Don't be silly. For example, my brother works on Excel spreadsheets that run into the tens of gigabytes; the largest suite was over 100 GB. Even with gigabit ethernet those take a LONG time to transfer; with 10 gigabit, the time is reduced significantly.
Don't be silly. For example, my brother works on Excel spreadsheets that run into the tens of gigabytes; the largest suite was over 100 GB. Even with gigabit ethernet those take a LONG time to transfer; with 10 gigabit, the time is reduced significantly.
At best that’s funny, at worst an absolutely ludicrous example to try and justify a position that in no way is representative of the home user scenario you discussed. I don’t like to suggest it smacks of desperation, but it seems quite an over reach.
Ignoring that for a moment, it seems unlikely he has a data set like that of his personal stamp collection, so i’ll suggest he is working on a commercial basis, GDPR/ISO certification/legality/insurance/good practice would suggest he is unlikely to be backing up commercial works to a personally owned consumer NAS on his hone network, as that’s just a horrible scenario, but ignoring all of that...
The backup time uncompressed on 100GB is less than 20 minutes at gigabit speeds (assuming the both devices are capable of reading/writing quickly enough), an incremental should be significantly quicker. So presuming he backs up once per day using incremental and once a week as a full backup, 20 minutes once a week hardy seems massive. The offsite backup being WAN limited is more likely to be a problem or pulling fresh 100GB data sets is more likely to be a problem.
He isn't right now because he doesn't have 10 GbE. He uses an external USB 3.1 drive.
Soldato
For everyone one person like your brother there's ten's of thousands that aren't. As @Jez mentioned, it's down to demand and there's very little in home/SOHO/SMB environments.
But there are also thousands of people like him, just in the UK. Now scale that up worldwide.
So neither you nor your brother have had or currently do use 10Gb in the home environment or aware of the minimal costs. You argue he needs it, but state he uses local USB3 (presumably flash based?) storage. You’ve not stated he has ever used, needed or quantified any actual benefit to 10Gb, or that what he has meets the hardware requirements to use 10Gb or is even permitted to use the set-up you describe and clearly he’s not a ‘home user’, SOHO perhaps, but he’s not even 1% of the market. If you want 10Gb you can do it as stated for £30 ish p2p, if you want to roll it out further with a switch, add £100 upwards, that’s really not expensive. Putting slow AHCI based flash on a 10Gb link with a NAS or server incapable of writing the the R/W operations to move it is way, way more expensive, from the sounds of it he’s more likely to need the ability to move the data set, hence the current usage scenario. Either way to suggest meaningful demand exists among home users is laughable, especially with the usage scenario you offer.
Soldato
I think there are plenty of guys on this forum who work in IT and dabble with Lab/Dev/Test environments at home who would see a benifit in moving to 10GB if it wasn't so expensive to buy new.
So I'll bite! Why don't you spec up a 10 Gb Ethernet network for four people? Two adults and two children. Everyone has their own PC, the adults in a shared study and the children in their bedrooms, and there's a HP Microserver. The server is distant from the PCs but local to the internet router. From the internet router to the study is 30m and the bedrooms require an additional 10m or so each from the study. There's also a printer in the study with a gigabit port to be connected to the LAN. All kit is to be brand new.
Put up.
Don
Keep it friendly gents 
Keep it friendly gents
But his imaginary home users needs the 10G for the internetz pointz!
Why don’t I? Perhaps because you’ve supplied an illogical/fictional set of circumstances in an attempt to justify a position that has already been discredited in previous posts by myself and others? Strangely it also lacks any obvious or stated benefit for 10Gb or for that matter any credible suggestion of hardware that really lends itself to 10Gb and pointless restrictions on sourcing goods, all of which suggests you aren’t really getting any of this.
Home user demand for 10Gb is virtually nonexistent, out of that tiny percentage of demand, the number of home users who who want 10Gb and actually have the hardware capable of making meaningful use of it is even smaller. Out of those individuals, the amount they’ve likely dropped on servers and workstations by that stage makes connecting your physical lab (servers and workstations) via 10Gb relatively cheap. However, most server/workstation boards come with dual LAN which lends itself to link aggregation nicely, assuming you have the server side IO (a cheap AHCI based SSD set-up will do nicely), without justification this sounds like a better and more efficient solution for your work of fiction.
Actually, it's a combination of two real scenarios. So you can take your 'discredited' and discredit it.
I bet they said that about gigabit ethernet too. You'll note that high-bandwidth wifi routers are increasingly popular as long as they're cheap. Speed sells, but the price has to be affordable.
So please, get off your high horse and ignore why someone might want it and try to answer the question of why 10 GbE is still so expensive in the first place.
I answered this further up;
https://forums.overclockers.co.uk/posts/32620029
There is such little demand that there is no driver on price. It really is as simple as that. When gigE surfaced there was a very different demand profile. This is the same with all technologies, they become cheap with volume and demand.
It's true that most people do not need 10GbE but it's also the case that the vast majority of home users wouldn't notice any issues with using 100 Mb/s Ethernet either. The fact is that 100 Mb/s is no good for most businesses and that is why GbE equipment is now cheap and ubiquitous. Any medium scale (or larger) business will have 10GbE backplanes though, so it is a little puzzling that it's still so expensive. An additional problem is that WiFi is now ubiquitous so most homes don't even use wired Ethernet to begin with.
I guess we can just hope the introduction of 2.5GbE and 5GbE will help matters.
I guess we can just hope the introduction of 2.5GbE and 5GbE will help matters.
Link aggregating two ports doesn't give you double usable bandwidth, unless you're using multi protocol SMB or dual IP addresses on the host device.
Are you sure about that? Everything I've ever tested will only ever saturate one link, unless the underlying protocol supports it, such as SMB (which I think is still in beta for this use) as mentioned above.
http://h22208.www2.hpe.com/eginfoli...18/5998-8161_ra_2620_mcg/content/ch04s11.html
Final performance is still going to be dependent on the exact use case but there's a bit more to it than just source/destination IP or MAC address. SMB has been multichannel since SMB 3.0.
Final performance is still going to be dependent on the exact use case but there's a bit more to it than just source/destination IP or MAC address. SMB has been multichannel since SMB 3.0.
So it’s a work of fiction with some silly restrictions thrown in as I said? By your own admission it’s a made up situation, one that lacks any actual usage case or justification. Funny way of saying I was right
As someone who grew up using ser/parnet and dial-up BBS’ with a 2400bps modem when the rest of the world was still using sneaker-net, and migrated through BNC, 10Mbit RJ45 and hubs through to 100Mbit (switches... yay), gigabit and 10Gb: No, they didn’t. The Wi-fi example falls flat on its face, Wi-fi is a big numbers game for people who don’t actually move data. Let’s say you get a real world 10Gb Wi-fi standard, how useful will that be on a 64GB tablet? Will your 8Mbit Netflix stream be any quicker? Even the 30Mbit 4K stream will be exactly as it was before and subject to the same bottleneck : Your WAN connection. Anyone who moves a lot of data does so via a wired connection, Wi-fi is for IoT and devices that physically can’t be wired.
Also you keep suggesting it’s not affordable, you still haven’t realised how much you need to spend on each PC to be able to move network data at anything remotely near 10Gb line speed, read the previous posts.
The very same post you cherry picked from clearly stated ‘home user demand for 10Gb is virtually nonexistent’. You are talking about a product that is largely enterprise orientated, that means it’s priced for enterprise. Why would any provider making a nice margin on hardware decide to slash prices and sell to the 1% of home users who want/need/can use 10Gb? In a few years when faster storage and more efficient chipsets with better levels of hardware offload start becoming affordable, it’ll filter down into workstation and ‘gaming’ class hardware, but the home user demand is so small and the difference in chipset prices so high, that it just makes no sense at present.
Last edited:
I started on 300 bps and 1200/75. I tapped 5 Mb thick ethernet. I strung 10 Mb ethernet BNC cables. So, please don't try that with me. I've seen networking go through several life cycles. We're more than overdue for 10 Gb to migrate from the datacentre to ubiquity.
If it were cheap there would be demand.