Participating in a whirlpool thread recently, as to whether a Linksys WRT54GP2 is better than a Billion 6404VGP, I chose the Linksys over the Billion, for a few clear reasons.
1. Many Linksys devices have Open Source firmware available, so you aren’t pestering the manufacturer for a fix, and you can customise the firmware as much as you want, don’t want wireless, turn it completely off, don’t want “x” feature, remove it.
2. I find Billion’s in general to be the lower quality device when compared to Linksys, Netgear, etc. The web interface for example, is a frames page. Whilst I won’t judge them solely on that, the product does seem “lower quality” than other products.
3. My experience with Tomato, which is open source firmware, the QoS has been somewhat acceptable, although we continually limit maximum upload to 80%, to save the extra 20% for VoIP calling (even though we are relatively low VoIP users)..
Anyway, a representative from Billion jumped in the thread, and bought in a feature Billion’s have, which was described as “VoIP InUse”.
This sounds like a great feature for Australian conditions, because we have pathetic upload speeds, artificially limited so Telstra can preserve the revenue it gets from its older, and more expensive business products, among other reasons.
I am keen on testing this feature out, essentially, its described as “halting” internet traffic while a VoIP call is in progress, so that VoIP will be crystal clear (and this paves the way for better codecs).
That’s ideal for us however, because when we don’t have a VoIP call happening, I certainly would love to be able to FTP a database for example, at the maximum speed of ~40KB/sec my connection is capable of, instead of limiting my connections capabilities to 30KB/sec permanently, “Just In Case” a call came in.
One of the bigger let downs with the Billion however, is I’ll have to drop my bandwidth monitoring logging feature, that identifies how much usage we have gone through and can alert us to when we are reaching that critical moment where our ISP might try and scam $3 from us.
But, it does support SNMP, and so I can recode my system to take advantage of that instead, though I will lose the data during Billion testing in my rstats file (which is bad, because rstats really does work well).
The compromise here is that for the sake of being able to use my full connection, I have to drop a few features that I sort of use for watching Exetel (they only update the usage meter every 12 hours, apparently due to the number of radius records, but I think they are just not integrating radius well, because it really isn’t that hard to process that much data, another topic).
In possible news coming in days, I’m expecting Exetel’s announcement of P2P caching, which should see a dramatic increase in the speed of connections Exetel wide, mainly because the demands placed on the links will, I believe, be dramatically reduced, so that could cause Exetel to hopefully revisit the off peak period and allow faster speeds to the users in the pool, or be prepared for more growth.
In fact, like others, I’m amazed others haven’t gone the caching path sooner. For a small $400,000, an ISP can potentially reduce a lot of load on their links!!
Enjoy!