how does a t1 relate to HDLC, frame-relay, ATM, etc?

[xyyz]

This has been confusing me for a long time and I don't know exactly how to explain my confusion.

Here's a try though....

Say a company wants to install a T1 what do you do to configure it on the router? I don't recall a "T1" setting. I know there are encapsulation methods. Frame-Relay, HDLC, PPP, etc. These can be configured to give the 1.544 bandwidth defined by a t1.

So, I'm totally confused. What exactly does a T1 mean here? Is it only a way of saying how much bandwidth or how many channels you are leasing from the SP?

Damn... I really don' t know how to explain my confusion.

Okay... do you use frame-relay, PPP, frame-relay, HDLC to get the T1 working?

This isn't helping is it?

Okay... can someone explain how a t1 or t3 or whatever relates to HDLC, PPP, ISDN, frame-relay, ATM, x.25 etc?

I want to know how these all work. I've seen a bandwidth setting on a router, but I haven't seen anything about a t1. I've seen encapsulation methods used, but I haven't seen a T1 method.

Explain this someone please.

 

[Garion]

A T1 (and a T3, for that matter) is the physical line (circuit) that connects you to the outside world. It is, in essence, a very fast, sophisticated phone line. Like a phone line, you can do different types of transports across a T1. Think of using a modem, voice, and faxing. They are all ways to send information across a phone line.

HDLC and PPP are encapsulation protocols which specify how data sent from your router is encapsulated into chunks of data which can be sent across the T1 line. They are used, in general, for point-to-point circuits. HDLC is a Cisco proprietary protocol, PPP is more generic and compatible across most routers. When you setup two routers to run a point-to-point T1/T3, you need to make sure they are both using the same protocol - HDLC or PPP.

ATM and Frame Relay are also ways of sending data, but they are used for non-point-to-point circuits (called, strangely enough, multi-point circuits). They are more complicated, as they have to specify not only how data is framed, but where it's destined to go across the Frame Relay or ATM backbones. It still travels across a T1 to get out of your site, but once it's out, it goes across an entirely different communications network than a point-to-point T1.

The reason that you usually don't see anything about T1's in a router config is that routers are built to handle multiple speeds of circuit using the same interface. For example, a standard serial port (IE, Serial0 in a config) can be connected to a modem, and ISDN modem, a 56Kb/s leased line CSU/DSU , a T1 CSU and a few other things. They don't get specific. The exception is when the serial port is part of a Wan Interface Card (WIC) and has a built-in T1 CSU/DSU (The "modem" for a T1) and you have to configure it. That's getting to be more common, but it used to be that almost everyone used external CSU/DSU's and standard serial ports, so you NEVER saw anything about a T1 in the config. The most you'd ever see would be a bandwidth 1544 statement, to handle routing protocol decisions based on interface speed.

One more quick lesson.. A T1 is also called a DS1 (More often in the voice than data world).

A DS1 is made up of 24 DS0 channels, each of which is a single voice line. So, you can handle 24 voice calls per DS1. You often see T1's split into voice and data - You can "bond" together DS0 channels to become a single bigger pipe. So, if you have 12 channels of voice and 12 channels of data, you'd end up with 768Kb/s of data.

So, the heirarchy goes:

DS0 = One voice line or 64Kb/s of data
DS1 = 24 DS0's or 1536Kb/s
DS3 (AKA T3) = 28 DS1's, or about 45,000Kb/s (or 45Mb/s)

A T3 is really handled EXACTLY like a T1. It's just the physical circuit. You can split it up, you can run ATM, Frame Relay, PPP, HDLC, etc. across it to make it all work.

Not magic, just knowledge!

- G

 

[xyyz]

so in otherwards the T1 is the connection from DTE to DCE?

so what advantage do you have... not advantage.. .but when do you decide when to use frame-relay opposded to PPP?

 

[Nutz]

<< so in otherwards the T1 is the connection from DTE to DCE? >>

Not exactly. The T1 is the actual circuit from the provider to the customer. The connection from a DTE to a DCE could be from the router to the CSU/DSU or something as simple as from a computer to a hub (depending on if you even consider a hub as a CSU/DSU-like device). IMO a CSU/DSU is basically a modem (ex. a Pairgain w/T1 card).

As found at CommWeb
(Digital Signal) A classification of digital circuits. The DS technically refers to the rate and format of the signal, while the T designation refers to the equipment providing the signals. In practice, "DS" and "T" are used synonymously; for example, DS1 and T1, DS3 and T3.

T-1 A 1.544 Mbps point-to-point dedicated, digital circuit provided by the telephone companies. The monthly cost is typically based on distance. T1 lines are widely used for private networks as well as interconnections between an organization's PBX or LAN and the telco. The first T1 line was tariffed by AT&T in January 1983. However, starting in the early 1960s, T1 was deployed in intercity trunks by AT&T to improve signal quality and make more efficient use of the network.

A T1 line uses two wire pairs (one for transmit, one for receive) and time division multiplexing (TDM) to interleave 24 64-Kbps voice or data channels. The standard T1 frame is 193 bits long, which holds 24 8-bit voice samples and one synchronization bit with 8,000 frames transmitted per second. T1 is not restricted to digital voice or to 64 Kbps data streams. Channels may be combined and the total 1.544 Mbps capacity can be broken up as required. See DS, T-carrier, bipolar transmission, D4 and ESF.

64 Kbps
T-Carrier Total Speed Channels
T1 1.544 Mbps 24
T2 6.312 Mbps 96
T3 44.736 Mbps 672

Voice/Data
Service Channels Speed
DS0 1 64 Kbps
DS1 24 1.544 Mbps (T1)
DS1C 48 3.152 Mbps (T1C)
DS2 96 6.312 Mbps (T2)
DS3 672 44.736 Mbps (T3)
DS4 4032 274.176 Mbps (T4)


EUROPE (ITU)

Voice
Service Channels Speed (Mbps)
E1 30 2.048
E2 120 8.448
E3 480 34.368
E4 1920 139.264
E5 7680 565.148


SONET CIRCUITS

Service Speed (Mbps)
STS-1 OC1 51.84 (28 DS1s or 1 DS3)
STS-3 OC3 155.52 (3 STS-1s)
STS-3c OC3c 155.52 (concatenated)
STS-12 OC12 622.08 (12 STS-1s, 4 STS-3s)
STS-12c OC12c 622.08 (12 STS-1s, 4 STS-3c's)
STS-48 OC48 2488.32 (48 STS-1s, 16 STS-3s)



<< so what advantage do you have... not advantage.. .but when do you decide when to use frame-relay opposded to PPP? >>


You would need to base that decision on how your circuit terminates, how your provider has their cirucuits routed, what sort of signaling/framing your CSU/DSU requires, etc. There are many things that can contribute to how your circuit will be made.

For example, my tailend circuit is just a straigh static. Hell a simple default route does what we need it to do. But for our encapsulation we use HDLC (defaults) because that's what our provider has directed us to use (they have a 4000 series we connect to). Thats it.

 

[xyyz]

so then the connection between the provider and customer is the T1...

does encapsulation come into play with this connection... or is encapsulation something that is more end to end... say how a router in LA communicates with one in SF... NOT how the router communicates from the CSU/DSU to the CSU/DSU at the telco (this is a CSU/DSU too right)? OR do I have it wrong...

so what else needs to be configured when setting up a t1?

take it step by step starting with "my company needs a t1"

 

[Garion]

Encapsulation defines how the two routers communicate with each other - "Do we speak French or Spanish". The data is the same, it's just the language.

To setup a point-to-point T1 you first:

Get it ordered and installed between the two sites
At both ends, connect your CSU/DSU to the SmartJack
At both ends, connect your CSU/DSU to the router
On your CSU/DSU, configure your framing and other paramaters (Although in 90% of the point-to-point data T1 networks default CSU setting will work)
On your router, go into the serial interface and configure:
Encapsulation (I usually prefer PPP - More options and more standard than HDLC, but most ISP's use HDLC out of the box, since it's default)
IP address and subnet mask
Next, configure routing on both the routers - In an ISP scenario, the POP router gets a single static route back to the subnet on the Ethernet of the remote router. The remote routers just gets a default route back to the POP router.

And life should be good.

- G

 

[ScottMac]

The "T1" is the connection from the provider to the customer. You can get it in three basic flavors: Channelized, Unchannelized, or ATM. In each case, it's not so much a single line as a conduit for a number of channels. Each channel of a T1 can have some completely different thing from anyplace. The line can be provisioned such that you have twenty-four/twenty-three 64/56K lines from twenty four/twenty-three different locations that are aggregated "in the cloud" (by the carrier) and delivered to your router as a single T1. The router would have to be configured so that it sees the total bandwidth as 23/24 channels (channelized) instead of one lump of 1.544Mbps (unchannelized).

ATM and Frame Relay differ from a traditional "T1" in that it is actively switched (where traditional "T1" is point-to-point and is a dedicated circuit- a logical two-pair of copper from point A to point B)...in Frame and ATM, the bandwidth is shared to some degree. The paths through the cloud are available for use (theoretically) by any transient circuit (end-to-end circuits are created and torn down as necessary). Even a "Permanant Virtual Circuit" is likely to be set up and torn down as necessary between the two end-point FR switches. Because the carrier can share these paths (instead of dedicating them to one user/two locations), FR is usually significantly less expensive.

When you specify the encapsulation (PPP, HDLC, SDLC, Frame Relay, ATM) at the router to attach to any of the above circuits, you're basically just deciding to put your mail in a blue envelope instead of a pink one (or orange one, or green one....). There are some advantages and disadvantages to all of 'em, but they've all been around long enough that all of that's been figured out and delt with. The only real hook to any of them is that if you send your mail in a blue (pink,orange,green) envelope, you have to tell the other side what color to look for....because it can only look for one color at a time (per circuit/channel/PVC...).

With Frame Relay and ATM, the circuit has to be specifically provisioned for Frame Relay or ATM, because the carrier has to connect you to the switched cloud/infrastructure instead of the permanant-end-to-end-circuit cloud. Since ATM and Frame Relay use different types of switching infrastructure, Frame is frame, and ATM is ATM (cells)...and have to be specifically ordered and provisioned as such.

To summarize:

Frame relay - Order Frame Relay service and configure your end-routers for Frame-Relay encapsulation (most will now auto discover all the other parms)

ATM - Order ATM (T1 is the minimum provisioning chunk), config the router for ATM (not really an encapsulation as much as an arcitechture).

Point-to-point (private or to the Internet) Pick a color of encapsulation (PPP, HDLC, SDLC...) and stick with it (or follow the provisioning per the agreement with your ISP).

Easy, huh ?

FWIW

Scott

 

[Nutz]

<< so then the connection between the provider and customer is the T1... >>

That's the simplified version of what I was trying to say earlier, so yes. The T carrier designation has more relavance in how much bandwidth it provides (or how many DS-0's it is comprised of) in today's communications industry.

The encapsulation is just how the bits are placed in order before they are transmitted down the wire. But the communications device at the other end has to be set up to read the encapsulation type so it knows how to de-encap it for processing. There are different versions of encapsulation in use today (created by various manufacturers and industry standards organizations)

Just an example:

Router A in NY enaps to HDLC DSU/CSU in NY
[]---------------------------------------------------------------{}
Circuit to LA...(stupid forum and no spaces).........||
Circuit to LA...(stupid forum and no spaces).........||
Circuit to LA...(stupid forum and no spaces).........||
[]---------------------------------------------------------------{}
Router B in LA reads the HDLC signal DSU/CSU in LA converts the DSX-1 signal back to HDLC

DSU/CSU NY changes the signal to DSX-1 for long transmission and the DSU/CSU in LA converts the DSX-1 signal back to HDLC (so Router B in LA can read the HDLC signal and make any routing decisions).

The big reason for signal encoding is for transmission. Look at what happens in the OSI model:

You type an email message. SMTP takes the data and passes it to the Presentation Layer.
Presentation codes the data as ASCII.
Session establishes a connection with the destination for the purpose of transporting the data.
Transport segments the data using TCP and hands it to the Network Layer for addressing
Network addresses the packet using IP.
Data-Link then encapsulates the packet in a frame and addresses it for local delivery (MACs)
The Physical layer sends the bits down the wire.

The idea of encapsulation basically the same.

 

[xyyz]

a few point that need some clarification please...



<< Get it ordered and installed between the two sites >>


and


<< T-1 A 1.544 Mbps point-to-point dedicated, digital circuit provided by the telephone companies >>



Okay... when you say "the two sites" , I get a bit confused. I know I gave the example of connecting LA to SF... but you have alot of companies that tote a "t1" but this doesn't connect to anything other than the telco. I'm confused... when you say that it' a "dedicated point-to-point" link. What are the two points? From the office to the telco... that counts as two points? If you have a connection between two offices... how is this a point-to-point link? There's no direct connection exclusively made between the two offices... it's not like they ran wire from SF to LA for the sake of this one circuit... right? Doesn't the information flow through the telco cloud? Or is the connection something like a static route?



<< On your CSU/DSU, configure your framing and other paramaters (Although in 90% of the point-to-point data T1 networks default CSU setting will work) >>



What's the default CSU/DSU setting? Does it have it's own protocol? So it doesn't use PPP or HDLC? What does it use then?

HDLC and PPP are used only for communication between the two DTE's? So what happens to the data once it leaves the DTE and goes to the DCE and out the DCE.... and yeah what happens to the data once it leaves the DCE.... how does it travel... what happens to it when it passes through the telco network?



<< The line can be provisioned such that you have twenty-four/twenty-three 64/56K lines from twenty four/twenty-three different locations that are aggregated "in the cloud" (by the carrier) and delivered to your router as a single T1. The router would have to be configured so that it sees the total bandwidth as 23/24 channels (channelized) instead of one lump of 1.544Mbps (unchannelized). >>



Why is it that you can have one extra 64k channel compared to one less 56k channel? What happens to the remaining bandwidth on the 56k channel connection?

How do you allocate each channel? Can you point me to cisco documentation? Where can I find the option to allow you to designate certain channels for voice and others for data? Is this done in the router... or does some other equipment handle it?

You also speak of how the 24/23 lines can come from 24/23 different locations. I don't understand this... can you explain this further?



<< That's the simplified version of what I was trying to say earlier, so yes. The T carrier designation has more relavance in how much bandwidth it provides (or how many DS-0's it is comprised of) in today's communications industry. >>



What the complex version then? I'd like to hear it... it might lessen my confusion.



<< DSU/CSU NY changes the signal to DSX-1 for long transmission and the DSU/CSU in LA converts the DSX-1 signal back to HDLC (so Router B in LA can read the HDLC signal and make any routing decisions). >>



DSX-1? What's that and what's a "long transmission"?

 

[ScottMac]

Regarding the 56K / 64K thing:

As mentioned, the T1 (or fractional part thereof) can be ordered as channelized or non-channelized. The channelized flavors use what's called 'Channel Associated Signalling" (CAS). If you use CAS, some part of the channel's bandwidth (about 8K) is used for signalling for that channel only.

Note that when we talk about "Channels," what we are actually talking about is a time-slot. From the beginning of one chunk of time, to the end of the same chunk, there are 24 fully synchronized time divisions. Each of those time divisions represent a "channel."

If the T1 is provisioned as a "fat pipe" (i.e., unchannellized), then one of the 24 available channels is designated to do the signalling for all the other (non-channelized) bandwidth. This is comparable to the US/Japan version of a PRI (like an ISDN flavor of T1) of 23 bearer channels + one signalling channel (23B+D is the usual label).

You can also have channelized T1 with 64K "clear channel" bandwidth, it's all in how it's provisioned.....typical is ESF framing with B8ZS. B8ZS is the setting/protocol that allows usage of the full 64K channel.

The CSU/DSU's on each end CAN be totally different settings, because the infrastructure of the cloud can re-arrainge things somewhat....but you'd be flirting with disaster: it's unlikely that the routers on each end could deal with the re-arrainged/reformatted data. About as extreme as you'd want to get would be to move the channel assignments of a frac-T1 to accommodate integrating some new circuits with existing....all that can be done in the cloud.

Regarding the point-to-point stuff: Even the "T1 to the Internet" that most companies have is point-to-point. If you wanted a T1 to the Internet, you wouldn't call the phone company and ask for a T1, because they're gonna wanna know where the other end is gonna be. That's why T1's to the Internet are ordered from the ISP...because after you give the ISP your address (along with a big check), THEY call the phone company, because THEY KNOW where both ends of the T1 are s'posed to be. Unless you are a special class of Internet-connectee, you will also need mail forwarding, DNS, and (perhaps) a multitude of other services as well as physical connectivity.

If you wanted to connect the Main Office to the "Branch Office", THEN you call the phone company (or whatever carrier) for a T1, because you know where both ends are s'posed to be.

SO, there ya go......

FWIW

Scott

 

[xyyz]

I have more questions on this comming soon...

but in the meanwhile can someone please suggest a good WAN book that I can use to prepare for my CCNA? I've been doing horribly this semester because semster 4 of the cisco academy deals with WAN's... and... well the material they have on WANs SUCKS... at least for me.

I need something that will cover WAN tech, PPP, ISDN and Frame Relay.

 

[xyyz]

bump.... comeon... I need the info... :/

 

[spidey07]

try searching cisco.com. A wealth of information there.

 

[xyyz]

<< try searching cisco.com. A wealth of information there. >>



I did... I found the Documentation CD link... but that's where they pull the academy text from... so it's not too much help.

 

[spidey07]

search the RFCs?

 

[Nutz]

Check my post here.

For starters, the Sybex books written by Todd Lammle are great. Also check out what Cisco Press has to offer. I highly recommend something like Newton's Telecom Dictionary . If you really want to get deep into it, then check out Computer Networks by Andrew S. Tanenbaum

Other must reads are:
Advanced Ip Network Design : CCIE Professional Development
by Don Slice

TCP/IP Illustrated, Volume 1: The Protocols
by W. Richard Stevens

Routing TCP/IP, Volume 1 : CCIE Professional Development
by Jeff Doyle

A couple of others that might do you good to check out are...
Practice of System and Network Administration, The
by Thomas A. Limoncelli

Lan Wiring
by James Grayson Trulove

For books in IT, the best place I've found to date is Bookpool. Their prices beat Amazon by a longshot!

 

[xyyz]

<< search the RFCs? >>



these are way too specific and complex for what I need... I just need someting on WAN tech, PPP, ISDN and Frame-Relay... that will give a detailed overview.

 

[Nutz]

Internetworking Technologies Handbook (Cisco Press)
ISBN 1587050013
http://www.bookpool.com/.x/64nwc26vq0/ss/1?qs=internetworking+technologies

This book has everything you just asked for and more!

 

[xyyz]

<< Internetworking Technologies Handbook (Cisco Press)
ISBN 1587050013
http://www.bookpool.com/.x/64nwc26vq0/ss/1?qs=internetworking+technologies

This book has everything you just asked for and more! >>



Funny that you mentioned this... I pulled it from my shelf last night... but it is a duplcaite of what's on the Cisco Documentation CD.