|
Most carriers aren't leaving route selection up to AS_PATH, but assign LOCAL_PREF depending on a variety of voodoo (customer vs. paid peer vs. settlement free, POP origin, link capacity, etc.)
|
# ? Dec 18, 2015 05:24 |
|
|
# ? May 31, 2024 22:18 |
|
adorai posted:out of curiousity, at what size of business (bank if it matters) should I be to (be worried about) get my own AS? At least 125 public IPs in use. You can get an ASN from ARIN easily, but you don't need it unless you have IP space, and you don't need/can't get IP space unless you can fill 50% of a /24. (Actually it might be 80% now)
|
# ? Dec 18, 2015 08:24 |
|
Powercrazy posted:If you host public web services and have IPv4 space, you should have your own AS number. They are easy to get unlike IP addresses. Also if you want your own portable IPv6 space, you'll need an ASN.
|
# ? Dec 18, 2015 09:55 |
|
adorai posted:out of curiousity, at what size of business (bank if it matters) should I be to (be worried about) get my own AS? An intent to dual home with two ISPs is enough to justify an AS for ARIN.
|
# ? Dec 18, 2015 13:11 |
|
Powercrazy posted:Is it an autonomous AP? Because if so, I'd say because it is sending tagged frames.
|
# ? Dec 18, 2015 13:42 |
|
Powercrazy posted:If you host public web services and have IPv4 space, you should have your own AS number. They are easy to get unlike IP addresses. 16 bit ASNs are the next big thing! madsushi posted:At least 125 public IPs in use. You can get an ASN from ARIN easily, but you don't need it unless you have IP space, and you don't need/can't get IP space unless you can fill 50% of a /24. (Actually it might be 80% now) Note: The IPs do not have to be on the public internet just as long as they're used
|
# ? Dec 18, 2015 14:15 |
|
Powercrazy posted:Hmm. In order of preference: Carrier specific prepend/suppression communities (usually only works if the guy 2 hops out you're trying to avoid is a peer, not customer, of your provider). TE prefix on your other carriers if you can (assumes you're advertising /23s or shorter). AS path poisoning your prefix on that transit (wouldn't do this long term)
|
# ? Dec 18, 2015 18:39 |
|
doomisland posted:bgpq3 Upgrade to something more current.
|
# ? Dec 18, 2015 18:45 |
|
In the world of a huge number of services being hosted on public cloud providers, and those services all using HTTPS, how are people ensuring the correct QoS is applied to different services? Say I have a web application that all the company employees practically live in, it's hosted on AWS and maybe it pulls files attached to records out of S3. If the marketing department uses a file transfer service that uses S3 at the backend, how are people ensuring that the large download is treated at a lower priority than the smaller requests to the business application? Are there firewall features that can look at how much traffic has been transferred in a certain time period in one session and decide it's a download, do I need to hope that the applications work in such a way that I can identify their requests by looking at the DNS hostnames, or is the correct answer to use something like AWS Direct Connect for the business app and let everything else happen over the Internet?
|
# ? Dec 19, 2015 21:09 |
|
Powercrazy posted:Only takes one transit AS to strip the duplicate path, and then it does nothing. Yeah I mean your upstream carrier, as soon as it gets more than one AS away you're out of luck.
|
# ? Dec 20, 2015 22:34 |
|
The ScreenOS master password: <<< %s(un='%s') = %u Rapid7 write-up They say ~25k internet-exposed ScreenOS boxes, although I assume (and hope) many were patched.
|
# ? Dec 21, 2015 02:23 |
|
Shame on your for using a NetScreen in TYOOL2015
|
# ? Dec 21, 2015 05:35 |
|
Researching 720p 30FPS+ video conference capability (mainly for 30fps screensharing actually).. Looking into H323/SIP P2P endpoints/codecs in 2016 and it's all still multi-thousand dollar dedicated hardware systems. Meanwhile teens are trailblazing live video broadcasts from their bedroom with $200 capture cards livestreaming 1080p60fps across the globe via youtube/twitch ... is there really no middle ground for live P2P HQ video chat? Computer Serf fucked around with this message at 09:04 on Dec 21, 2015 |
# ? Dec 21, 2015 08:04 |
|
A pair of xbones with kinekt and Skype does head tracking fwiw.
|
# ? Dec 22, 2015 14:05 |
|
Panda Time posted:Researching 720p 30FPS+ video conference capability (mainly for 30fps screensharing actually).. I wouldn't consider that an apples and oranges comparison. One is a platform designed around the distribution of living streaming content to as many users as possible while injecting ads to generate revenue to cover the incredible costs of doing this. The other is trying to rely on god knows what kind of internet circuit between two points and hoping for the best. They have to make money somewhere so... hardware it is.
|
# ? Dec 22, 2015 14:57 |
|
None of the software solutions do great echo cancellation either, which the hardware systems do a pretty good job on.
|
# ? Dec 23, 2015 03:12 |
|
We use something similar to these: http://www.forbes.com/sites/roberthof/2014/02/06/google-launches-999-meeting-room-in-a-box/ It's basicaly a PC that boots straight to hangouts. Works better than any paid solution I've seen. And I believe you don't need a Google account to join a hangout anymore either. (That may be a paid feature with Gapps for work)
|
# ? Dec 26, 2015 06:13 |
|
Lol Forbes. "Turn off your ad blocker to continue."
|
# ? Dec 26, 2015 13:12 |
|
All right ciscogoons, I've got one that has me stumped. Setting up a new MPLS circuit. Vendor configures BGP, I configure BGP, we connect, all is well, except I'm not getting all the routes he's advertising to me. My router: 172.16.55.1 ASN 100 Vendor router: 172.16.55.2 ASN 65333 What I'm receiving and putting into the routing table: code:
code:
So I decide to do a wireshark capture and the first message from him is a BGP update: code:
|
# ? Dec 28, 2015 20:04 |
|
DigitalMocking posted:All right ciscogoons, I've got one that has me stumped. Your local AS is 100? Do you have allowas-in turned on?
|
# ? Dec 28, 2015 20:24 |
|
You have an import-list? Or a prefix list? You learning those routes via a different IGP? Have you tried to debug the bgp process, and the routing table to see why the routes aren't being imported even though they are being received? Also yea this: ragzilla posted:Your local AS is 100? Do you have allowas-in turned on?
|
# ? Dec 28, 2015 20:24 |
|
DigitalMocking posted:All right ciscogoons, I've got one that has me stumped. What do your route-maps and import policies look like?
|
# ? Dec 28, 2015 20:24 |
|
ragzilla posted:Your local AS is 100? Do you have allowas-in turned on? I'm an idiot. Thanks.
|
# ? Dec 28, 2015 21:02 |
|
What is everyone using for inventory tracking and does it support parent/child relationships for piece parts (line cards / modules)? We currently use an Excel spreadsheet that I built with respective tabs for each site, drop-down menus for most of the column information, and it has several functions and formulas for populating contract information from a sheet that has all of the appropriate information provided from the vendor. I want something internally accessible that runs on a LAMP environment so I can query it directly for various checks, population of devices for scripting, etc. Basically a good host database that scales well with custom fields. I looked into racktables.org, but I have an IPAM solution (phpIPAM) and don't want to migrate away from it and in order to populate devices and their respective IP's, the IP subnets had to be defined as well, which just doubles up the work. I'm looking at using the custom fields within phpIPAM but I'd rather not extend/hack my IPAM solution into an inventory tracker as that isn't the projects original intention. Anytime I go looking for a solution I end up saying "I should just convert the excel document into a php/mysql setup and be done with it" because what's out there either doesn't function in enough detail or comes with a bunch of extra features I don't want/need. I figured I would ask again before just building something myself.
|
# ? Dec 29, 2015 14:58 |
|
H.R. Paperstacks posted:What is everyone using for inventory tracking and does it support parent/child relationships for piece parts (line cards / modules)? We currently use an Excel spreadsheet that I built with respective tabs for each site, drop-down menus for most of the column information, and it has several functions and formulas for populating contract information from a sheet that has all of the appropriate information provided from the vendor. I set up netdot about a year and a half ago and just half-assed most of it, but its turned out to be surprisingly useful once you start putting some time into it. https://osl.uoregon.edu/redmine/projects/netdot There's whole sections I just don't use, but for asset tracking and referencing back to our internal asset tags as well as contracts it works great. We also use the IPAM religiously, which was a huge step up from excel spreadsheets.
|
# ? Dec 29, 2015 20:05 |
|
http://racktables.org/
|
# ? Dec 29, 2015 21:17 |
|
H.R. Paperstacks posted:What is everyone using for inventory tracking A roll of numbered stickers and an excel sheet managed by a small dedicated internal team. With some extra support from fairy dust and good feelings. It works for a multi billion dollar company, it can work for you!
|
# ? Dec 29, 2015 22:57 |
|
I wish someone had told me about translatorx a long time ago. What a great program for collaboration trace file analysis.
|
# ? Dec 31, 2015 16:33 |
|
It get used in TVOICE for course work , and, maybe you could use it for your IRL job / work, but that exam also expects you know how to read some of the traces as they are without that utility.
|
# ? Dec 31, 2015 17:04 |
|
Yeah the exams don't use it but its great for irl.
|
# ? Dec 31, 2015 17:32 |
|
I'm reviewing summarization, and I think the guy making the video made a mistake. He said that networks on 10.1.1.0, 10.1.2.0, 10.1.3.0, and 10.1.4.0 would be a 24 bit mask, and that since the first five bits of the third octet are all the same, the summarization would lead to a 29 bit mask. I'm thinking he screwed up and said 24 when he should have said 16. Am I right?
|
# ? Dec 31, 2015 20:06 |
|
I'm rusty at this but I don't think you can summarise 10.1.1.0 thru 10.1.4.255. You can do 10.1.0.0/19 to get 4 /24 subnets, but that doesn't include 10.1.4.0/24
|
# ? Dec 31, 2015 20:32 |
|
The video said the network ID would be 10.1.0.0, and the mask would be /29. Maybe I should watch some other videos. Most of the videos seem okay, and they're free on Linda, but Laz Diaz misspeaks a lot.
|
# ? Dec 31, 2015 20:39 |
|
10.1.0.0/19 would be 10.1.0.0 - 10.1.31.255. If you need four /24 networks, you would use a /22. If you specifically needed 10.1.1.0/24 through 10.1.4.255, you would have to use 10.1.0.0/21 and have some unused space. 10.1.0.0/22 would be 10.1.0.0 -10.1.3.255, so if you needed 10.1.4.0/24, you'l need to shift over one more bit.
|
# ? Dec 31, 2015 21:00 |
|
Jesus I can't type. Not sure where I got the /19 from up there. Sorry if that confused anyone.
|
# ? Dec 31, 2015 21:02 |
|
It sounds like the 3rd and 4th octets were flip flopped and he's talking about the first four usable in 10.1.0.0/29 (.1 - .4 of .1 - .6 usable)
|
# ? Dec 31, 2015 21:07 |
|
Does anyone have a reputable site to purchase Cisco odds and ends at a consumer level? For instance, I need a power cable for my 2801 router at home.
|
# ? Jan 4, 2016 17:02 |
|
Isn't that just a regular IEC C13 connector?
|
# ? Jan 4, 2016 17:04 |
|
True, I guess I don't care if it's the Cisco approved cable since its for home. Just bought some generics.
|
# ? Jan 4, 2016 18:54 |
|
|
# ? May 31, 2024 22:18 |
|
Yeah just get a $5 one from your nearest electronics store.
|
# ? Jan 4, 2016 19:31 |