Multisite Distributed VoIP Deployment

I briefly went over the Multisite Distributed VoIP Deployment section in the SRND Guide. It has very similar attributes and requirements as the centralized model except it is a lot more versatile due to having a call processing cluster at each site. That eliminates the need to depend on the WAN or whatever media of inter-site connectivity for calls as much. I also began reading about the Cisco Intercompany Media Engine which seems to dynamically interconnect different enterprise organization phone numbers and create a network similar to how internet was when it was first created. Instead of just IP services, the network is used to connect Unified Communications within all the Enterprises. There's a lot more to learn about this technology and I may be off a little bit with my understanding of it at this point in time.

VoIP Site Based Design Models

I spent sometime this morning reading through the various VoIP deployment models and how to implement high availability with each model while trying to get over this cold I currently have. So far I only went through the Campus model and the model that shows centralized call processing with multiple sites. In my current work environment we currently use the latter which is a centralized site that provides all the call processing which is then distributed to our multiple remote call centers. We're working on converting this into a more highly available solution which is a distributed model.

I plan on reading up on the other site based design guides that shows the distributed model along with a few others. I'm starting to see that there are multiple ways to provide redundancy using tools such as SRST (Survivable Remote Site Telephony), Distributed Communications Manager Express (CME), and many others.

Finally moving back to the VoIP World

I haven't posted in a while but I finished up the TCP/IP Vol 1 book going through the subjects that I feel I needed to review before going back to VoIP. Mainly the interior routing protocols and overview regarding redistribution and basic static routing. This was all for making sure I had foundation for building on my VoIP knowledge to tackle the CCNP: Voice and maybe the CCIE: Voice. How can you sufficiently route voice over IP packets if you don't know how to route IP packets on its own?

I feel comfortable enough with IP routing to start diving deep into the Voice world and I have. It doesn't appear that the new CCNP: Voice reading material is quite done yet so I'm starting with Cisco Solution Reference Network Designs (SRND) material. These are really good free reading material providing best practices and design for all types of network solutions from security, voice, routing/switching, wireless, etc. Also moving to a new place leaves you pretty broke for a while and I know I'm going to have to drop some $$$ for additional lab equipment, books, and the 4 test to become CCNP: Voice certified. So I'm saving up money now while I scavenge the internet for free material!

I started with the Cisco 8.x CUCM SRND last week and I'm already at page 148 (1356 pages total) putting in about an hour of study a day.

Today I started on the security section for both the access layer and phone level. Talks about many basic things such as Man in the Middle Attacks, port security, disabling phone PC ports, etc. Tomorrow I'll read through material regarding Access Lists (ACLs) in a Unifed Communication environment.

End of the EIGRP Chapter (TCP/IP Vol I)

I cruised through the EIGRP chapter which refreshed my memory on a bunch of topics that escaped me. Such as the default metric calculations used when calculating the best routes along with topics such as Feasible Successors, SIA (Stuck-in-Active), and k values. I went through the configuration exercises which were surprisingly not that complex considering the book is supposed to help prepare you for CCIE. All they asked were to configure 5 routers with EIGRP using the process-id 5, create a authentication key between router A and B, and summarize the routes as much as possible.

Tomorrow I'll knock out the troubleshooting section and report back on what that was like time permitting. This has been one of the most helpful chapters so far since we still use EIGRP in our network but not near as much since migrating to a different vendor which uses strictly BGP for connecting WAN's.

RIPv2 and EIGRP

I briefly went through the RIPv2 review questions,configuration labs, and troubleshooting exercises over the last few days in the TCP/IP Vol 1 book. I also went through the basics and beginning of IGRP and how it moved to EIGRP. I will be spending the next week or so going through the this chapter so you may not see a post from me until then when I reach the configuration exercises. In other news I'll be moving soon so you really may not see a post until about 3 weeks or so due to the craziness going on right now!

RIPv1 TCP/IP Vol I Config Exercises

Today's lab involved RIPv1, 6 routers, 2 switches, and 4 configuration tasks. All together the lab was around an hour. The first part was to simply configure the network for RIPv1, simple enough. It's nice not having to mess with static routes all over the place! The next task was make the serial link between RTC and RTD send unicast updates rather than broadcasts. I did this by configuring the serial interfaces as passive along with setting up RTC/RTD with neighbor commands.

Next up was too change how often RTC and RTD sent each other updates, this was done with the timer basic command. I picked some random timers and used the ? command to help me a little. I set this across the board on all 6 routers because from what I understand the timers have to be universal across the network in order to work., someone correct me if I'm wrong though. The last task was to prevent RTA from talking to the 192.168.4.0 network and RTB from talking to the 192.168.5.0 network. I used an access-list and the off-set command. I told both routers that on any incoming RIP update for the particular network listed in their access-list, set the hop count to 16 or invalid. 16 hops is considered an invalid/infinite route with RIP.

RIPv1Characteristics

I went over the beginning pages in the Vol I book regarding the RIP version 1 protocol. There are some interesting things that should be noted about version 1 compared to version 2. I'll list what I remember:

1. Version 1 is a classful protocol rather than a classless protocol.
2. Since it's classful RIPv1 can become problematic when running the protocol over a discontiguous network.
3.RIPv1 pass the entire network class that it knows about when advertising routes rather than the specific subnet unless the subnet is directly connected. Again this points back to numero uno.
4. The only subnets it will list in the routing table are the subnets that's directly connected to the router.
5. RIPv1 assume any subnets that it knows about will be the same subnet size throughout the entire network. Variable subnetted networks could easily cause a lot of problems.