10 Gigabit networks

[V-Spec]

Since (some of) you guys seem to struggle to think of uses for this I'll use my own case as a quick example. I go away (a lot) for extended periods of time. When I travel I like to take movies/music/games with me to help pass the time. Invariably I will forget to transfer them from my server (WHS) until the last minute and then always get the comedy box telling me it will take 3 hrs to transfer my files. At that point (with minutes few to wheels) the difference between 3 hours and 18 minutes is a LOT!

I'd put money on you not even being close to maxing out 1G with WHS between two machines.

Now granted I couldn't saturate a 10Gb pipe with my current rig - I'd be lucky to get anything over 300MB/s onto my creaking Raid 0 SSDs on the laptop, let alone an external enclosure. But the point is that even on my old rig (3 years+ XPS m1730) my network is the limiting factor. I could and would use a faster connection if it were available.

1G is not a limiting factor in anybodies home network comprising of just a few machines.

Theoretically - you should be able to plug two machines together across a 1 meter length of cat5e with 10G network cards, pull a file from one to the other and providing you have capable SSDs - immediatley get 10Gps of throughput...

Fact is - real world networking just isn't like that, stuff like WHS uses TCP which is notorious for not windowing or ramping up properly in the presence of abundant bandwidth.
When transferring files at speeds of over 1G, your CPU will be high which will introduce latency in the application layer - which will limit TCP and the lower network layers from reaching their max performance..

This is why 10G links or bundles of 10G links exist to connect cities and massive enterprise networks together, not two machines.

The bit I find surprising is that people on this Forum, of all places, seem willing to accept that limitation. Just look at the excitement over boot times in the HDD section now that SSDs are common place. If 10GbE was available at a reasonable price, would you really turn it down because you seldom use it and are happy to wait?

Well, there will come a point (probably in the next 7-10 years) where we'll see 10G either come down to consumer prices or replace 1G.. Until then it doesn't bother me - I can't even max out 1G with my Crucial M4s so why am I bothered about 10G?


And to be honest - I'm not even sure we even will see 10G network cards become readily available for the consumer, I think it's more likely we'll see more focus on faster and lower latency wireless - WiFi that can do 4-5Gbps and is extremely low latency.

 

[HazardO]

Ah yes, Wikipedia, that fountain of knowledge that is 100% correct at all times and is certainly never updated by random people around the world. No guarantees that 6 will work with 10GBASE-T, and many many reported problems.

A good read.



1000BASE-T. It was slated as having more available bandwidth available above and beyond 1000BASE-T, but at the time, no-one expected the next logical step to be 10GbE.

As mentioned 10GBASE-T will work over CAT6 at a reduced distance, even CAT5e can be used upto ~40m. CAT6a is ideal though.

 

[Skidilliplop]

But the point is that even on my old rig (3 years+ XPS m1730) my network is the limiting factor. I could and would use a faster connection if it were available.

No, your network design is the limiting factor. I have iSCSI SANs running Gigabit and can max out even new SSDs with ease. It's all about how you move the data.

To max out the 10G link you'd need to run multiple sessions, if you're running multiple sessions anyway, why not run them over multiple paths?

If it's taking you 3 hours to copy data in a single stream over Gigabit (assuming the network is the limitation) using 3 sessions over 3 gigabit NICs would reduce that to about an hour. Over 4 NICs would reduce it further to about 45 minutes.
The kind of goal you've set out for yourself is already achievable using current 1Gbit technology. At a fraction of the cost of 10GbE.

10GbE was developed, like everything else, to solve a problem. Problems arising from short comings in Gigabit Ethernet. But as shown, your particular problem can easily be solved without it and at lesser cost. So where's the incentive to create a new product to solve a problem that's already been solved.