Preview of what's coming next to Asuswrt

[RMerlin]

Howdy folks,

These past few years, Asus has typically released one major firmware update every year. In 2012 they introduced an update which added AiCloud, and they eventually implemented it across older routers that were powerful enough to support it. In 2013, they added TimeMachine, OpenVPN, and VPN client support.

2014 seems no different. Asus is currently finalizing various major changes to their Asuswrt firmware, which should be introduced in the next couple of months.

At this point, it is unsure which router model will get which new feature. Keep in mind that some of these features require more CPU power or memory than is available on older routers such as the RT-N16. I would expect most (if not all) of these features to be available on at least the newer dual core routers such as the RT-AC56U and RT-AC68U, in addition to their next flagship routers announced at CES and Computex.

Just a preview of some of these changes people should expect in the coming months:

• Whole new QoS engine (I tested it with speedtest.net + a concurrent torrent, and it worked beautifully, with a minimum of configuration required - my torrent dropped to next to 0 as the speedtest.net reported my full 30 Mbits link capability).
• Improved Networkmap
• Improved realtime traffic monitoring of your network clients
• Improved Parental Control settings (you will be able to control more than just the time at which they can access the Internet).
• New Internet security features, based on multiple technologies including DPI (this is what Asus called AiProtection at the recent Computex), bringing protection against port scanning, known malware menaces, etc...

There's currently no ETA from Asus beside "in the coming months", and I assume these features might still change before their release (which is why I won't go into further details at this point), but this will give you an idea of what to expect in the near future.

 

[cmillar6]

It would be nice if they overhauled the GUI a little bit. A little eye candy when you log into your router always makes for a more pleasing experience.

 

[RMerlin]

It would be nice if they overhauled the GUI a little bit. A little eye candy when you log into your router always makes for a more pleasing experience.

Well I tried it, and while some liked it, others lynched me for it LOL!

Quite honestly, the Asuswrt UI is already a bit heavy, making it react a bit slowly. Too many images for instance. Thankfully enabling caching for these helped quite a bit, and Asus has also integrated that a few revisions ago.

When it comes to management interfaces, the simpler is the better. Personally, I kinda liked the interface on the WRT1900AC, aside from a few quirks here and there.

 

[srinivasvaradaraj]

Interesting, I wonder if the changes go down to newer/drivers for wireless or if it remains the same. Just curious, not that the current devices/firmwares aren't delivering.

 

[cmillar6]

Well I tried it, and while some liked it, others lynched me for it LOL!

Quite honestly, the Asuswrt UI is already a bit heavy, making it react a bit slowly. Too many images for instance. Thankfully enabling caching for these helped quite a bit, and Asus has also integrated that a few revisions ago.

When it comes to management interfaces, the simpler is the better. Personally, I kinda liked the interface on the WRT1900AC, aside from a few quirks here and there.

Yes, I was indirectly referencing the interface look Linksys did with their firmware. They did a good job on visual appearance however its too stripped down feature wise for me.

 

[L&LD]

Normally, I don't care about the look of the gui. As long as the underlying functionality is there.

However, I got curious with all this talk about the WRT1900AC and had to see what this was about and found a setup video to watch.

All I can say is that I greatly prefer the Asus gui instead.

 

[sinshiva]

QoS for mips?

 

[RMerlin]

QoS for mips?

I doubt the MIPS-based routers have the necessary CPU power to handle the new QoS engine, but I could be wrong.

 

[XSXS]

Yeah keep the GUI simple, no eye candy for me!

 

[RMerlin]

An additional fact: the new QoS engine is compatible with HW acceleration.

 

[Milkman]

hmmmmmmmmm - Hope it supports the AC66u. I like the router and would hate to upgrade right now,

 

[connorm]

An additional fact: the new QoS engine is compatible with HW acceleration.

how the hell almost all the qos mechanisms im aware of require HW acceleration or "fastnat" (basically its offloading) to be disabled

however it is possible to keep offloading or "hw accleration" on while still running QOS. but its more complicated then that

for example on my ubnt edgerouter lite i have ipv4 offloading enabled.

i have a custom fq_codel + htb script set to limit my upstream only.

so because i have the script for upstream only. my downstream will still recieve the ipv4 offloading while my upstream will not.

this is a much more logical way of doing qos and disabling offloading as needed. basically it keeps running in the background and certain rules break it.

rather then with asus requiring a full reboot to turn offloading on or off.

IF asus has qos rules that remain active while hw acceleration is on. that seems pretty big.

theres no way that mips wont be able to handle it. the fact it can be enabled means there qos mechanism works with hw acceleration on which will make it an easier job on the CPU

so rt-ac66u where its limited to 100mbit +/- with QOS on (hw acceleration off)

if it could have hw acceleration on then it could simply run higher then 100mbit under less load.

if they have the technology to do it i think its related to the code more then raw processing power.

 

[RMerlin]

theres no way that mips wont be able to handle it. the fact it can be enabled means there qos mechanism works with hw acceleration on which will make it an easier job on the CPU

The new QoS is DPI-based, hence the need for a faster CPU.

 

[connorm]

i will be interested to see the result of this, sounds like it could be hit or miss. fairly exited to see it either way though

i doubt any qos mechanism can beat fq-codel + htb. if someone can beat these results they deserve a cookie

http://n1.netalyzr.icsi.berkeley.edu/summary/id=43ca253f-6039-194516f9-c73a-469a-ad1e



with fq-codel + htb



normal connection with no QOS rules applied

my downstream is slightly affected because i turn gso and gro off.

when i had them on i still got full 107 down. but with them off my downstream is slightly more stable even with no qos rules applied to it.

also my brother was watching netflix when i did the first speedtest lol

u should tell asus to start using newer linux kernels and try out fq-codel......... my script really isnt even that complicated

#!/bin/bash

ethtool -K eth1 gso off gro off
ethtool -K br0 gso off gro off

WAN_INTF=eth1
WAN_UP_SPEED=5300kbit

TQDISC=fq_codel

FQ_CODEL_QUANTUM_UP=300

FQ_CODEL_TARGET_UP=5ms

HTB_QUANTUM_UP=1500

TQDISC_OPTS_UP=
if [ "$TQDISC" == fq_codel ]; then
TQDISC_OPTS_UP+="quantum $FQ_CODEL_QUANTUM_UP "
TQDISC_OPTS_UP+="target $FQ_CODEL_TARGET_UP "
fi

WAN_IFB="ifb_${WAN_INTF}"
TC=/sbin/tc
IP=/sbin/ip

$TC qdisc del dev $WAN_INTF root >/dev/null 2>&1
$TC qdisc del dev $WAN_IFB root >/dev/null 2>&1

[ "$1" == clear ] && exit 0

$TC qdisc add dev $WAN_INTF root handle 1: htb default 10
$TC class add dev $WAN_INTF parent 1: classid 1:1 \
htb quantum $HTB_QUANTUM_UP \
rate $WAN_UP_SPEED ceil $WAN_UP_SPEED
$TC class add dev $WAN_INTF parent 1:1 classid 1:10 \
htb quantum $HTB_QUANTUM_UP \
rate $WAN_UP_SPEED ceil $WAN_UP_SPEED
$TC qdisc add dev $WAN_INTF parent 1:10 handle 100: $TQDISC $TQDISC_OPTS_UP

$IP link add $WAN_IFB type ifb >/dev/null 2>&1
$TC qdisc add dev $WAN_IFB root handle 1: htb default 10

$IP link set $WAN_IFB up
$TC filter add dev $WAN_INTF parent ffff: \
protocol all prio 10 u32 match u32 0 0 flowid 1:1 \
action mirred egress redirect dev $WAN_IFB

 

[RMerlin]

u should tell asus to start using newer linux kernels and try out fq-codel......... my script really isnt even that complicated

This is up to Broadcom, not Asus. That's why every Broadcom routers also use the exact same kernel version for a given SDK version.

 

[connorm]

but the only limitation to using a different kernel is that you would loose HW acceleration correct?

just to clarify, is hw acceleration basically just offloading? because i cant tell the different between hw acceleration on my rt-ac66u and offloading on my edgerouter.

they seem to be the EXACT same thing, but while the asus is using broadcom my router is using an 64 bit cavium octeon processor so there is possibility they both do hw acceleration or offloading differently.

i dont know why asus wouldnt simply forget hw acceleration and just use the newer proven qos methods givin all the new ARM processors. soon almost no one will need HW acceleration because in a year or two the processors will be able to handle 1gbps without it.

maybe asus should look at cavium chips or another vendor who can do hw acceleration on newer kernels

 

[RMerlin]

but the only limitation to using a different kernel is that you would loose HW acceleration correct?

They would probably also lose official support from Broadcom if they ran into any issue with the code. This might be fine for an open source project like DD-WRT, but for a manufacturer like Asus that kind of technical support is essential.

just to clarify, is hw acceleration basically just offloading? because i cant tell the different between hw acceleration on my rt-ac66u and offloading on my edgerouter.

they seem to be the EXACT same thing, but while the asus is using broadcom my router is using an 64 bit cavium octeon processor so there is possibility they both do hw acceleration or offloading differently.

Hard to say since only Broadcom knows exactly what goes in CTF. What little I know is that it bypasses part of Netfilter (essentially the FORWARD chain, maybe other parts as well). It probably also does other things, but Broadcom isn't sharing any info...

maybe asus should look at cavium chips or another vendor who can do hw acceleration on newer kernels

Asus does have non-Broadcom devices (like the RT-N56 and RT-N65), but they seem to be heavily invested on the Broadcom platform for all their high-end devices (with the RT-AC87U being an hybrid of sort, being Broadcom and Quantenna).

Meanwhile, Netgear seems to be working with quite a few different providers lately, with Mediatek/Qualcomm appearing on some of their new products.

 

[connorm]

good info Merlin. thanks

but isnt ctf different then hw acceleration?

cut-through-forwarding

 

[mikmock]

2014 & Merlin

As long as you are involved with ASUSWRT I believe we will be fine.
With your tweaks asuswrt-merlin has been an awesome product.

BTW...i started with the RT-N66U and recently added a AC68U as my primary router which is great but i still use the N66U(35_4-sdk5) for my 2.4GHz AP. it has unparalleled range(i have tested against my AC68, d-link & cisco).
I would be surprised if they achieve that range again.

Thanks for all your hard work Merlin

 

[RMerlin]

good info Merlin. thanks

but isnt ctf different then hw acceleration?

cut-through-forwarding

Asus calls it "HW acceleration", and I suspect it does more than just cut-through-forwarding. Some theories point at possibly allowing direct access (perhaps through DMA) to switch buffers. Also, the newer generations of BCM470x have a new technology called Flow Accelerator, which is also handled by the ctf module.