On Fri, 23 Jun 2006, Charles Wu wrote:

Right or wrong, in the context of throughput efficiency, the documentation I have seen regarding N-stream leads me to believe that frame concatenation is the main method utilized by the protocol. Would you care to expand/enlighten further (I am sure there are a lot of other inquisitive types like me who like to know how the insides of their "black box" ticks =)

Charles, Mikrotik chose not to reveal all the details on their NStreme protocol. However, some things that I do know is that it is basically the 802.11 type MAC. Part of the benefit is that the protocol overhead is reduced (a byproduct of the packet concatenation probably), the distance limitations (timing) has been removed and they have implemented an (optional) polling mechanism. These three combined will improve the overall throughput of the link. There are probably other things that make it a better solution, but this is what jumps to mind.

Additionally, they have NStreme2, which is a whole different animal. This protocol is a dual radio system (point to point only) that is designed for backhauls. This protocol is not based on the 802.11 stuff (as far as I know). There is no need for polling with NStreme2. The protocol overhead is very low with NStreme2 and it is robust in the face of interference. With a dual radio (2 on each end of the link), you create a full-duplex capable link.

One of the real limitations of either NStreme or NStreme2 is that they require an Atheros based card. That may not be a big deal to some, but it is something worth mentioning.

Butch Evans
Network Engineering and Security Consulting
Mikrotik Certified Consultant
WISPA Wireless List: wireless@wispa.org


Archives: http://lists.wispa.org/pipermail/wireless/

Reply via email to