causing throughput issues on a 2811 that's already doing a lot? I regret that I don't recall what sh proc cpu output was, or if I even ran it at all. It was kind of hectic just to get it fixed at the time. Well, a few months later (last week), the chicken came home to roost when their IPSec tunnel started proxy ARP puking stuff to our side that temporarily took out parts of our internal LAN. I have requested a 2911 replacement for the 2811 because I have seen the 2811 cpu load max out a few times when passing lots of traffic. I am hoping it will allow us to go back to this VLAN setup again, but I've never heard whether dot1q adds any overhead. It's small, but plain 802.1Q adds in a 4-byte (32-bit) header after the normal Ethernet header and before the actual Ethernet payload. That tag has to go somewhere. :p Also, I'm not privy to the details of your IPSec Tunnel, but that can introduce additional overhead as well. I'm not sure about your specific 2811/2911 and IOS combination (to know if this feature is there or not), but you might also consider setting ip tcp adjust-mss and ip mtu values on your tunnel interfaces to signal the true maximum-transportable-size of the various traffic types over the tunnel. I've also been bit by this bug before (http://networknerd.wordpress.com/2010/06/11/cisco-ipsec-mtu-bug/) that affects the MTU calculations of tunnels, for which the source address is specified by an interface, after a reboot. Worth knowing about. Cheers, jof
Re: Why use PeeringDB?
On Wed, Jul 18, 2012 at 8:43 AM, Chris Grundemann cgrundem...@gmail.com wrote: I am currently working on a BCOP for IPv6 Peering and Transit and would very much appreciate some expert information on why using PeeringDB is a best practice (or why its not). All opinions are welcome, but be aware that I plan on using the responses to enhance the document, which will be made publicly available as one of several (and hopefully many more) BCOPs published at http://www.ipbcop.org/. It's a nice resource for finding out which networks are in which facilities. As someone seeking out and setting up peering sessions, it's useful to be able to search out networks that also have a couple common POPs, so that one can call or email them and ask about potential interconnection. It's certain cut down on emails that are just requests for information (Where do you have sites? We're in these metros..., Looks like we'd be good potential peers, what's your policy like?). Overall -- I really like it! Cheers, jof
Re: Why use PeeringDB?
On Wed, Jul 18, 2012 at 9:59 AM, Zaid Ali z...@zaidali.com wrote: The goal is Source of truth for any peer to know information at the Exchange points as well as peering coordinator information. I think it is a great tool for the peering community and definitely useful. Cons: Will it be the next RADB? There needs to be a sustainable community to keep it running since it is a volunteer effort. Good point. I suspect that enough large users (with money, developers, hosting, etc.) are enjoying it that it has reached a critical mass of a semi-core service that wont have a hard time getting some support going forward. --j
Re: technical contact at ATT Wireless
On Thu, Jun 28, 2012 at 1:50 PM, Christopher Morrow morrowc.li...@gmail.com wrote: of course, but you aren't supposed to be doing that on their network anyway... so says the nice man from sprint 4 nanogs ago. That, and if you are tunneling in, it's good practice to forward over any DNS traffic as well (or all, depending on the application). That way, if you have internal names or special resolvers setup, you'll hit that as well. Cheers, jof
Re: Peer1/Server Beach support for BGP on dedicated servers
On Sat, May 19, 2012 at 3:23 AM, Anurag Bhatia m...@anuragbhatia.com wrote: Was wondering if there's anyone from Server Beach/Peer1 here. We have a dedicated server with them which we primarily use for DNS. I am adding support for anycasting on that one but seems like Peer1 is not supporting BGP at all. NOC support told me that they can announce our block and statically pass us but cannot hear BGP announcement from our router. Was wondering if someone else had similar issue? Generally, most dedicated hosting (renting/leasing the exclusive use of a computer in their facility) outfits aren't setup to speak BGP to individual servers/customers. Such a request is usually infrequent enough that it doesn't warrant setting up the added hardware. While you could have your provider announce your space for you, you'll loose the fine-grained control over how that route gets announced once the routing is out of your hands. This is important and if doesn't works then I would have to find a new place for dedicated server somewhere in California. I would instead recommend looking for a colocation provider that will host small installations (1 - 5 U), but is also savvy enough to speak eBGP with their customers. Cheers, jof
Re: Squeezing IPs out of ARIN
On Wed, Apr 25, 2012 at 8:46 AM, Kenneth McRae kenneth.mc...@dreamhost.comwrote: I have never provided the names of end users.. How the address space would be utilized? Definitely.. But not the names of end users... Probably because you are an end user. If you're talking about AS26347, I don't think there is any re-assigned space in there. Do you ever assign users CIDR blocks of IP space for their own use? If it's just the transitory use of IPs in an operational network you control, then that sounds like end user use to me, even though you may sell the use of those IPs. If you have questions about this stuff, the ARIN NRPM is a great resource: https://www.arin.net/policy/nrpm.html Cheers, jof
Re: Squeezing IPs out of ARIN
On Tue, Apr 24, 2012 at 10:32 AM, ad...@thecpaneladmin.com wrote: Anyone have any tips for getting IPs from ARIN? For an end-user allocation they are requesting that we provide customer names for existing allocations, which is information that will take a while to obtain. They are insisting that this is standard process and something that everyone does when requesting IPs. Has anyone actually had to do this? Indeed. It's worked this way for a long time. When starting a new organization, there's a bit of a chicken and egg problem with IP space. If anyone could get IP space just for asking for it, it would have been consumed too quickly. So, organizations must first get some space assigned to them from an upstream provider and begin using it. At some point the current usage and growth rate of the assigned space will justify a direct allocation. Then, you can renumber into your new space and be totally independent. Cheers, jof
Re: Squeezing IPs out of ARIN
On Tue, Apr 24, 2012 at 11:14 AM, Owen DeLong o...@delong.com wrote: That's not entirely true. What you say applies to one possible way for an ISP to get an allocation. It does not apply at all to end-users. Even for end-user allocations, they would still need to fulfill the requirements of 4.3.3 in the ARIN NRPM (https://www.arin.net/policy/nrpm.html#four33), no? I suppose for immediate need assignments, this can be short circuited, but from what I know those are pretty rare. Am I missing something? Cheers, jof
Re: About Juniper MX10 router performance
On Sun, Apr 22, 2012 at 9:05 PM, Md.Jahangir Hossain jrjahan...@gmail.com wrote: Dear valued member: Wishes all are fine. i need suggestion from you about Juniper MX10 router performance. i want to buy this router for IP Transit provider where i received all global routes . Do you have some specific questions about it? You should be able to comfortably take in one to two full BGP feeds as RIBs, and hold an Internet-sized FIB. It's basically an MX80 (internally), but with some software modifications to limit its performance. You probably need to make sure that you're also purchasing the S-MX80-ADV-R software license in your base bundle (to support the full-scale L3 routing), but the licensing is based on the honor system. Cheers, jof
Re: About Juniper MX10 router performance
On Sun, Apr 22, 2012 at 9:48 PM, Md.Jahangir Hossain jrjahan...@gmail.com wrote: Thanks jonathan for your reply . Actually i have not specific question , i need suggestion about this product if i purchase this as IP Transit provider. Only someone with the knowledge of your business and requirements can answer this for you. If you'd like to take this up off-list, I'm happy to share what I know. Generally, we try to keep the discussion to the list (that is sent to the many subscribers) on topics that are of interest to all network operators. For Juniper-specific questions, the j-nsp mailing list is pretty great (http://puck.nether.net/mailman/listinfo/juniper-nsp). Cheers, jof
Re: airFiber (text of the 8 minute video)
On Thu, Mar 29, 2012 at 12:33 PM, Oliver Garraux oli...@g.garraux.net wrote: I was at Ubiquiti's conference. I don't disagree with what you're saying. Ubiquiti's take on it seemed to be that 24 Ghz would likely never be used to the extent that 2.4 / 5.8 is. They are seeing 24 Ghz as only for backhaul - no connections to end users. I suspect this is just due to cost and practicality. ISPs, nor users will want to pay 3k USD, nor widely utilize a service that requires near-direct LOS. I could see this working well in rural or sparse areas that might not mind the transceiver. I guess point-to-multipoint connections aren't permitted by the FCC for 24 Ghz. The whole point of these unlicensed bands is that their usage is not tightly controlled. I imagine hardware for use still should comply with FCC's part 15 rules though. AirFiber appears to be fairly highly directional. It needs to be though, as each link uses 100 Mhz, and there's only 250 Mhz available @ 24 Ghz. Being so directional, I'm not sure that cross-talk will as much of an issue, except for dense hub-like sites. It sounds like there's some novel application of using GPS timing to make the radios spectrally orthogonal -- that's pretty cool. If they can somehow coordinate timing across point-to-point links, that would be great for sites that co-locate multiple link terminations. Overall, this looks like a pretty cool product! --j
Re: airFiber (text of the 8 minute video)
On Thu, Mar 29, 2012 at 2:37 PM, Joel jaeggli joe...@bogus.com wrote: Cost will continue to drop, fact of the matter is the beam width is rather narrow and they attenuate rather well so you can have a fair number of them deployed without co-channel interference. if you pack a tower full of them you're going to have issues. This is exactly the kind of case that I'm thinking about (central towers). The novel thing Ubiquiti seems to do is TDMA-like channelization (like with Airmax), or by changing the coding scheme over the air to maintain orthogonality (what it sounds like this new product may be doing). --j
Re: Concern about gTLD servers in India
On Sat, Mar 10, 2012 at 10:45 AM, Bill Woodcock wo...@pch.net wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On Mar 10, 2012, at 8:05 AM, Suresh Ramasubramanian wrote: Sure, if you can find a datacenter that's capable of handling all the traffic, and has staff who are able to provide efficient remote hands for huge racks of extremely powerful servers . Honestly, we haven't even gotten that far when we've offered to deploy servers (for instance for domains like .IN) inside India. The bribes that were requested in exchange for giving us permission to deploy a free service were, uh, both prohibitive and ludicrous in their enormity. This. This and the import duties on hardware and the requirement for licensing to operate as an ISP makes placing even a modest deployment a lot more work compared to deploying in other neighboring countries. I would presume that Verisign decided that it just wasn't worth the effort to deploy into India. It obviously has a gigantic user base for which getting into local ISPs and IXPs would probably save on transit costs. Perhaps if some local root operators could donate some space/power/connectivity, Verisign-grs could colocate a gTLD cluster there? Cheers, jof
Re: WW: Colo Vending Machine
On Fri, Feb 17, 2012 at 10:35 AM, Jay Ashworth j...@baylink.com wrote: Please post your top 3 favorite components/parts you'd like to see in a vending machine at your colo; please be as specific as possible; don't let vendor specificity scare you off. This is a riot! I'd love to have something like this at facilities I'm in. Some useful stuff that comes to mind: - Rack screws of various common sizes and threadings - SFPs, GBICs, etc. - Rollover cable / DE-9-8P8P adapter - Screwdrivers - Cross-over Ethernet, patch cables - zip ties, velcro tape, etc. - Label tape Cheers, jof
Re: WW: Colo Vending Machine
On Fri, Feb 17, 2012 at 10:55 AM, Leo Bicknell bickn...@ufp.org wrote: In a message written on Fri, Feb 17, 2012 at 01:35:15PM -0500, Jay Ashworth wrote: Please post your top 3 favorite components/parts you'd like to see in a vending machine at your colo; please be as specific as possible; don't let vendor specificity scare you off. USB-Serial adapters. Preferably selected so they are driverless on both OSX and Windows. :) Does such a device exist? I've yet to run across one. Personally, I would recommend those based on FTDI chips or Prolific PL2303 -- both have support for Linux, Windows, and OSX. --j
Re: 802.11 MAC Point Coordination Function
On Wed, Feb 15, 2012 at 8:13 PM, Jeremy jba...@gmail.com wrote: I'm doing some research on 802.11 quality of service, congestion control, etc. I'm trying to find some information on the Point Coordination Function, a polling based access control method, but I'm having a hard time finding much in the way of vendor support. I have access to some cisco 1242's, 1140's and 1252's and I've been searching the Cisco's site and can't find a real answer on whether or not it's supported let alone how to configure it. Does anyone have any experience with this? Does Cisco have some special name for it aside from PCF? Any help would be appreciated! I know of no such feature in classic Aironet APs. Everything I've run across only does classic 802.11-style DCF along with all the stations. You might check out the Aironet 1550, which supports forming dynamic mesh-like topologies, though I can't speak to it as I've never laid hands on the hardware or software. Other vendors and hackers are implementing alternative media-access control and coordination functions, though. You might check out Ubiquiti's AirMax-speaking hardware or ieee80211_tdma in FreeBSD for software MAC-capable drivers. Cheers, jof
Re: Wireless Recommendations
On Wed, Feb 15, 2012 at 7:50 PM, Faisal Imtiaz fai...@snappydsl.net wrote: Is that because of Channel Spacing ? or some other reason ? I would presume channel spacing. In FCC-land, there are only 3 non-overlapping 20 Mhz bandwidths available. --j
Re: Wireless Recommendations
On Wed, Feb 15, 2012 at 8:41 PM, Joel jaeggli joe...@bogus.com wrote: On 2/15/12 20:14 , Mario Eirea wrote: This is my guess too, i guess there is some bleed over from their antenna arrays. Even the most directional sector antenna in the world has a back lobe... and there there's the clients... Agreed. There is rarely a thing as a perfectly-directional antenna (not without a lot of shielding, I would presume). Since I would presume that all the radios are controlled by the same host, perhaps it could coordinate the 802.11 DCF and sequence CTS frames so that the various client and AP radios remain as spectrally orthogonal as possible. There's not much you can do about the clients transmitting RTSes, but it can be predicted to a certain extent. there's no magic bullet you simply can't do it all in one ap with the space available. Agreed. More, lower-power APs means better spectral efficiency and overall resilience. --j
Re: Wireless Recommendations
On Tue, Feb 7, 2012 at 11:19 AM, Arzhel Younsi xio...@gmail.com wrote: Xirrus say that they can support 640 clients with this device: http://www.xirrus.com/Products/Wireless-Arrays/XR-Series/XR-4000-Series I heard about it a couple weeks ago, didn't try it yet. That's a pretty neat product -- it seems like it takes care of spectrally isolating clients by utilizing 4 - 8 radios per AP-box and 8 - 24 directional sector antennas. I feel like this addresses the suggestions that I and others gave to utilize more APs rather than a big central one, but it just packages it all into one box with many antennas. Cheers, jof
Re: Hijacked Network Ranges
On Tue, Jan 31, 2012 at 10:19 AM, Grant Ridder shortdudey...@gmail.com wrote: Hi, What is keeping you from advertising a more specific route (i.e /25's)? Most large transits and NSPs filter out prefixes more specific than a /24. Conventionally, at least in my experience, /24's are the most-specific prefix you can use and expect that it will end up in most places. Some shops with limited router processing or table storage capacity will filter even more restrictively, so a bigger aggregate is worth announcing as well. Cheers, jof
Re: Hijacked Network Ranges
On Tue, Jan 31, 2012 at 10:00 AM, Kelvin Williams kwilli...@altuscgi.com wrote: We've been in a 12+ hour ordeal requesting that AS19181 (Cavecreek Internet Exchange) immediately filter out network blocks that are being advertised by ASAS33611 (SBJ Media, LLC) who provided to them a forged LOA. [ ...snip...] Ugh, what a hassle. I've been there, and it's really no fun. Our customers and services have been impaired. Does anyone have any contacts for anyone at Cavecreek that would actually take a look at ARINs WHOIS, and IRRs so the networks can be restored and our services back in operation? Have you tried the contacts listed at PeeringDB for AS19181? Check out: as19181.peeringdb.com Additionally, does anyone have any suggestion for mitigating in the interim? Since we can't announce as /25s and IRRs are apparently a pipe dream. If you fail to get AS19181 to respond, you might consider contacting *their* upstreams and explaining the situation. Cheers, jof
Re: Wireless Recommendations
On Mon, Jan 30, 2012 at 12:46 PM, Jim Gonzalez j...@impactbusiness.com wrote: Hi, I am looking for a Wireless bridge or Router that will support 600 wireless clients concurrently (mostly cell phones). I need it for a proof of concept. I've had some great luck with a variety of vendors, though never with this many clients on one AP. For a stable 802.11 stack, I've found Cisco AP1142N's to be great. That said, I'm not sure what you're trying to do here, but I think you'll be disappointed with any AP with 600 *active* stations associated to it. No AP can work around the congestive collapse of hundreds of stations all transmitting RTS frames at once. If you can split up your many stations across a swath of APs, bridging down to a couple L2 Ethernet LANs, I think you'll get something much more scalable. Cheers, jof
Re: Populating BGP from Connected or IGP routes
On Mon, Jan 23, 2012 at 12:46 PM, Eric C. Miller e...@ericheather.com wrote: Hi all, I'm looking for a best practice sort of answer, plus maybe comments on why your network may or may not follow this. First, when running a small ISP with about the equivilent of a /18 or /19 in different blocks, how should you decide what should be in the IGP and what should be in BGP? I assume that it's somewhere between all and none, and one site that I found made some good sense saying something to the following, Use a link-state protocol to track interconnections and loopbacks only, and place all of the networks including customer networks into BGP. Secondly, when is it ok, or preferable to utilize redistribute connected for gathering networks for BGP over using a network statement? I know that this influences the origin code, but past that, why else? Would it ever be permissible to redistribute from the IGP into BGP? This is one of those questions where the answer will depend heavily on who you ask. In my opinion, I would - Keep externally-learned eBGP routes in one table. The Internet table. - Keep internal links (loopbacks, single-homed (to me) customers, networks containing next-hops outside your AS) in an IGP (like OSPF or IS-IS). These routes should very rarely get exchanged outside the AS. - Where possible, have multi-homed customers speak BGP to your AS and just treat those routes as those you'll provide transit for (re-announcing them to other external peers) -- In cases where customers multi or single-home with their own address space that they'd like you to address, put very specific filters and tagging on the routes. This way, you can perform careful filtering on allowing those routes to cross the boundary from IGP to EGP (and onto your external peers). Cheers, jof
Re: bgp question
On Wed, Jan 18, 2012 at 5:58 AM, Deric Kwok deric.kwok2...@gmail.com wrote: ls it supporting equally multipath in different bgp connections? Most software routing protocols have support for this in their RIBs, but the actual forwarding ability of the underlying kernel will determine the support for this. What platform do you route on? Cheers, jof
Re: enterprise 802.11
On Sun, Jan 15, 2012 at 3:36 PM, Greg Ihnen os10ru...@gmail.com wrote: Since we're already top-posting… I've heard a lot of talk on the WISPA (wireless ISP) forum that 802.11g/n starts to fall apart with more than 30 clients associated if they're all reasonably active. I believe this is a limitation of 802.11g/n's media access control (MAC) mechanism, regardless of who's brand is on the box. This is most important if you're doing VoIP or anything else where latency and jitter is an issue. To get around that limitation, folks are using proprietary protocols with polling media access control. Ubiquiti calls theirs AirMax. Cisco uses something different in the Canopy line. But of course then you've gone to something proprietary and only their gear can connect. So it's meant more for back-hauls and distribution networks, not for end users unless they use a proprietary CPE. Since you need consumer gear to be able to connect, you need to stick with 802.11g/n. You should limit to 30 clients per AP. You should stagger your 2.4GHZ APs on channels 1, 6 and 11, and turn the TX power down and have them spaced close enough that no more than 30 will end up connecting to a single AP. 5.8GHz APs would be better, and you'll want to stagger their channels too and turn the TX power down so each one has a small footprint to only serve those clients that are nearby. Stay away from mesh solutions and WDS where one AP repeats another, that kills throughput because it hogs airtime. You'll want to feed all the APs with Ethernet. After working in some WISP-like and access environments, I con corroborate that this is pretty much true. It becomes worse the lower the SNR is and the more that clients are spread out. It just makes the 'hidden node' problem worse. Making APs as low power and local as possible is good advice. Where possible, feed everything with hardlines back to your Ethernet switching environment. If client roaming and client-client traffic is important, using a central controller that can tunnel 802.11 frames over whatever wired L2 network you like is a good win. It means that to clients they can associate and/or authenticate to one AP and roam from place to place while keeping the same session to the controller. As far as vendor gear goes, if roaming and client-client stuff isn't as important, Ubiquiti UnFi is great stuff for the price. Next rung up in my book would be Meraki, followed by Cisco or Aruba. Good luck! Cheers, jof
Re: Linux Centralized Administration
On Thu, Jan 12, 2012 at 1:02 PM, Paul Stewart p...@paulstewart.org wrote: Hey folks. just curious what people are using for automating updates to Linux boxes? Today, we manually do YUM updates to all the CentOS servers . just an example but a good one. I have heard there are some open source solutions similar to that of Red Hat Network? There's no tool I could recommend that would be very close to RHN. However, for solving the problem of keeping packages up to date and systems in a known-state, I would recommend checking out some configuration management tools. There are several popular ones nowadays, though I personally prefer Puppet or Chef. Both are tools that allow administrators to declare what a system should look like, and abstract away the hard work of making that happen on a variety of platforms. In both cases, it's possible to monitor how well those tools are working and what they're doing in the background so that you can get an idea of what's up to date and what's not. Are you just trying to solve for making sure that packages are up to date? Making sure that running daemons are also up to date? Cheers, jof
Re: bgp question
On Tue, Jan 10, 2012 at 2:43 PM, Deric Kwok deric.kwok2...@gmail.comwrote: Hi all When we get newip, we should let the upstream know to expor it as there should have rule in their side. how about upstream provider, does they need to let their all bgp interconnect to know those our newip? If no, Can I know how it works? If they don't have rules each other, ls it any problems? It depends on your upstream ISPs. Conventionally, some choose to place exact filters in place on BGP announcements that exactly match IP space that is registered with a RIR or LIR, some build those filters from IRR sources, and others just filter on the number of prefixes your sending (to avoid sending a whole table out on accident). I'm sure there are some other filtering schemes in place around the world. In the case of exact filters, you'll need to contact your upstream ISPs and ask them to update their filters. In the case of IRR-sourced filtering information, update the prefixes that you originate with your IRR provider. And in the case of max-prefix filtering, ask your ISP what they have their equipment set to. Cheers, jof
Re: subnet prefix length 64 breaks IPv6?
On Sat, Dec 24, 2011 at 6:48 AM, Glen Kent glen.k...@gmail.com wrote: SLAAC only works with /64 - yes - but only if it runs on Ethernet-like Interface ID's of 64bit length (RFC2464). Ok, the last 64 bits of the 128 bit address identifies an Interface ID which is uniquely derived from the 48bit MAC address (which exists only in ethernet). SLAAC could work ok with /65 on non-Ethernet media, like a point-to-point link whose Interface ID's length be negotiated during the setup phase. If we can do this for a p2p link, then why cant the same be done for an ethernet link? I think by point-to-point, Alexandru was referring to PPP-signalled links. In the case of Ethernet and SLAAC, the standards define a way to turn a globally unique 48-bit 802.3 MAC-48 address into an EUI-64 identifier by flipping and adding some bits. This uniquely maps conventional MAC-48 addresses into EUI-64 addresses. I imagine this was chosen because the IEEE is encouraging new standards and numbering schemes to use the 64-bit schemes over the older 48-bit ones. Presumably to avoid exhaustion in the future (like we're seeing with IPv4). The result of which is that with the standards we've got today, we can easily map a piece of hardware's globally unique MAC address into a globally unique 64-bit identifier -- which happens to cleanly fit into the second half of the v6 address space. I suppose one could make an argument to use /80 networks and just use the MAC-48 identifier for the address portion, but given the vastness of v6 space I don't think it's really worth the extra savings of bit space. So, to address your original question, in v6 networks with netmask lengths greater than 64 bits nothing breaks per se, but some of the conventional standards and ideas about what a network is in that context are broken. While it's not possible to have hosts uniquely pick addresses for themselves, one can use other addressing mechanisms like DHCPv6 or static addresses. --j
Re: Multiple ISP Load Balancing
The best applications for analyzing paths, that I've seen, have been in-house development projects. So, admittedly, I don't have much experience with commercial products for route optimization. Projects I've seen that analyze best paths to Internet destinations via multiple ISPs add instrumentation to content-serving applications to log stream performance details to a database or log collection system along with a timestamp. Another database keeps a periodic log of RIB data that lists the specific next-hops out of the AS. Another log keeps a running log of UPDATEs. From joining up all of this information, you can figure out the ISP you're taking to a destination (at a given time) and how the stream performed. Then, add some logic to inject routes to try out different next-hop ISPs for some destinations. Then, compare the newer ISP-path to the older one and see which performs best. Where best means something specific to your application (optimizing for latency, cost, etc.) Cheers, jof
Re: Inaccessible network from Verizon, accessible elsewhere.
On Sat, Dec 10, 2011 at 11:49 AM, NetSecGuy netsec...@gmail.com wrote: I have a Linode VPS in Japan that I can't access from Verizon FIOS, but can access from other locations. I'm not sure who to blame. The host, 106.187.34.33, is behind the gateway 106.187.34.1: From FIOS to 106.187.34.1 (this works). traceroute to 106.187.34.1 (106.187.34.1), 64 hops max, 52 byte packets 4 so-6-1-0-0.phil-bb-rtr2.verizon-gni.net (130.81.199.4) 9.960 ms 9.957 ms 6.666 ms 5 so-8-0-0-0.lcc1-res-bb-rtr1-re1.verizon-gni.net (130.81.17.3) 12.298 ms 13.463 ms 13.706 ms 6 0.ae2.br1.iad8.alter.net (152.63.32.158) 14.571 ms 14.372 ms 14.003 ms 7 204.255.169.218 (204.255.169.218) 14.692 ms 14.759 ms 13.670 ms 8 sl-crs1-dc-0-1-0-0.sprintlink.net (144.232.19.229) 13.077 ms 12.577 ms 14.954 ms 9 sl-crs1-nsh-0-5-5-0.sprintlink.net (144.232.18.200) 31.443 ms sl-crs1-dc-0-5-3-0.sprintlink.net (144.232.24.37) 33.005 ms sl-crs1-nsh-0-5-5-0.sprintlink.net (144.232.18.200) 31.507 ms 10 sl-crs1-kc-0-0-0-2.sprintlink.net (144.232.18.112) 57.610 ms 58.322 ms 59.098 ms 11 otejbb204.kddnet.ad.jp (203.181.100.45) 196.063 ms otejbb203.kddnet.ad.jp (203.181.100.13) 188.846 ms otejbb204.kddnet.ad.jp (203.181.100.21) 195.277 ms 12 cm-fcu203.kddnet.ad.jp (124.215.194.180) 214.760 ms cm-fcu203.kddnet.ad.jp (124.215.194.164) 198.925 ms cm-fcu203.kddnet.ad.jp (124.215.194.180) 200.583 ms 13 124.215.199.122 (124.215.199.122) 193.086 ms * 194.967 ms This does not work from FIOS: traceroute to 106.187.34.33 (106.187.34.33), 64 hops max, 52 byte packets 4 so-6-1-0-0.phil-bb-rtr2.verizon-gni.net (130.81.199.4) 34.229 ms 8.743 ms 8.878 ms 5 so-8-0-0-0.lcc1-res-bb-rtr1-re1.verizon-gni.net (130.81.17.3) 15.402 ms 13.008 ms 14.932 ms 6 0.ae2.br1.iad8.alter.net (152.63.32.158) 13.325 ms 13.245 ms 13.802 ms 7 204.255.169.218 (204.255.169.218) 14.820 ms 14.232 ms 13.491 ms 8 lap-brdr-03.inet.qwest.net (67.14.22.78) 90.170 ms 92.273 ms 145.887 ms 9 63.146.26.70 (63.146.26.70) 92.482 ms 92.287 ms 94.000 ms 10 sl-crs1-kc-0-0-0-2.sprintlink.net (144.232.18.112) 58.135 ms 58.520 ms 58.055 ms 11 otejbb203.kddnet.ad.jp (203.181.100.17) 205.844 ms otejbb204.kddnet.ad.jp (203.181.100.25) 189.929 ms otejbb203.kddnet.ad.jp (203.181.100.17) 204.846 ms 12 sl-crs1-oro-0-1-5-0.sprintlink.net (144.232.25.77) 87.229 ms sl-crs1-oro-0-3-3-0.sprintlink.net (144.232.25.207) 88.796 ms 88.717 ms 13 124.215.199.122 (124.215.199.122) 193.584 ms 202.208 ms 192.989 ms 14 * * * Same IP from different network: traceroute to 106.187.34.33 (106.187.34.33), 30 hops max, 60 byte packets 6 ae-8-8.ebr2.Washington1.Level3.net (4.69.134.105) 2.230 ms 1.847 ms 1.938 ms 7 ae-92-92.csw4.Washington1.Level3.net (4.69.134.158) 2.010 ms 1.985 ms ae-62-62.csw1.Washington1.Level3.net (4.69.134.146) 1.942 ms 8 ae-94-94.ebr4.Washington1.Level3.net (4.69.134.189) 12.515 ms ae-74-74.ebr4.Washington1.Level3.net (4.69.134.181) 12.519 ms 12.507 ms 9 ae-4-4.ebr3.LosAngeles1.Level3.net (4.69.132.81) 65.957 ms 65.958 ms 66.056 ms 10 ae-83-83.csw3.LosAngeles1.Level3.net (4.69.137.42) 66.063 ms ae-93-93.csw4.LosAngeles1.Level3.net (4.69.137.46) 65.985 ms ae-63-63.csw1.LosAngeles1.Level3.net (4.69.137.34) 66.026 ms 11 ae-3-80.edge2.LosAngeles9.Level3.net (4.69.144.143) 66.162 ms 66.160 ms 66.238 ms 12 KDDI-AMERIC.edge2.LosAngeles9.Level3.net (4.53.228.14) 193.317 ms 193.447 ms 193.305 ms 13 lajbb001.kddnet.ad.jp (59.128.2.101) 101.544 ms 101.543 ms lajbb002.kddnet.ad.jp (59.128.2.185) 66.563 ms 14 otejbb203.kddnet.ad.jp (203.181.100.13) 164.217 ms 164.221 ms 164.330 ms 15 cm-fcu203.kddnet.ad.jp (124.215.194.164) 180.350 ms cm-fcu203.kddnet.ad.jp (124.215.194.180) 172.779 ms cm-fcu203.kddnet.ad.jp (124.215.194.164) 185.824 ms 16 124.215.199.122 (124.215.199.122) 175.703 ms 175.700 ms 168.268 ms 17 li377-33.members.linode.com (106.187.34.33) 174.381 ms 174.383 ms 174.368 ms In doing a little probing right now, from various source addresses, I'm unable to reproduce the problem. I've seen failures similar to this one (where the source address matters; some work, some don't) when multi-port LAGs or ECMP paths have a single link in them fail, but are still detected and forwarded over as if it was up. This can happen, for example, if you run a LAG with no channeling protocol (like LACP or PAGP), that hashes source and destination IPs to pick a link (to ensure consistent paths per-path, and with ports, per-flow). If one of those links fails in the underlying media or physical path, but the link is still detected as up, packets to some IPs (but not others) will just drop on the flor. Now, in this particular case, it doesn't seem like the path to both destinations seem like they even take the same path (so that previous hypothesis is pure conjecture). Perhaps routes were actively
Re: Internet Edge and Defense in Depth
I would argue that collapsing all of your policy evaluation and routing for a size/zone/area/whatever into one box is actually somewhat detrimental to stability (and consequently, security to a certain extent). Cramming every little feature under the sun into one appliance makes for great glossy brochures and Powerpoint decks, but I just don't think it's practical. Take a LAMP hosting operation for example. Which will scale the furthest to handle the most amount of traffic and stateful sessions: iptables and snort on each multi-core server, or one massive central box with some interface hardware and Cavium Octeons. If built properly, my money's on the distributed setup. Cheers, jof
Re: IPv6 prefixes longer then /64: are they possible in DOCSIS networks?
On Mon, Nov 28, 2011 at 10:43 PM, valdis.kletni...@vt.edu wrote: On Tue, 29 Nov 2011 00:15:02 EST, Jeff Wheeler said: Owen and I have discussed this in great detail off-list. Nearly every time this topic comes up, he posts in public that neighbor table exhaustion is a non-issue. I thought I'd mention that his plan for handling neighbor table attacks against his networks is whack-a-mole. That's right, wait for customer services to break, then have NOC guys attempt to clear tables, filter traffic, or disable services; and repeat that if the attacker is determined or going after his network rather than one of his downstream customers. It's worked for us since 1997. We've had bigger problems with IPv4 worms that decided to probe in multicast address space for their next target, causing CPU exhaustion on routers as they try to set up zillions of multicast groups. Sure, it's a consideration. But how many sites are *actually* getting hit with this, compared to all the *other* DDOS stuff that's going on? I'm willing to bet a large pizza with everything but anchovies that out in the *real* world, 50-75 times as many (if not more) sites are getting hit with IPv4 DDoS attacks that they weren't prepared for than are seeing this one particular neighbor table exhaustion attack. Any of the guys with actual DDoS numbers want to weigh in? Agreed. While I don't have any good numbers that I can publicly offer up, it also intuitively makes sense that there's a greater proportion of IPv4 DDOS and resource exhaustion attacks vs IPv6 ones. Especially on the distributed part; there's a heck of lot more v4-only hosts to be 0wned and botnet'ed than dual-stacked ones. That said, I think the risk of putting a /64 on a point-to-point link is much greater than compared to a (incredibly wasteful) v4 /24 on a point-to-point. Instead of ~254 IPs one can destinate traffic towards (to cause a ARP neighbor discovery process), there's now ~18446744073709551614 addresses one can cause a router to start sending ICMPv6 messages for. For links that will only ever be point-to-point, there's no good reason (that I know of) to not use a /127. For peering LANs or places that have a handful of routers, /112's are cute, but I would think the risk is then comparable to a /64 (which has the added benefit of SLAAC). I imagine the mitigation strategies are similar for both cases though: just rate-limit how often your router will attempt neighbor discovery. Are there other methods? Cheers, jof
Re: ASA log viewer
On Sat, Nov 19, 2011 at 4:51 PM, Duane Toler deto...@gmail.com wrote: Hey NANOG! My employer is deploying CIsco ASA firewalls to our clients (specifically the 5505, 5510 for our smaller clients). We are having problems finding a decent log viewer. Several products seem to mean well, but they all fall short for various reasons. We primarily use Check Point firewalls, and for those of you with that experience, you know the SmartViewer Tracker is quite powerful. Is there anything close to the flexibility and filtering capabilities of Check Point's SmartView Tracker? For now, I've been dumping the logs via syslog with TLS using syslog-ng to our server, but that is mediocre at best with varying degrees of reliability. The syslog-ng server then sends that to a perl script to put that into a database. That allows us to run our monthly reports, but that doesn't help us with live or historical log parsing and filtering (see above, re: SmartView Tracker). It sounds like you've already got a pretty good aggregation setup going, here. I've had great luck with UDP Syslog from devices to a site-local log aggregator that then ships off log streams to a central place over TCP (for the WAN paths) and/or TLS/SSL. It sounds like you may have something similar going here, though I'd be curious to know where you've had this fall down reliability-wise. If a customer called to help us troubleshoot connection issues over the past few days, there's no way to review the logs and figure out what happened back then. Every CCIE we've talked to, and Cisco themselves, seem to not care about firewall traffic logs or the ability to parse and review them. We know about Cisco Security Center, but that seems incapable of handling logs, etc. CS-MARS would've been great, but that's overpriced and now discontinued anyway. We'd hate to spend the time writing our own app if there's a viable product already available (we're willing to pay a reasonable price for one, too). I don't know of any great commercial products, as I've only built homegrown tools for various organizations. I'm curious though, what kinds of features are you looking for? Searching log data? Alerting on events based on log data? Cheers, jof
Re: ASA log viewer
On Sat, Nov 19, 2011 at 5:32 PM, Duane Toler deto...@gmail.com wrote: On Sat, Nov 19, 2011 at 20:04, Jay Ashworth j...@baylink.com wrote: - Original Message - From: Duane Toler deto...@gmail.com My employer is deploying CIsco ASA firewalls to our clients (specifically the 5505, 5510 for our smaller clients). We are having problems finding a decent log viewer. Several products seem to mean well, but they all fall short for various reasons. We primarily use Check Point firewalls, and for those of you with that experience, you know the SmartViewer Tracker is quite powerful. Is there anything close to the flexibility and filtering capabilities of Check Point's SmartView Tracker? Is your problem the aggregation proper, or the mining? Do the ASA's log to syslog? Cheers, -- jra -- Yep, we log to syslog, and the issue is the mining. Not that I/we *can't* grep/regex/sed/awk/perl our way thru the log files. It's just that it's overly tedious. Especially when compared to Check Point's product (given that they are aiming to compete...). I'd second Mike's suggestion then -- check out Splunk. They make a commercial log viewing, searching, and reporting product that's pretty awesome. They license based on log volume, and the pricing scales somewhat logarithmically. So, I would consider your log volume and budget before sinking too much time into it. There's a free trial installation and license that's available if you want to try it out. Cheers, jof
Re: ASA log viewer
On Sat, Nov 19, 2011 at 5:46 PM, Duane Toler deto...@gmail.com wrote: On Sat, Nov 19, 2011 at 20:30, Jonathan Lassoff j...@thejof.com wrote: On Sat, Nov 19, 2011 at 4:51 PM, Duane Toler deto...@gmail.com wrote: Hey NANOG! My employer is deploying CIsco ASA firewalls to our clients (specifically the 5505, 5510 for our smaller clients). We are having problems finding a decent log viewer. Several products seem to mean well, but they all fall short for various reasons. We primarily use Check Point firewalls, and for those of you with that experience, you know the SmartViewer Tracker is quite powerful. Is there anything close to the flexibility and filtering capabilities of Check Point's SmartView Tracker? For now, I've been dumping the logs via syslog with TLS using syslog-ng to our server, but that is mediocre at best with varying degrees of reliability. The syslog-ng server then sends that to a perl script to put that into a database. That allows us to run our monthly reports, but that doesn't help us with live or historical log parsing and filtering (see above, re: SmartView Tracker). It sounds like you've already got a pretty good aggregation setup going, here. I've had great luck with UDP Syslog from devices to a site-local log aggregator that then ships off log streams to a central place over TCP (for the WAN paths) and/or TLS/SSL. It sounds like you may have something similar going here, though I'd be curious to know where you've had this fall down reliability-wise. We considered that, but didn't want to burden small customers with a classic scenario of ok well you have to have our other box in your room and have to deal with procurement, maintenance, upkeep, monitoring, blah blah. Recent ASA code (8.3-ish, 8.4? i forget) had syslog-tls built in and finally able to ship logs out across the lowest security zone, which was quite a nice addition. Ah, this totally makes sense now. I can see why you'd want to use features that are already on your ASAs. Sounds like a bug to me, though. I wonder what Cisco calls syslog-tls though. Syslog-like packet bodies, over a TLS-wrapped TCP socket? Sorry to hear it's been so unreliable -- I guess that's why I'm biased towards just running generic PCs and open source software for this kind of stuff; when bugs happen, you're actually empowered to debug and fix problems. I'd like to fully search on an 'column', a la 'ladder logic' style., as well as have the data presented in an orderly well-defined fashion. I know that sounded like the beginnings of use XML! but oh dear, not XML, please. :) Poor syslog is just too flat and in a state of general disarray. The bizarre arrangement of connection setup, NAT, non-NAT, traffic destined to the device, originating from the device, traffic routing across the to another zone, etc. ... it's very nonsensical, verbose, and frankly maddening. This does indeed sound like a good application for splunk. They have ways of defining custom logging formats that will parse out simple column and message types so that you can construct queries based on that information. There's some more information here in Splunk's docs on custom field extraction: http://docs.splunk.com/Documentation/Splunk/latest/Knowledge/Managesearch-timefieldextractions Cheers, jof
Re: Cable standards question
On Mon, Nov 14, 2011 at 7:12 AM, Jon Lewis jle...@lewis.org wrote: On Mon, 14 Nov 2011, Sam (Walter) Gailey wrote: My question is this; Is there an appropriate standard to specify for fiber-optic cabling that if it is followed the fiber will be installed correctly? Would specifying TIA/EIA 568-C.3, for example, be correct? I'm envisioning something like; The vendor will provide fiber connectivity between (building A) and (building B). Vendor will be responsible for all building penetrations and terminations. When installing the fiber-optic cable the vendor will follow the appropriate TIA/EIA 568 standards for fiber-optic cabling. At minimum, I think you should probably specify the type and number of fibers you want. i.e. Based on the distance and gear you'll be using, do you need single-mode, or will multi-mode do (as well as the core/cladding diameter)? Generally, but not always, fiber uses one strand for transmit and another for receive, so a typical fiber run is done using duplex fiber. Some optics can transmit and receive over one strand using different wavelengths. You might even specify how you want the fiber terminated (SC, LC, cables hanging from the wall, fiber patch panel, etc.). I'd agree with this. I wouldn't worry about the standard so much as the practical aspects of a run. Once you have an idea of the approximate distance of the run, you can figure out which optics you plan on using. This will determine what physical connectors you'll need and what your approximate link budget will be. Based on that information, you can figure out which type to ask for (9um/125um single-mode, most likely), a range of path loss that you're comfortable with, and the physical termination you'd like at either end. Cheers, jof
Re: Firewalls - Ease of Use and Maintenance?
On Wed, Nov 9, 2011 at 12:44 PM, Nick Hilliard n...@foobar.org wrote: On 09/11/2011 19:07, C. Jon Larsen wrote: put the main portion of the conf in subversion as an include file and factor out local differences in the configs with macros that are defined in pf.conf Easy. As I said, it's not a pf problem. Commercial firewalls will do all this sort of thing off the shelf. It's a pain to have to write scripts to do this manually. Agreed. This is rather a pain to have to do manually each time (either scp'ing or scripting). It's unfortunate that there's not a conventional script or mechanism for doing this. I have plenty of scripts from past commercial work that do this, but they're sadly tied up license-wise. I've had good luck, pf-wise, with creating a ruleset that is just identical between hosts. By keeping the interface naming/numbering scheme consistent across two hosts, the same configuration can just work on both. Cheers, jof
Re: Firewalls - Ease of Use and Maintenance?
On Wed, Nov 9, 2011 at 5:24 AM, Nick Hilliard n...@foobar.org wrote: On 09/11/2011 12:22, Richard Kulawiec wrote: You will find it very difficult to beat pf on OpenBSD for efficiency, features, flexibility, robustness, and security. Maintenance is very easy: edit a configuration file, reload, done. There are several areas where pf falls down. One is auto-synchronisation from primary to backup firewall (not really a pf problem, but it's important for production firewall systems). I've found that this works decently well, via pfsync. It sends out multicast IP packets with multi-valued elements describing the state of the flows it has in its table. If you're having pf inspect TCP sequence numbers, there's a bit of a race condition in failover with frequently or fast-moving TCP streams. As the window of acceptable sequence numbers moves on the active firewall, they're slightly delayed in getting replicated to the backup(s) and installed in their state tables. Consequently, on failover, it's possible for some flows to get blocked and which have to be re-created. I've hit this and dug into it recently, so if you're having a problem, I'd be happy to chat offlist. Cheers, jof
Re: Firewalls - Ease of Use and Maintenance?
It really depends on what constraints you have. Do you care about: cost? performance? support? Personally, for cost-constrained applications of 1 Gbit/s or less (assuming modestly-sized packets, not all-DNS for example), I like OpenBSD/pf or Linux/netfilter and generic x86 64-bit servers. It's cheap, deeply customizable and since everything touches a CPU, it allows for deep traffic inspection. The tradeoff is that there's no support from major vendors, but there are many smaller but very experienced consulting shops that can integrate any patches and fix and issues that may arise. What kinds of things are you looking for? Cheers, jof On Tue, Nov 8, 2011 at 3:06 PM, Jones, Barry bejo...@semprautilities.com wrote: Hello all. I am potentially looking at firewall products and wanted suggestions as to the easiest firewalls to install, configure and maintain? I have a few small networks ( 50 nodes at one site, 50 odd at another, and maybe 20 at another. I have worked with Cisco Pix, ASA, Netscreen, and Checkpoint (Nokia), and each have strong and not as strong features for ease of use. Like everyone, I'm resource challenged and need an easy solution to stand up and operate. Feel free to ping me offline - and thank you for the assistance. Barry Jones - CISSP GSNA Project Manager II Sempra Energy Utilities (760) 271-6822 P please don't print this e-mail unless you really need to.
Re: IPv6 Availability on XO
On Mon, May 23, 2011 at 4:39 PM, Ryan Rawdon r...@u13.net wrote: I've heard some mixed reports of XO's IPv6 availability - some that they have full deployment/availability, but others like the answer back from our XO reseller that XO does not offer IPv6 on circuits under 45mbit/s. What is the experience of NANOG on this matter, particularly with XO connectivity under 45mbit/s? Interesting. Perhaps they haven't plumbed native v6 throughout their network? For comparison, I'm currently running some native IPv6 over XO in the San Francisco Bay Area (homed off of an XO router in Fremont, CA). The circuit is GigE. Cheers, jof
Re: what about 48 bits?
Excerpts from John Peach's message of Sun Apr 04 08:17:28 -0700 2010: On Sun, 4 Apr 2010 11:10:56 -0400 David Andersen d...@cs.cmu.edu wrote: There are some classical cases of assigning the same MAC address to every machine in a batch, resetting the counter used to number them, etc.; unless shown otherwise, these are likely to be errors, not accidental collisions. -Dave On Apr 4, 2010, at 10:57 AM, jim deleskie wrote: I've seen duplicate addresses in the wild in the past, I assume there is some amount of reuse, even though they are suppose to be unique. -jim On Sun, Apr 4, 2010 at 11:53 AM, A.B. Jr. skan...@gmail.com wrote: Hi, Lots of traffic recently about 64 bits being too short or too long. What about mac addresses? Aren't they close to exhaustion? Should be. Or it is assumed that mac addresses are being widely reused throughout the world? All those low cost switches and wifi adapters DO use unique mac addresses? Sun, for one, used to assign the same MAC address to every NIC in the same box. I could see how that *could* work as long as each interface connected to a different LAN. Maybe the NICs shared a single MII/MAC sublayer somehow? I've never borne witness to this though. Re: MAC address exhaustion, if the the second-to-least significant bit in the first byte is 0 (Globally Unique / Individually Assigned bit), then the first three bytes of the MAC should correspond to the manufacturer's Organizationally Unique Identifier. These are maintained by the IEEE, and they have a list of who's who here: http://standards.ieee.org/regauth/oui/index.shtml I haven't ever programmatically gone through the list, but it looks like a lot of the space is assigned. Cheers, jof
Re: Using private APNIC range in US
Excerpts from Jaren Angerbauer's message of Thu Mar 18 09:22:40 -0700 2010: Thanks all for the on / off list responses on this. I acknowledge I'm playing in territory I'm not familiar with, and was a bad idea to jump to the conclusion that this range was private. I made that assumption originally because the entire /8 was owned by APNIC, and just figured since the registrar owned them, it must have been a private range. :S It sounds like this range was just recently assigned -- is there any document (RFC?) or source I could look through to learn more about this, and/or provide evidence to my client? There's a couple of relevant documents you could refer them to: IANA's IPv4 Address Space Registry ( http://www.iana.org/assignments/ipv4-address-space/ ), which will show you a listing of which registries and various entities are assigned /8 chunks of IPv4 space. There's some interesting names and historical registrations in there (including 1.0.0.0/8's recent allocation to APNIC) There's also an RFC, RFC1918 that sets aside some IPv4 space for private, ad-hoc use. http://www.faqs.org/rfcs/rfc1918.html This is also a good lay reference: http://en.wikipedia.org/wiki/Private_network Have fun, jof
Re: news from Google
Excerpts from Charles Wyble's message of Thu Dec 03 10:44:49 -0800 2009: 8.8.8.8 6.6.6.6 would have been really really funny. :) Nice IPs from Level 3, huh? 6.6.6.6 belongs to the US Army. --j
Re: Layer 2 vs. Layer 3 to TOR
Excerpts from David Coulson's message of Thu Nov 12 13:07:35 -0800 2009: You could route /32s within your L3 environment, or maybe even leverage something like VPLS - Not sure of any TOR-level switches that MPLS pseudowire a port into a VPLS cloud though. I was recently looking into this (top-of-rack VPLS PE box). Doesn't seem to be any obvious options, though the new Juniper MX80 sounds like it can do this. It's 2 RU, and looks like it can take a DPC card or comes in a fixed 48-port GigE variety. I like the idea of doing IP routing to a top-of-rack or edge device, but have found others to be skeptical. Are there any applications that absolutely *have* to sit on the same LAN/broadcast domain and can't be configured to use unicast or multicast IP? --j
Re: San Francisco Power Outage
Well, the fact still remains that operating a datacenter smack-dab in the center of some of the most inflated real estate in recent history is quite a castly endeavor. I really wouldn't be all that surprised if 365 Main cut some corners here and there behind the scenes to save costs while saving face. As it is, they don't have remotely enough power to fill that facility to capacity, and they've suffered some pretty nasty outages in the recent past. I'm strongly considering the possibility of completely moving out of there. --j On 7/24/07, Patrick Giagnocavo [EMAIL PROTECTED] wrote: On Jul 24, 2007, at 6:54 PM, Seth Mattinen wrote: I have a question: does anyone seriously accept oh, power trouble as a reason your servers went offline? Where's the generators? UPS? Testing said combination of UPS and generators? What if it was important? I honestly find it hard to believe anyone runs a facility like that and people actually *pay* for it. Sad that the little Telcove DC here in Lancaster, PA, that Level3 bought a few months ago, has weekly full-on generator tests where 100% of the load is transferred to the generator, while apparently large DCs that are charging premium rates, do not. Cordially Patrick Giagnocavo [EMAIL PROTECTED] -- Jonathan Lassoff echo thejof | sed 's/^/jof@/;s/$/.com/' http://thejof.com GPG: 0xC8579EE5