What's the speed of gigabit ethernet?

workfrom925workfrom925 Member Posts: 196
Obviously the speed of gigabit Ethernet is gigabit, 1,000,000,000 bits/second. However, in which direction? Since most hardware nowadays support full duplex, does it mean, if two hosts are sending each other data, the combined speed is 2,000,000,000 bits/second over the link?

Comments

  • Node ManNode Man Member Posts: 668 ■■■□□□□□□□
    I'll take a stab at this one:

    Ethernet uses 3 wires in each direction. 1 to 3, 2 to 6 and 3 to 1. (google cisco cross over pin outs). However i understand that modern equipment uses autosync (is that the right name?) So I am going to go out on a limb and say its a gig in each direction.

    I hope Im right icon_smile.gif
  • ptilsenptilsen Member Posts: 2,835 ■■■■■■■■■■
    Yes, it's a gig in each direction while running at full duplex. The eponymous data rate of ethernet media is always unidirectional, meaning that is the data rate for both directions in full duplex.
    Working B.S., Computer Science
    Complete: 55/120 credits SPAN 201, LIT 100, ETHS 200, AP Lang, MATH 120, WRIT 231, ICS 140, MATH 215, ECON 202, ECON 201, ICS 141, MATH 210, LING 111, ICS 240
    In progress: CLEP US GOV,
    Next up: MATH 211, ECON 352, ICS 340
  • pcgizzmopcgizzmo Member Posts: 127
    Both answers above are correct but you will rarely get the full throughput of Gigabit. There will always be some overhead and latency.
  • ptilsenptilsen Member Posts: 2,835 ■■■■■■■■■■
    In fact, you will never get full throughput. Even connectionless transport protocols have some overhead from error checksums, and generally even application-layer protocols designed for connectionless transport protocols will have some on top of that. It is the fundamental difference between throughput and data rate.

    On that note, since the connection handshake and any error correction will produce overhead and opposite-direction traffic, it is important to note that concurrent opposite-direction traffic using connection-oriented protocols will always compete. E.g., if I use SMB to send a large file from a dedicated drive to a dedicated drive on a remote system, the throughput will go down (slightly, mind you) if that remote system then sends me a file (again, to and from dedicated drives to avoid disk I/O competition). The converted byte throughput of the transfer will already have been less than the 1gbps data rate, and the bi-directional traffic will be an even lower percentage.

    In real-world scenarios, you might see 900-950mbps going one way and 800-900 going both ways, using very good hardware on both ends with transfers not competing for non-network resources.

    So, summarizing that lengthy addendum, data rate (simply, BPS of the medium) in and of itself can get to the full amount in both directions. Throughput will not be the full amount in either direction, and even though the medium is full duplex, concurrent traffic in each direction will usually affect the throughput of both directions.
    Working B.S., Computer Science
    Complete: 55/120 credits SPAN 201, LIT 100, ETHS 200, AP Lang, MATH 120, WRIT 231, ICS 140, MATH 215, ECON 202, ECON 201, ICS 141, MATH 210, LING 111, ICS 240
    In progress: CLEP US GOV,
    Next up: MATH 211, ECON 352, ICS 340
  • theodoxatheodoxa Member Posts: 1,340 ■■■■□□□□□□
    I usually get about 80 Mbps on a 100 Mbps connection. I haven't tested the speeds over Gigabit yet, but at that speed, other bottlenecks (Hard Disk, Memory, etc...) might limit the speed you can actually achieve.
    R&S: CCENT CCNA CCNP CCIE [ ]
    Security: CCNA [ ]
    Virtualization: VCA-DCV [ ]
Sign In or Register to comment.