00:52.55 | plarsen | jsmith-away, <plarsen> ~ask |
00:52.55 | plarsen | <javabot> The Ask To Ask protocol wastes more bandwidth than any version of the Ask protocol, so just ask your question. |
00:52.56 | plarsen | lol |
01:15.12 | GGD | hey plarsen |
01:15.18 | GGD | jsmith-away, |
01:15.23 | plarsen | greetings |
01:15.34 | GGD | and salutations |
01:16.21 | GGD | how goes it excited for summit? |
01:18.25 | plarsen | Sure - if I get to go i am ;) |
01:18.30 | plarsen | Limited space for RH'ers .... |
01:18.47 | GGD | oh? |
01:18.58 | plarsen | But I like a lot of the stuff that will be going on there. Pretty exciting stuff. |
01:18.59 | GGD | i figured that it would be crawling with RH'ers |
01:19.17 | plarsen | There'll be plenty of RH'ers - but the whole company isn't going ;) |
01:20.32 | GGD | ahh.. if it did who would mind the store? |
01:23.47 | plarsen | lol - something like that yeah |
01:24.54 | GGD | lol |
01:31.09 | GGD | hey jsmith |
01:31.16 | jsmith | Howdy GGD |
01:31.22 | GGD | :) |
01:31.30 | GGD | how are things in your neck of the woods |
01:34.01 | GGD | question for you |
01:34.06 | GGD | in regards to VOIP |
01:35.10 | GGD | how do you put a "safety net" on Voip when you have a whole bunch of switches? |
01:36.12 | plarsen | GGD, define safety net |
01:36.51 | GGD | well if the power goes out in the building how does one dial out on VOIP? |
01:37.13 | GGD | with a regular line the line is still live.. unless i mis understood someplace |
01:37.26 | GGD | dial out meaning 911 etc |
01:38.04 | plarsen | GGD, you need to do what "they" do with the regular phone lines - battery backup |
01:38.12 | plarsen | GGD, switches are part of that. |
01:38.26 | plarsen | GGD, keeping all equipment in a central location will of course help. |
01:38.37 | plarsen | GGD, make sure your uplink demarkation is battery backed up to. |
01:38.39 | plarsen | too* |
01:38.57 | jsmith | Power over ethernet can help |
01:38.58 | GGD | true |
01:39.07 | jsmith | Having an analog phone line is a cheap insurance policy, though |
01:39.12 | GGD | but there are like 500-600 switches i think |
01:39.15 | GGD | yea |
01:39.26 | GGD | 400-500 rather |
01:40.01 | plarsen | GGD, each NOC should be battery backed up. You could choose that only part of the network should be operational in case of a power failure - the answer is still the same - UPSes |
01:40.27 | plarsen | And as jsmith said, make sure your phones are POE enabled. |
01:40.43 | GGD | agreed |
01:40.45 | GGD | plarsen, |
01:41.28 | GGD | but your phone would be connected to a regular 24 port managed switch right? |
01:42.02 | plarsen | GGD, providing POE and being connected to a UPS, yeah the run of the mill switch |
01:42.23 | plarsen | GGD, although I would imagine most switches being managed or at least allow for VLAN segmenting. |
01:42.28 | GGD | yes |
01:42.39 | GGD | correct |
01:43.08 | plarsen | The good news is you can connect a lot of switches to a single consumer grade UPS |
01:44.27 | GGD | so if you had the swtich in the room and had that connected to an ups the phones in the room can be also connected to it? |
01:45.41 | plarsen | If the switch does POE you just backup the switch |
01:45.53 | plarsen | Anything you connect to it gets power from the switch |
01:46.02 | GGD | ahh |
01:46.08 | plarsen | Remember to use the power consumption of all the connected phones when you dimention the UPS size ;) |
01:46.11 | GGD | thats alot of ups's |
01:46.24 | GGD | 5-10 phones? |
01:46.30 | plarsen | If you have that many switches you have a looooooot of phones |
01:46.33 | plarsen | 2000+ |
01:46.34 | plarsen | ? |
01:46.35 | jsmith | PoE switches draw power :-) |
01:46.52 | GGD | 5-10 phones per switch |
01:47.32 | plarsen | That's still 3000-5000 phones. I think 10-15 server grade UPSes aren't bad if the premise is thta you cannot afford to be without phones in case of a power failure |
01:47.35 | jsmith | has a 24 switch, with 12 of the ports being PoE |
01:47.44 | plarsen | And last time I saw a SIP phone boot it took minutes ;) |
01:48.07 | plarsen | yeah - get larger switches. 24 ports is small today. The more ports the better. |
01:48.40 | jsmith | plarsen: Depends on the phone. Polycoms take minutes. My Linksys phone takes less than 5 seconds. Guess which one has a firmware based on Linux? |
01:48.57 | plarsen | GGD, and make sure you do not just connect the switches in serial. Daisy-chain them in a loop or you're going to regret it very fast. |
01:49.09 | plarsen | lol |
01:49.17 | plarsen | I saw an upgrade being pushed out at work once. |
01:49.21 | GGD | all of them go to a larger switch |
01:49.22 | plarsen | They were out 15+ minutes |
01:49.39 | GGD | which goes to a core type switch |
01:50.47 | GGD | does that make sense? |
01:51.32 | plarsen | GGD, again, depends on the configuration. If you have that many switches one or two will be out. You should not depadize the whole network just because a switch gives up |
01:52.11 | GGD | ok... |
01:52.16 | plarsen | There's a switch protocol which name is escaping me, that is meant to deal with loops, so that if you loose one switch another route is automatically picked and no connectivity is lost. |
01:52.18 | jsmith | Good ol' spanning tree protocol... |
01:52.25 | plarsen | :) there we go. Thanks jsmith |
01:52.27 | GGD | stp |
01:52.41 | jsmith | learned one or two things in his former life as a data center junky |
01:52.46 | GGD | LOL |
01:53.02 | GGD | *bows* to the all great and powerful fedora god |
01:53.13 | jsmith | I ain't all that and a bag of chips |
01:53.26 | GGD | i just wanted to throw that in there |
01:53.29 | GGD | you know |
01:54.18 | GGD | gosh i would love a white board to diagram all this out *sighs* |
01:56.55 | GGD | i mean i think we can do this the only question that was weighing was the power issue |
01:57.33 | plarsen | GGD, make it a business issue. Find out what the cost is for being without power and use that to justify the solution |
01:57.45 | plarsen | GGD, or convince yourself that it's not that big of a deal ;) |
01:58.19 | jsmith | It's all about risk management, not playing "whack a mole" trying to make everything redundant |
01:58.21 | GGD | we have power outages frequently in the buildings |
01:58.36 | GGD | business issue? i don't know what you are meaning plarsen |
01:58.51 | plarsen | GGD, business cost of having no phones |
01:58.58 | GGD | whack a mole? gosh haven't played that in quite a while |
01:59.22 | GGD | thats a total deal breaker plarsen as we use them as well as email etc to communicate |
01:59.40 | GGD | i wanted to tie the two together in a more seamless way |
02:00.12 | plarsen | GGD, you still need to quantify. It makes the relative cost of the UPSes and maybe some labor time in moving thigns around more viable and defendable. |
02:00.58 | GGD | in what way? |
02:01.41 | GGD | gosh i feel so much like a noob |
02:01.55 | plarsen | GGD, let's say you calculate the cost per hour of no phone for your company is $100K - if the cost of getting UPSes and proper rendundancy is $150K it's paying for itself the second time you have a power outage. Those are things the C-guys understand. |
02:02.33 | GGD | ahh |
02:02.45 | GGD | all though you are paying 50k more |
02:02.55 | GGD | you are saving 75k when the power dies |
02:02.59 | plarsen | No - after the 2nd power outage, it's FREE :) |
02:03.01 | GGD | 50 rather |
02:03.08 | GGD | ok..... |
02:03.22 | plarsen | You can actually put it up as saving the company money which is always a good selling point |
02:03.52 | GGD | and since the backend is mostly if not all open source.. |
02:03.57 | GGD | that saves money too |
02:04.01 | plarsen | GGD, make sure you get the operational side of the UPSes in there too. batteries needs to be maintained, and you most likely need to invest in a management system too. |
02:04.24 | GGD | understood |
02:05.02 | plarsen | GGD, server grade UPSes are all networkable. I would definitely recommend going that route even though they charge like $150 for a NIC for those things (well, there's actually a computer on the NIC but that's a different issue) |
02:05.39 | GGD | so that when a battery acts up you have an alarm that goes off... |
02:06.14 | plarsen | GGD, that and you have maintenance data on your batteries. You know when to replace them; how much time you have on battery (that's another aspect you need to put into your calculations) |
02:06.39 | plarsen | GGD, goto the apc site. They have online calculators that basically allow you to say: Wattage and uptime required = size of UPS |
02:07.09 | GGD | gotcha |
02:07.55 | GGD | i can do that |
02:07.57 | plarsen | There's a lot of power management solutions out there too. |
02:08.04 | plarsen | Depending on how far you want to take it. |
02:08.24 | plarsen | With a central IP management console, you can check and verify the status of all the UPSes in your ship - even if they're accross the country |
02:08.39 | GGD | its sorta like ebay... there is always someone selling something out there :) |
02:08.41 | GGD | yes |
02:08.48 | plarsen | You can schedule maintenance, calibration and know when batteries will not be able to fulfill your SLA |
02:08.49 | GGD | making it easier to manager |
02:09.00 | GGD | manage |
02:09.52 | plarsen | Yes - APC of course sells their own solution with call-home features and all kinds of funky stuff. But you can use apcupsd too. |
02:12.51 | GGD | true |
02:37.44 | GGD | gosh its been a long day |
02:38.05 | GGD | time to get for bed thank you plarsen and jsmith for helping me through this |
02:47.20 | *** join/#fredlug ggd_droid (~AndChat@18.sub-174-252-106.myvzw.com) |
02:47.36 | ggd_droid | Test test |
02:51.42 | *** join/#fredlug plarsen (~plarsen@pdpc/supporter/professional/plarsen) |
13:36.24 | *** join/#fredlug plarsen (~plarsen@pdpc/supporter/professional/plarsen) |
17:38.28 | *** join/#fredlug ggd_droid (~AndChat@222.sub-174-252-118.myvzw.com) |
17:43.12 | ggd_droid | Hey smith |
17:43.54 | jsmith | Howdy howdy |
21:54.35 | *** join/#fredlug plarsen (~plarsen@pdpc/supporter/professional/plarsen) |