LAN's vs WAN's

[Carlos Gil]

Hello,

I apologize if these questions have been answered before. I tried looking through the forums, but I haven't been able to find the answers I am looking for. If anyone can set me straight on this, I would definitely appreciate it.

First, is "serial link" synonymous with "WAN", such that I could contrast serial connections and LAN's? If Ethernet is not a serial connection, then what is it?

Second, why do serial connections need a clock source when Ethernet does not?

Thanks.

[Jon Marshall]

Carlos

Have a look at this thread for difference between ethernet and serial -

https://supportforums.cisco.com/message/3075380#3075380

In answer to your general question about LAN and WAN yes serial connections are traditionally used to interconnect sites ie. WAN. Ethernet has been traditionally used to interconnect devices within a single building or a campus network primarily because ethernet can give much greater speeds than serial. So within a LAN yes, ethernet is the technology used.

However the lines are blurring these days in terms of ethernet. You can interconnect multiples sites together using ethernet so it is not strictly accurate to say ethernet is only LAN. Ethernet has proven to be a very adaptable technology and it continues to develop with speeds of 40 and 100Gbps on the horizon.

Jon

[Joseph W. Doherty]

Disclaimer

The  Author of this posting offers the information contained within this  posting without consideration and with the reader's understanding that  there's no implied or expressed suitability or fitness for any purpose.  Information provided is for informational purposes only and should not  be construed as rendering professional advice of any kind. Usage of this  posting's information is solely at reader's own risk.

Liability Disclaimer

In  no event shall Author be liable for any damages whatsoever (including,  without limitation, damages for loss of use, data or profit) arising out  of the use or inability to use the posting's information even if Author  has been advised of the possibility of such damage.

Posting

First, is "serial link" synonymous with "WAN", such that I could contrast serial connections and LAN's? If Ethernet is not a serial connection, then what is it?
Second, why do serial connections need a clock source when Ethernet does not?

Serial links are most often used in WAN topologies, and because of that they might be considered synonymous, but they really aren't.

What's Ethernet?  Ethernet is a mainly a L2 specification, it defines how groups of bits are "packaged" and processed.

Comparing serial links and Ethernet is somewhat like comparing apples and oranges.

The simple answer to your second question is "that's by design", a full and complete answer would be both lengthy and complex.

[Peter Paluch · Accepted solution]

Joseph,

As you mentioned, this topic is very extensive and impossible to cover in a small place like this. Nevertheless, I am somewhat reluctant to accept the following statement:

What's Ethernet?  Ethernet is a mainly a L2 specification, it defines how groups of bits are "packaged" and processed.

To me, Ethernet is very strongly specific about the Layer1 as well. In fact, most evolution in Ethernet is done primarily in the Layer1 as we move from 10Mbps to 100Gbps and beyond. The frame format has essentially not changed since the DIX Ethernet2 specification.

Comparing serial links and Ethernet is somewhat like comparing apples and oranges.

Personally, I would say that comparing Ethernet and serial links is similar to comparing Ethernet frame and UTP cabling. Serial links are a particular realization of a physical layer, but the framing can be selected - HDLC, PPP, X.25, LAPB, Frame Relay at least in Cisco routers. Because of their technical specialities, serial interfaces were best suitable for connections to modems to adapt the digital signal for transmission over longer distances. As high speed transmission over long distances was impossible at the time, the speeds of these serial technologies have never met the speed of LAN technologies. The duality of speed/reach was obvious: either high speeds (LAN) and only hundreds of meters at most between nodes in a network, or low speeds (WAN, serials) and reach up to kilometers.

Serial interface is simply a hardware component on a much lower technical level than the Ethernet which is a complete Layer1+Layer2 technology.

Best regards,

Peter

[Carlos Gil]

Thanks guys, you definitely helped clarify that for me.

Jon, from the post you linked, I read that

Each frame uses a 64 bits preamble to allow to potential receivers to sync with the frame.

It sounds as if Ethernet is similar to asynchronous serial communication because of that (56-bit) preamble. If so, this is the reason why Ethernet has no clock source, if I'm understanding correctly.

[Peter Paluch]

Carlos,

It is Jon's reply that contains the link - just to keep the credit to the rightful person.

Regarding your question - the situation about Ethernet being synchronous or asynchronous is somewhat convoluted. 10Mbps versions of Ethernet (10Base2, 10Base5, 10BaseT) were truly asynchronous. They used the 8-byte preamble (or 7-byte preamble + 1-byte start-of-frame delimiter) at the start of each frame to allow for bit synchronization of the sender and receiver.

Since the 100Mbps version of Ethernet, each station constantly keeps sending and receiving a so-called IDLE symbol if it is not sending frames. This IDLE symbol is used to keep the scramblers/descramblers on both interconnected devices synchronized, as these versions of Ethernet use a scrambling mechanism to further spread the 0s and 1s in the sent data. Also, the IDLE symbol is used to monitor that the link is still up (the Fast Link Pulses used during the autonegotiation are sent/received only in the link initialization phase, and are not sent afterwards). As a side result, the stations are constantly in bit synchronization. The preamble in 100Mbps and faster Ethernet versions is still kept for backward compatibility purposes but it is actually useless.

Note that the IDLE symbol is exchanged between each two physically adjacent PHY controllers in Ethernet - so in most cases, this is the network card on one side and the particular port controller in the switch on the other side of the cable.

Best regards,

Peter

[Peter Paluch]

Carlos,

A small addition: regarding the clock source, even Ethernet must have a clocking signal. However, Ethernet is not carrying the clock signal in a separate channel, rather, it uses self-clocking line coding mechanisms that guarantee sufficient density of signal changes, and allow to recover the clocking at the receiver.

Best regards,

Peter

[Carlos Gil]

Peter,

Very good info. Just to clarify though, 10 Mbps Ethernet wouldn't use coding mechanisms right? I'll have to research further on Ethernet clocking, because it's definitely as complicated as you've both indicated.

[Peter Paluch]

Carlos,

Just to clarify though, 10 Mbps Ethernet wouldn't use coding mechanisms right? I

I am not sure if I understand you here. The link coding is the particular procedure how to express a binary 0 or 1 (or even groups thereof) by a specific change in the carried signal. Logically, every technology that sends data over some medium must use some kind of link coding. Naive coding mechanisms like +1V = 1, 0V = 0 do not work well because of several deficiences (circuit designers and signal processing specialists would know better how to explain this - I am not so well versed in hardware), so there are more sophisticated link coding mechanisms. 10Base2 and 10BaseT use Manchester coding, 100BaseTX uses MLT-3, 1000BaseT uses PAM-5 (you may easily find all these on Wikipedia).

The scrambling is a procedure where the data being transmitted are XORed with a stream of pseudorandom bits so that the density of logical 0s and 1s runs is further dispersed, and statistically, the number of 0s and 1s in a data stream is equal. This also eliminates significantly higher RF emissions at any particular frequency possibly caused by a specific sequences of 0s and 1s in the transmitted data. Scrambling has been incorporated into Ethernet since its 100Mbps variant - the 10Mbps variants did not make use of scrambling.

Best regards,

Peter

[Joseph W. Doherty]

Disclaimer

The   Author of this posting offers the information contained within this   posting without consideration and with the reader's understanding that   there's no implied or expressed suitability or fitness for any purpose.   Information provided is for informational purposes only and should not   be construed as rendering professional advice of any kind. Usage of  this  posting's information is solely at reader's own risk.

Liability Disclaimer

In   no event shall Author be liable for any damages whatsoever (including,   without limitation, damages for loss of use, data or profit) arising  out  of the use or inability to use the posting's information even if  Author  has been advised of the possibility of such damage.

Posting

Peter, of course you're correct, Ethernet is very specific about its L1, in its many Ethernet 802.3 specification variants, but for someone asking the difference between serial interfaces and Ethernet, they might be confused by delving into such specifications; e.g. DIX framing formats vs. other Ethernet framing formats.  (Thankfully [Ethernet framing] an issue we generally don't need to contend with anymore.  Not like the good olde [Novell] days - laugh.)

A reason I described Ethernet as mainly L2, is because many, I think, involved with day-to-day networking would think of Ethernet primarily as a L2 protocol.  As you noted, evolution in Ethernet has been principally at L1, which perhaps shows the essence of Ethernet is its L2 part of its specification, but when we think of serial interfaces, we're often dealing with L1 considerations.  As you also noted, for a typical serial interface, we need to select a L2 protocol.

In any case, I do apologize is my over simplification has mislead, but both Jon's and your posts do indeed show the complexity of these subjects.