*) ping times improved on Nv2 high data rate wireless links;
I’ve got a test link to try new thinks.
Signal is 70/70 TX/RX in the two chains. Radios are ubiquiti SR71-15. And the antennas are the wonderful Jirous JRC24 DUPLEX. Distance is ~15km. If you want more details i can give you radio mobile profile of the net.
Is a transparent link using WDS where i have got a couple of VLAN upon it.
NV2 in version <= 5.16 was TOTALLY unstable. latency and jitter simply do the link unusable.
With Nstreme enabled WITHOUT frame policy and whit polling i’ve got 100mbps send/receive and 50/50 both. Ping, latency and jitter is simply wonderful (1ms or 2).
I’ve upgraded to 5.17 to try the quoted feature. Now NV2 is true that it work better, also bandwidth test is more stable and powerful, i’ve got 130mbps send or receive and 65 both side. BUT this is the result of ping (BETTER THAN OTHERS VERSIONS):
Ok guys. With those results NV2 cannot be used in production environment. For me, and for anyone with a little bit of head, jitter and latency ARE MORE important than bandwidth.
Only to remark the difference with nstreme, this is the result:
I have issue with graphing tool in winbox only with resources. RB433AH 5.17.
Interface graphing in winbox is OK. On www everything is OK. Every resource in winbox is 1px high.
What happened with e-mail notification of new versions? It’s only its mentioned in another post that I became aware this new version is out… its the 6th of June now…
*) ping times improved on Nv2 high data rate wireless links;
What is exactly done to improve the ping times? A bit more technical background on what has actually done by technicians would be helpful to understand how and when this help ping times and if I really need this upgrade.
Does it mean the link as a whole is improved? Or is icmp protocol filtered and given high priority?
ping is only a package, so if ping is improved are other packages traffic improved too? If only ping is improved than that’s useless. Ping is used to see how a link performs. If ping now is stable where the rest of the link still would suffer latency and jitter the tool is useless and I see no reason for this improvement if only gimmick.
Please an explanation.
I agree, the ping time are better on interfaces with some moderate traffic. But still very big jitter, when with very low or no traffic. But this is the problem with waiting for some data to send just one big packet istead of few smaller.
I can also agree to improovments on stabilising transfer rates and speeds. It does not improove the maximum trafer rate too much, but it seems to be more stable, so more predictable version to version.
*) ping times improved on Nv2 high data rate wireless links;
In ptmp is it essential to upgrade both AP and stations to 5.17 to take full advantage of the benefit? Or is it enough to upgrade either AP or CPE only?
Vyatta FAQ states reason: http://www.vyatta.com/files/pdfs/Vyatta_6_4_FAQ_Final.pdf - performance, stability and need for frequent updates, IPS/IDS requires quite a lot CPU/memory so i can imagine to have it on new 36-core routerboard platform, but definately not on current boards. And if you need to run RouterOS on x86, then another box running just IPS/IDS is better idea.
yeah I know right, Funny.
