1. Home
  2. Cisco Systems
  3. Network Clock question

Network Clock question

Here's a question regarding the NO NETWORK-CLOCK-PARTICIPATE command.

I have multiple T1s to multiple sites. All of them come back into a central 3840 router. Some links are timed via telco-provided clock (loop timing). Some links have no timing and thus have the CLOCK-SOURCE INTERNAL on that controller running, and on the distant end of that particular link, the controller is loop timed.

Now, the NO NETWORK-CLOCK-PARTICIPATE command, in which scenarios should this be used? As I understand it, the command causes the wic card to use its own timing, rather than the network. The positive form of the command means it will use the network for timing - which I take it that a site set up with the NETWORK-CLOCK-PARTICIPATE WIC 0 command should use the recovered timing signal (loop) from my central router which is participating with its own clock having been set up with NO NETWORK-CLOCK-PARTICIPATE command.

Looking at traffic pattern graphs, the ethernet traffic on the router mirrors the serial interface errors which is indicative of the clock on the that interface having sync issues.

My root question is, within a hub and spoke network, where all your sites are recovered timing and having the central site router providing timing via its internal clock, what should each site router and what should the central router be set to with relation to the NETWORK-CLOCK-PARTICIPATE command? My assumption is the sites should all me NETWORK-CLOCK-PARTICIPATE and the central router should have the negated form of the command.

-pooch

Reply to
poochworx
Loading thread data ...

Your assumptions are all wrong. Each router needs a timing source for T1's. The timer on a router is not stable enough to provide this timing, so you need to recover this clock from at least one T1 on EACH router. If one side is loop-timed and the other end is internal-timed than I can pretty much guarantee that the internal-timed router is going to have issues. Why, because the carrier is NEVER going to use you end of the link for timing. Now if you have 2-T1's from the same carrier on the router, you can pick one of them to recover the clock and the other can be internal timed. Why? because the two T1's are getting clock from the carrier equipment so they are timed correctly on the carrier end, and since the carrier is providing time from the same source, the other end only needs to recover clock from one of the T1's.

Scott

Reply to
Thrill5

Hi Pooch,

Its been a long long time (15+ years?) since I have been involved in provider networks that need the user to manage the clock settings, however some things never seem to change when considering this an an option. My base rules for this scenario say - 1. For each SINGLE Data Circuit (be it point-to-point or multi--drop), use 1 and 1 only Clock source point. All other points for that Data Circuit should use LOOP or RECEIVER derived clocking (and for multi-point the MASTER site should be the clock source). The golden rule is - Wherever possible use a NETWORK derived clock as the Master Clock, not a clock source from your own device. The Network clock is much more accurate and stable than your device. So in this case the RECEIVED Clock Source becomes the MASTER source for this Data Circuit.

  1. The ONLY time one needs to consider other Clock Source options is when you are using TDM (Time Division Multiplexing) environments with Tail circuits attached, in which case then your Network device HAS to be the source point, for all clocks but then I haven't seen a scenario like this for many many years now.

Cheers...............pk.

Reply to
Peter

Im still a bit confused, Im familiar with the timing and associated issues using the network for recovered timing. However, im still not quite getting the NETWORK CLOCK PARTICIPATE COMMAND as it relates to teh recovered timing and internal timing.

I knew each t1 needed a clock and usually, the clock is recovered, however, there are situations where the customer needs to provide clock.

So, my question originally is still about the NETWORK CLOCK COMMAND as it realtes to the internal and external timing for a dual t1 wic card.

Reply to
poochworx

I can't give you an authoritative answer however I have found out a few things about this recently.

The key thing is that the more recent T1 and E1 cards, when installed in a more recent chassis, have additional functionality. These cards are the MFT or Multi-Flex Trunk cards. They do individual DS0 channel pass through, in real time, between different T1/E1 ports.

My view is that this makes the router able to provide some of the services that might be expected from a Telco.

What I know:- If you have a MFT card in a router and are using it as a PRI then you MUST configure

network-clock-participate wic 0 !

Reply to
Bod43

That makes sense. I did more research as well. The ISR routers do indeed provide more services on the T1 cards. I pulled this from CISCO's site:

*************************** Usage Guidelines This command is used for ATM segmentation and reassembly or digital signal processing and Cisco 3660, Cisco 3725, and Cisco 3745 routers.

This command applies to any network module with T1/E1 controllers to provide clocks from a central source (MIX module for the Cisco 3660) to the network module and to the port on the network module. Then that port can be selected as the clock source with the network-clock-select command to supply clock to other ports or network modules that choose to participate in network clocking with the network-clock-participate command. This command synchronizes the clocks for two ports.

On the Cisco 3700 series, you must use the network-clock-participate command and either the wic wic-slot keyword and argument or the slot slot-number keyword and argument.

-------------------------------------------------------------------------------- Note If the AIM takes its clock signals from a T1 or E1 controller, it is mandatory to use the network-clock-select and network-clock-participate commands for ATM. The clocks for the ATM and voice interfaces do not need to be synchronous, but improved voice quality may result if they are.

-------------------------------------------------------------------------------- Note The only VWICs that can participate in network clocking are digital T1/E1 packet voice trunk network modules (NM-HDV), and Fast Ethernet network modules (NM-2W, NM-1FE. and NM-2FE).

************************************

So, with all my circuits I just have to figure why only two of teh links are slipping and the others arent.

Reply to
poochworx

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.