6

I know that cable connection (DSL) is asymmetric, there is a certain bandwidth on the channel, and for a better user experience, they allocate a bigger portion to the downstream. I suppose Ethernet cables have a certain bandwidth as well, or fiber-optic cables.. Why are they symmetric?

  • 1
    The DSL allocation is purely arbitrary and mostly assymetric to only require cheaper hardware at the end user, and because that is the most common use case. You can always get symmetric DSL. – PlasmaHH Jul 21 '18 at 16:26
11

The very first Ethernet standards were designed for using a single, shared coax wire. There's no benefit in making the connection asymmetric in any way.

Additionally, network nodes are generally viewed as equal peers from the technology design perspective (in contrast to specific network design). There's simply no motive in making a general-purpose network technology not symmetric.

ADSL had a completely different design goal - it was intended as a high-speed distribution network link using legacy cabling. There's are clear uplink/downlink roles and it's no general-purpose network link.

The roles in xDSL have changed somewhat over time and VDSL 2 can actually adapt to that, featuring different downlink/uplink bandwidth allocations.

8

Q: Why is ethernet symmetric (when compared to ADSL's asymmetry).
A: This is best explained by looking the technical reason for ADSL's asymmetry.

Quoting directly from wikipedia:

There are both technical and marketing reasons why ADSL is in many places the most common type offered to home users. On the technical side, there is likely to be more crosstalk from other circuits at the DSLAM end (where the wires from many local loops are close to each other) than at the customer premises. Thus the upload signal is weakest at the noisiest part of the local loop, while the download signal is strongest at the noisiest part of the local loop. It therefore makes technical sense to have the DSLAM transmit at a higher bit rate than does the modem on the customer end. Since the typical home user in fact does prefer a higher download speed, the telephone companies chose to make a virtue out of necessity, hence ADSL.

In short, ADSL's asymmetry reduces cross-talk at the DSLAM, where a large bundle of cables terminate close together. If the designers choose a faster customer upload speed, they would need a stronger signal to increase signal-to-noise ratio (thus increasing cross-talk at the DSLAM).

Ethernet was designed for a completely different purpose and doesn't need the asymmetry. Ethernet isn't restricted to a residential market with legacy cabling; furthermore, asymmetry in ethernet would needlessly complicate many ethernet designs.

4

Ethernet is a local area network technology, and it was designed with no assumption that traffic will be higher in one direction than the other.

It was developed long before the web or Internet was available.

4

The cabling for ADSL uses telephone grade cabling and upstream and downstream traffic share the same copper pair. As there is limited bandwidth, the designers chose to give the downstream more capacity to make more efficient use of the bandwidth, something that fits the usage patterns of home users. ADSL was designed for a specific use case – to connect home users to the Internet.

Ethernet on the other hand was designed to be a universal standard to support many use cases. The device on the end of an Ethernet cable could be a PC, server or even a device running peer-to-peer software. You can't make any assumption as to whether the device would need more sending or receiving capacity, so providing symmetric capacity makes Ethernet a far more universal standard.

Early Ethernet was half-duplex. Not setting limits on the send and receive bandwidth made it a lot more flexible. Devices could use as much send or receive bandwidth as there was available. It also operated on a shared multi-access cable, setting upstream and downstream limits in this case doesn’t make sense. Modern Ethernet is now full duplex, so the physical medium itself provides an efficient symmetric data path without the overhead and complexity of CSMA/CD.

  • 1
    I beg to disagree with the half-duplex statement - ADSL uses frequency-division multiplexing and is generally full-duplex capable. – Zac67 Jul 20 '18 at 20:00
  • The medium is half-duplex, it is a shared bus, only 1 pair of wires. I see your point though – Karl Billington Jul 20 '18 at 20:02
  • Recheck G.922 - it's full duplex. You could call TDM half-duplex but not FDM. Completely different technology (hybrids and echo compensation) but each 1000BASE-T lane and 100BASE-T1 use single pairs as well and are undoubtedly full duplex. – Zac67 Jul 20 '18 at 20:10
  • 1
    I was referring to the same physical cable being shared for upstream and downstream, perhaps half-duplex isn't the correct term in this case so reworded – Karl Billington Jul 20 '18 at 20:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.