How fast should my gigabit ethernet network be?

1st, not all home routers have fast processors, such that they actually switch traffic at true gigabit speeds. As an experiment run an Ethernet cable between each Mac and turn off WiFi so it is not used by accident.

2nd, make sure your Ethernet cables are good. I once had a cable that I got with some device which was really only 4 wires instead of 8 and as a result it only did 100megabit. If you have a cable that is miss wired, even if an 8 wire cable, the interface may revert to 100megabit.

3rd, FireWire protocol is much more optimized for file transfer, so it is possible to edge out other interfaces of comparable speeds.

I don't have the equipment to try this but my understanding is that if 2 computers are connected to a switch which is in turn connected to a modem/router then initial connection will be at router speed but subsequent file transfer should be at switch speed.

It is asking a lot to expect modem/routers to do much at their maximum speed as their single processor is trying to do so many things at once but switches are optimised for data transfer speed subsequent to connection.

It should be possible to do a switch speed test without the router connected.

Further reading at:

https://encrypted.google.com/search?q=switch+router+speed&as_qdr=all&newwindow=1 &num=100&filter=0

Tom's Hardware links are helpful.

Host-local interconnects will be faster than network interconnects. FireWire 800 has a theoretical bandwidth of ~786.432 megabits per second. Gigabit is theoretically slightly faster, but you won't get anywhere near that theoretical maximum with the usual baggage.

Unlike the local path, your remote copy test is running through an IP network stack (that's contending with a potentially lossy network) and a file server client and a remote file server, and only then along to the disk, and the network stack and the remote file server are extra layers that will slow down the aggregate performance.

On balance, I'd expect better performance from FireWire 800 than Gigabit Ethernet, if similar disks are used. I'd definitely expect better performance from Thunderbolt, too.

Related: "goodput", "where's my bandwidth?" and measuring performance (caution, an advert may autoplay), and further down the rabbit hole, measuring gigabit performance.

BobHarris is also correct; IP networks encountering Ethernet-level errors or configuration problems can decrease performance. Sometimes massively. This can show up as duplex mismatches, and as connections running at 100 megabits per second or slower, and as bad cables and cabling faults.

One option for troubleshooting some of this — if your present router/switch lacks LEDs or a management interface that can see the negotiated settings — can be either an unmanaged switch, or (for better visibility) a managed switch. The latter are more expensive, of course.

Run some benchmarks here, and see what you're getting for performance. Post the details, as well as a general description of the Mac and storage hardware involved. (Somebody using a Mac with internal SSD will have different I/O performance than somebody with an old USB 2 disk, for instance.)

The fastest network drives around are usually part of a Fibre Channel Storage Area Network. The FC SAN eliminates the IP network stack and related parts, and the FC SAN is also spec'd with much lower bit error rates than are normally found on Ethernet networks. (Part of why IP goes slower is due to contending with those higher network error rates.)

Neville Hillyer wrote:

I don't have the equipment to try this but my understanding is that if 2 computers are connected to a switch which is in turn connected to a modem/router then initial connection will be at router speed but subsequent file transfer should be at switch speed.

Resolving the path to the destination address via ARP or via router for off-subnet traffic is going to be negligible in a data transfer of any size, and ARP caching will mean the connections operate at full speed. The overhead of shoveling the bits of any decent-sized file will be far larger than a few packets flying around to figure out the network path. Typical benchmarking practice will also deliberately "warm up the caches" with a few transfers, but the ARP caches should completely mask even an underperforming IP router here.

Yes. you can do that by link aggregation but that would ask you to have more than 2 ethernet ports. Say, if you have 4 ethernet ports, the resulting logical link will have a theoretical bandwidth of 4 Gbit/s using four CAT6 cables.

Since your house is wired in cat5e, it is inherently slower no matter what method you use. So I'd rather put NAS next to your iMac using thunderbolt if possible because the cable length also affects your transfer rate. It depends on how far you want to put your NAS next to your iMac.

Suppose you put your NAS in 2F of the house and iMac in 1F, you won't get much out of bandwidth even if your NAS is using USB3, or any other kind of ethernet cable like CAT6 or CAT7 --- or vice versa, your NAS in 1F, then your iMac will be slower in 2F.

You can rewire your house cables to be CAT6 structured if possible. It won't be too hard.

Mr. Latte wrote:

Since your house is wired in cat5e, it is inherently slower no matter what method you use.

Both Cat 5 and Cat 5e are rated for 100 MHz, and both are capable of supporting gigabit speeds.

Cat 6 adds support for 10 GbE, which isn't provided by Apple on any current Mac system configurations.

While you can add 10 GbE to various Mac systems (which might help here), adding a FC SAN also usually then becomes a good alternative.

You can rewire your house cables to be CAT6 structured if possible. It won't be too hard.

If and when rewiring a network — which can be no small hassle in various locations — consider dragging something faster than stock Cat 6. Cat 6a, or some combination of 6a and maybe also optical. (During one network installation project back when Cat 5 was commonplace, 350 MHz Cat 5e was installed — and yes, there were different things all labeled Cat 5e back then — and that cabling turned out to be rated above Cat 6 (at 250 MHz), when Cat 6 was finally approved. Which means that that particular 5e installation — now sometimes called Cat 5enh 350 MHz in retrospect — should be able to support 10 GbE, assuming the cable vendor's rating can be trusted, and assuming the cabling hasn't degraded.)

That "might help here"? Rewiring? In general, consider investigating the network performance before dragging more and newer wiring, and see what's going on. If the installation is getting a reasonable fraction of gigabit, or if it's running unusually slowly. That investigation might lead to dragging new wire (if there's a cable flaw somewhere), or it might show that the OS X file server is just slowing things down, for instance.

"Both Cat 5 and Cat 5e are rated for 100 MHz, and both are capable of supporting gigabit speeds."

In general 100 MHz is not equivalent to 100 bits per second. They are related but in general the frequency has to be several times the bit rate. There are a few exceptions such as ADSL where the bit rate is higher than the frequency.

I have often seen high specification cable bent much sharper than its specified minimum bend radius which can reduce its throughput. The following can also affect performance:

• running next to metal or power cables
• extra inline plugs/sockets

Yes. MrHoffman is right.

CAT6 may be an overkill for most households right now. CAT6 is a bit pricy too.

Some people think CAT6 is a big scam to the network world especially when consumers see the advertising packages of CAT6.

To my understanding, there are people who think that CAT6 outperforms CAT5e in a small margin about 2~3% faster when implementing in gigabit environment but it may also depends on the quality of the copper cord, thus making the difference almost negligible.

Getting a Packet Trace (QA1176) might be interesting, too.

Cartoonguy wrote:

At the end of the day, how can I get a NAS set up that has effective transfer speed? Or perhaps I just can't.

What are your plans for this NAS? How much data do you plan to copy on a regular basis? What kind of data? They are the kinds of things that sound great in theory, but are a royal hassle in practice.

It is pretty unusal to see a house wired for ethernet. If that wiring was done to the same quality standards as the rest of the typical, modern home, you should be lucky that it works at all.

I am doubtful that your yellow bend radius in you Unicom box meets the standard and I also wonder what the degradation is for their type of termination.

Reflections from non-standard terminations are best shown with a time domain reflectometer but a simple test is to try different routes/cables/terminations.

I would try a large zip file which takes at least 30 seconds over the fastest route.

TCP/IP connection...

You could try to expand the size of the TCP/IP data buffer. Look at line to see if you are getting errors. The ifconfig command will give you error stats.

You need to think about the TCP/IP protocol too. I do not know the protocol well enough make formal postulates. I think there is a frame which established how many packets that can be sent before expecting a response. There is a time to wait until the line is clear to send.

Bending radius

Most Category 5 cables can be bent at any radius exceeding approximately four times the outside diameter of the cable.^[6][7]

http://en.wikipedia.org/wiki/Category_5_cable

Robert

It is not difficult to find different bend radius specifications. In 5 minutes I found:

4 x OD

5 x OD

8 x OD

10 x OD

12 x OD

one inch

two inches

Some of these relate to:

fixed

flexible

installed

installing

These complications have resulted in a recommendation that two inches be adopted as the minimum for all small data cables.

FireWire will be fast because it's a short bus and shorter buses can be fast, and FireWire got a low error rate (which means the software doesn't have to clean up from errors quite as often as happens in a network stack), and you're going through the I/O stacks multiple times. (Compare disk to host memory back down through the network stack and across the network (slow, physically long, shared) and back up the remote network stack into the file server and back down to the disk. Versus disk I/O (fast) to host memory to disk I/O (fast).)

A Terminal.app command such as system_profiler SPNetworkDataType will tell you if your systems are both configured and running GbE, but based on the calculations (below), I'd suspect you are. Here's an example of a box that's negotiated and running GBe, extracted from the output of that command:

Ethernet:

MAC Address: aa:bb:cc:dd:ee:ff

Media Options: Full Duplex

Media Subtype: 1000baseT

820 megabytes is roughly 8 gigabits. Over 30 seconds, that works out to be about 0.2 Gbps for the transfer, which isn't particularly out of the question for GbE performance involving a host-to-host (presumably) AFP file transfer.