Serial Cable Speed

[jake052399]

In my home lab I have 2 2501's connecting to each other through a back to back serial cable to simulate a WAN between two separate LAN's.

My question is what is the fastest speed I can get out of this sort of connection. Is this set by the clockrate? I would like to share the Internet connection from the one side.

Thanks

[Richard Burts]

Jacob

The speed of the connection is directly controlled by the clock rate and is indirectly controlled by the kind of cable inteface that is used. The serial interrface of the 2501 supports several cable standards. If the cable is RS-232 you will have difficulty getting speeds much above 64 K where if you have a V.35 cable you can get speeds of T1 or E1.

HTH

Rick

[stephtchoko · Accepted solution]

Hi,

the fastest speed is 4000 000 bps (4Mbps).

I confirm thatthis fast speed is set by clock rate.

GN-2500-YDE(config-if)#clock rate ?

Speed (bits per second)

1200

2400

4800

9600

19200

38400

56000

64000

72000

125000

148000

250000

500000

800000

1000000

1300000

2000000

4000000

<300-4000000> Choose clockrate from list above

GN-2500-YDE(config-if)#end

http://www.cisco.com/en/US/tech/tk713/tk317/technologies_configuration_example09186a00800944ff.shtml

Regards.

[asifarham383]

Serial cable speed refers to the data transmission rate of a serial communication link, which sends data one bit at a time over a single channel. It is usually measured in bits per second (bps), such as 9600 bps, 115200 bps, or higher, depending on the cable type and hardware. Serial cables are commonly used to connect devices like routers, switches, and computers for configuration or data transfer. Compared to parallel communication, serial communication is slower in sending individual bits but more reliable over long distances because it reduces signal interference and timing issues.

[Joseph W. Doherty]

I suspect, as did @Richard Burts , OP was really asking about a WAN serial cable, like, as Rick specifically mentioned, a V35.

Further, as @stephtchoko describes, yes link speed/bandwidth depends on the clock rate setting (and I also vaguely recall the max being 4,000,000 Mbps, on 2500s).  Although, not mentioned in his reply, but it's within the Cisco reference, clock rate is only configured on just one end of the link.

Compared to parallel communication, serial communication is slower in sending individual bits but more reliable over long distances because it reduces signal interference and timing issues.

It has nothing to do with speed to sending individual bits or reliability across distance.

Serial vs. parallel boils down to whether you have just one transmission path or multiple paths.  More or less, you can increase throughput, linearly, as you add concurrent paths, assuming each additional path offers the same bandwidth.  The problem with multiple paths, media often increases in cost too; likely at a less than lineral rate, but as distance increases, the incremental costs can become substantial.  Further, other physical considerations may arise.

Practically, it's aways trade off between performance vs. cost.

Not disregarding what I noted, above, media choice, design, modulation rate, etc., do often differ much between serial vs. parallel, but that's not because they have to differ, but a specific application of either, makes more sense doing serial or parallel a specific way.  For example, RS-232 is serial, V35 is serial as is 100G Ethernet.  Or, decades ago, a parallel Centronics printer cable (1.2 Mbps), PCIe v.7, 16 lanes, almost 2 Tbps.

[Elliot Dierksen]

I'll add my way back machine recollections to this. With Cisco gear, the device that has the DCE (data communication equipment) cable is the one that sets the clock rate. The DTE (data terminal equipment) recovers the clock from the DCE. I believe there is something to do with pin-outs that the DCE/DTE needs, but it has been a decade or more since I dealt with a serial WAN connection. In gear other than Cisco, the DCE can recover clock from the DTE. In general though it would make sense that the DCE (provider side) is the one to supply the clock.

[Joseph W. Doherty]

@Elliot Dierksen , yes, normally the Telecom's device, the DCE, provided the clock, and customer's router, like a 2500, the DTE, accepted the clock signal.  However, in a lab, like for two described 2501s, back-to-back, one, and only one, needs to provide the clock signal.  Doesn't matter which; don't recall if a V35 had a specific DCE vs. DTE ends, but vaguely recall it didn't.

[Elliot Dierksen]

Yes, @Joseph W. Doherty the V.35 cables are either DCE or DTE. There are (or were) special single cables you could order that were DCE one side and DTE on the other side. I still have some of those in my lab.

[Joseph W. Doherty]

Laugh, it has been over a decade since I've seen a V35 cable, and that was back when I was a PT instructor teaching college network classes, actually using a pair of ISRs (low end 2800 series), back-to-back.  So, I was just reviewing V35 cables, connectors and pin outs.

A router would normally have DTE sockets while CSU/DSU would have DCE sockets, so a typical V35 connector would have also have a DTE plug on one end and a DCE plug on the other.  In other words, a standard V35 cable could not be used between a pair of routers.

However, I recall if you used a gender changer on the DTE plug, creating a DCE<>DCE cable, the two back-to-back routers worked fine (again after setting one to generate the clock).

The above is also similar to when using serial RS232 connectors, which also used DTE and DCE conversations.

Lastly, although copper Ethernet may use RJ45 as same connector on both ends, and doesn't generally use DTE/DCE terminology, a similar issue is being addressed by what kind of device used MDI vs. MDI-X sockets, and having straight through cables vs. cross-over cables (although most of those issues have be eliminated by auto MDI(X) sockets).