Delivering PTP SLA using a PTMP Wireless Architecture for IP Video

On May 1st the first shipment of the new Radwin 5000 PTMP product will start shipping from their New Jersey warehouse.  At first glance the Radwin 5000 is similar your typical wifi base station but on closer inspection the Radwin 5000 is the first of its kind.

Standard wifi base stations have a limited capacity delivered in a first come first served basis.  Subscribers connecting take as much or as little of that capacity as they want.  Throughput is not guaranteed per subscriber unit which means that if your neighbor decides to start downloading torrents from the internet and takes up 90% of the capacity of the base station you and the rest of the subscribers will have to contend for the remaining 10%.  This might be OK in broadband service networks where you have usage cycles and over subscription but in a constant streaming applications such as IP video surveillance where dedicated bandwidth is needed at all times it doesn’t.

Enter the Radwin 5200 base station.  The 5200 is a 200 Mbps (100 Mbps currently) MIMO radio, deployed in sectors that can support up to 16 subscriber units with each SU guaranteed a fraction of the capacity of the base station.  Radwin does this by employing time slots.  The number of time slots allocated per SU depends on the licensed throughput of that device.  Currently each time slot equates to 6.4 mbps and with 16 time slots the total capacity of the base station is currently 102.4 mbps.

Here is a table of the maximum number of time slots allocated to each SU:

  • 10 Mbps SU = 2 time slots (max)
  • 20 Mbps SU = 4 time slots (max)
  • 50 Mbps SU = 8 time slots (max)

So if the base station has 16 time slots you can mix and match SUs to meet your needs.  For example, you can have two dedicated 50 Mbps SUs or eight dedicated 10 Mbps SUs or one 50 Mpbs and four 10 Mbps, etc.  There are many permutations but the key thing to remember is that each SU will have a guaranteed SLA.  I am also told by Radwin that you can specify, in software, the actual number of time slots you allocate to a specific SM.  So if you wanted to give a 10Mbps SU one time slot instead of the two then you could do so and increase the number of devices supported per base station.

What if I need more subscribers?  Well glad you asked.  You will notice that I used the word currently throughout this post.  This is because in the current release of software for the Radwin 5000 is limited to a maximum aggregate capacity of 100 Mbps.  However in 8/2011 Radwin will release a free firmware upgrade doubling the capacity of the base station to 200 Mbps.  This is accomplished by increasing the maximum channel size to 40 MHz from the current limitation of 20 MHz.  So by doubling capacity Radwin is essentially doubling the mbps/time slot figure, halving the number of time slots needed per SU and doubling the number of SUs supported per base station.  If that is confusing here are the numbers you can expect.

  • 12.8 Mbps / time slot
  • 204.8 Mbps capacity per base station
  • 10 Mbps SU = 1 time slots (max)
  • 20 Mbps SU = 2 time slots (max)
  • 50 Mbps SU = 4 time slots (max)

As you can see the Radwin 5000 changes the way we can design high throughput networks.  Use the multiple point to points when you need more than 50 Mbps but below that the Radwin 5000 should come out more cost effective, easier to install and more efficient to manage.