Superhub : Brand new superhub problem!

[qasdfdsaq]

Originally Posted by caph

I'm going to give you the benefit of the doubt, but please don't come back aggressively at me when you don't know what you are talking about.

If you're unsure or have a query about anything I've said then just ask, but please don't "tell" me what is clearly incorrect, it makes you look foolish and it wastes my time replying to you. It also detracts from the value of this forum.

I would suggest you do the exact same thing. You're completely wrong so please stop talking nonsense and misleading people

Benefit of the doubt ... switch segregates traffic based on IP addresses attached to each port.

False. An ethernet switch segregates traffic based on ethernet addresses to each port.

Therefore each port only receives directed traffic that applies to it. Do me a favour and install Wireshark on your PC, then start it running on your LAN interface. Next, start any traffic on any other port on your LAN and watch it magically appear in Wireshark

That is exactly what it does not do. Do me a favour. Verify your own facts before spouting nonsense. How about you go install Wireshark on your own PC then watch other traffic on your LAN not appear at all. I've already have it installed on all my PCs and spent enough time sniffing Superhub traffic several years ago. Then I started playing around with this kinda stuff for a while and got bored of the Pooperhub. Yes, that is a 4x10Gb internet connection.

due to the fact that the Superhub has a cheap hub and not in fact an intelligent switch.

Completely wrong. It is a layer 2 ethernet switch and not a hub. Most fundamentally it's full duplex and by definition cannot be a hub. Furthermore if it were a hub, why does it have a Broadcom BCM53114KFBG switch chip individually connected to each of its network ports? The same chip used in the Netgear GS1053E ProSafe Switch? Why do the Netgear base specifications indicate "Bridging" (a switch function) and "Spanning tree" (another switch function) and 802.1d MAC bridge (i.e. ethernet switching standard)in it's feature list?

Get a clue. Superhub is a switch and is subject to all the vulnerabilities of a badly programmed switch.

---------- Post added at 09:39 ---------- Previous post was at 09:35 ----------

Originally Posted by Qtx

On a side note, my sky hub (SR101) only has 100 and not gigabit ports. Obviously a switch type operation though as today I was ftp'ing some stuff between 2 machines at 11mb/s while downloading at 3.5mb/s and streaming a hight bitrate mkv to a wireless laptop from a hard wired NAS.

Not had a superhub but I would expect it to work in a similar way and not repeat all the packets out all all the ports, ala a hub.

Hubs have been outdated for a decade. You'd be hard pressed to find a modern one to buy new even if you tried. I've never even seen a gigabit hub in existence, ever.

---------- Post added at 10:03 ---------- Previous post was at 09:39 ----------

Originally Posted by ferretuk

I'm running DD-WRT on a WR1043ND. I don't use the wireless side though, so can't vouch for its performance in that area...

I use a 2543ND, which IMO is the cheapest and best value any-band (well it's dual band but one-at-a-time) router with gigabit ports. Personally I stick with Atheros based kit not only because of superb open-source and developer community but they're also far more reliable and less buggy than Broadcom, Realtek or Ralink stuff.

The 1043 and 2543 share the same CPU speed and architecture and the 2543 manages 220Mbps WAN to LAN throughput, so I wouldn't expect the 1043 to be much slower.

In my opinion if you want the best value 2.4Ghz performance and the flexibility of occasionally switching to 5Ghz, get the TL-WR2543ND.

If you want future-proofing and simultaneous dual-band get the TL-WDR3600 or 4300. The hardware NAT acceleration on those new processors gives 800Mbps+ WAN to LAN throughput.

---------- Post added at 10:04 ---------- Previous post was at 10:03 ----------

Originally Posted by sniper007

Really? How so? In that they over or under estimate WAN to LAn throughput?

Under, but my comments mostly revolve around their wireless testing. WAN to LAN isn't too far off, though sometimes is.

I tend to get wireless results that are 2-3x higher than they do under the same conditions, and not to mention they spent ages testing with a broken setup with missing antennas, which doesn't give me much confidence either...

[Qtx]

I still have an old cisco rack mount 20 or 24 port 100mb hub here somewhere. They were build like tanks considering their size lol.

[sniper007]

@qasdfdsaq thanks for the advice. The TPLink routers look good. I like the bang per buck they appear to give and the gradual price rises as the features increase. I am off to read up on them now.

I still want to solve the above problem though. It can't be a general problem surely as other people would have seen it? Maybe it is the combination of streaming to a PS3 that does it. I wonder if someone could test streaming to a PS3.

---------- Post added at 13:45 ---------- Previous post was at 12:39 ----------

@qasdfdsaq I just found your post here from a search where you reply to the TPLink rep from HQ. It's quite interesting that he states the - presumably "on paper" - WAN to LAN throughput as around half of what you tested it to be. Very interesting indeed. For me personally, I'm not bothered about dual band wireless as I have too many clients on 2.4ghz band, so....the TL-WR1043ND looks to be a bargain at £38.11.... if....and only if it does in fact have WAN to LAN throughput of which matches the WR2543ND. You state above you think it would be similar since it uses the same hardware internally. How confident of that are you? Serious bang per buck right there with the 1043ND.


EDIT:

Just seen that the specs for the 1043ND vs 2543ND. The CPU is actually a newer revision on the older 1043ND it seems ? I would assume the performance of WAN to LAN throughput should be near identical and if anything better on the 1043ND? Yet tests seem to put it at 120mbps.

1043ND:
System on chip: AR9132 rev 2
CPU speed = 400mhz
Flash = 8mb
Ram = 32mb
Ethernet: Realtek RTL8366RB


2543ND:
System on chip: AR7242
CPU speed = 400mhz
Flash = 8mb
Ram = 64mb
Ethernet: Realtek RTL8367R



Cheers

[qasdfdsaq]

It's a slightly newer revision but the same architecture (MIPS 24K). It's only when you move to newer architectures (e.g. MIPS 74K) that you see big differences in "per megahertz" performance. The newer gen new architecture CPU in the 3600 is 20-30% faster at the same clock speed than the AR7242 so any differences between the two revisions of the same architecture will be smaller than this.

The 2543ND can be had for about £38 apparently including delivery but most places sell it for about £45. The extra full-service wireless stream on that will give you slightly improved range and reliability even on devices that can only use 2 streams.

I still think the Superhub problem is due to it being confused by MAC/ARP/IP mismatches but you would have to use Wireshark or similar to diagnose it. And you *cannot* do this using the Superhub because it *is* a switch and not a hub.

---------- Post added at 06:25 ---------- Previous post was at 06:22 ----------

Originally Posted by Qtx

I still have an old cisco rack mount 20 or 24 port 100mb hub here somewhere. They were build like tanks considering their size lol.

If you want to talk size and built like tanks... How about a dozen HP 192-port 10/100 switches? That was our "pre-2007" hardware. Since then all our switches in new buildings are gigabit ported with two 10-gigabit uplinks on each floor

[sniper007]

I finished off testing this by using a different playback device on my network also via the superhub. Plays back fine. The problem is the PS3 believe it or not. Whether this is a hardware problem with the actual PS3 or if it is some kind of incompatability with the superhub I do not know. I tried resetting the superhub to factory default settings and it did not make a difference. Would be good if someone else can test streaming full 1080 rips to a PS3 via PS3MS using superhub.

[caph]

Originally Posted by qasdfdsaq

I would suggest you do the exact same thing. You're completely wrong so please stop talking nonsense and misleading people

Do ME a favour and stop regurgitating the "network switch" Wikipaedia page. It makes you look like a monkey see monkey do.

Wireshark shows traffic going to port 4 on port 3. Argue with me I dare you. It may have chips that are used in more expensive ports, but it acts as a hub and no matter how many words you pattern match on a Wiki page, that fact remains.

Monkey see, monkey fail.

---------- Post added at 23:35 ---------- Previous post was at 23:30 ----------

Originally Posted by sniper007

I finished off testing this by using a different playback device on my network also via the superhub. Plays back fine. The problem is the PS3 believe it or not. Whether this is a hardware problem with the actual PS3 or if it is some kind of incompatability with the superhub I do not know. I tried resetting the superhub to factory default settings and it did not make a difference. Would be good if someone else can test streaming full 1080 rips to a PS3 via PS3MS using superhub.

That's not consistent with it working fine when streaming via the other switch you tried though? Did it definitely work fine via the switch when streaming exactly the same media? This guy had the same problem but it was an encoding issue that the PS3 struggled with after a while.

[qasdfdsaq]

Originally Posted by caph

Wireshark shows traffic going to port 4 on port 3. Argue with me I dare you. It may have chips that are used in more expensive ports, but it acts as a hub and no matter how many words you pattern match on a Wiki page, that fact remains.

Wireshark shows no traffic going to port 4 on port 3. It does not act as a hub and no matter how many times you try to lie to people it will still be a switch.

Since you've presented absolutely no facts or evidence of relevance and seem intent on insults around, you clearly have no idea how to hold a civil debate or any clue about the subject at hand. Let me refer you to level two of this.

Monkey see, monkey fail.

[sniper007]

UPDATE:
So I took the PS3 round to my friends house who has also a superhub and a PC hardwired to it running PS3 Media Server. We streamed a 1080p high bitrate rip to MY PS3 on his network also via superhub and it all worked fine. So basically I have ruled out the PS3. The stuttering and ping timeouts ONLY happen when using my superhub at my house with my PS3. I don't know where to go next with this. I have even reset the superhub to factory defaults and it still does it.

[Sephiroth]

Perhaps you could compare router (SH) settings between your setup and your friend's. You can come back to us to consider the relevance of any differences.

[qasdfdsaq]

Even in factory defaults? Sounds like a Superdud to me (dud hub)

[sniper007]

Originally Posted by qasdfdsaq

Even in factory defaults? Sounds like a Superdud to me (dud hub)

Yes even straight out of the box factory defaults it has a problem using my superhub.
If I brought my friends superhub round my house and plugged it in, or mine to his house to test...would it work? i.e. Are they bound to a particular household or UBR based on mac address?

[qasdfdsaq]

UBR based on mac address.

Though it can be specific to a port/segment on the UBR as well.

[Mr K]

Ask them to give you an NTL250 modem. Rock solid and reliable for me. Never seem to have any of the issues that keep being reported with the Superdud.

[raging bull]

Sorry to say Mr K, the 250 has gone the same way as the Dodo!
Virgin have warned me the 256 I have will be automatically replaced by SHub if it dies.

[Mr K]

Well I'll be keeping my 250 till it dies. Unlike the Dodo it's very much alive. Don't see the need for anything more than 20mb. Rather have a reliable connection which it doesn't seem the sh can provide.