FreeSpaceOptic.com

FreeSpaceOptic.com is home for manufacturer-neutral information on Free-Space Optics.
 


 

                                                                                                          

Hybrid FSO + RF?

Why go to the trouble of deploying two links, different technologies, different learning curve? Isn't there a "silver bullet" solution?

The short answer is "no". If you need FastE or GigE throughput, and the distance is >250 meters (weather interference issue starts about there depending on your locale) and you require virtually 100% availability, then backing up a Free-Space Optics link with an RF link makes sense.

You are in fact killing two birds with one stone: you can mitigate not only potential weather outage (mainly fog for FSO), but also potential hardware failure (transceiver dies, LAN fiber damaged, workman bumps unit, power circuit goes down).

The design of your hybrid system needs to very carefully consider some critical issues:

    What throughput do you need on the RF redundant path?

    Do you need a full FastE or GigE backup when it will be active very little?

    Licensed or unlicensed?

    If unlicensed, 2.4, 5.8, 24, or 60 GHz?

    What failover technology will be used? more >>

        I. Layer one? (specialized hardware solution)       

        II. Layer two?

            a. Unbalanced EtherChannel

            b. Spanning Tree (not recommended)

            c. Rapid Spanning Tree

        III. Layer three?

            a. BGP

            b. OSPF

            c. Least Cost Routing

A major problem that plagues dual pathed links using layer two failover protocols is "flapping". This is a condition that materializes during gradual onset of a weather event (e.g. fog) that causes the Free-space Optics link to rapidly "flap" in and out of service initiating rapid repetitive switching between the two paths, basically confusing the layer two failover protocol, and ultimately "hanging" the ports.

Flapping potential needs to be mitigated by forcing the link to run on the redundant path for a specified minimum number of seconds before returning to the primary path. Unfortunately STP and RSTP don't include this as a configurable variable. Unbalanced EtherChannel does; and certain FSO transceivers as well.

Layer three protocols, although more complex and costly to implement and manage, don't suffer from flapping as they are able to contend with packet-by-packet routing.

Figure 1.

Typical GigE Layer 1 dual-pathed hybrid FSO + RF link. Can provide virtually 100% availability. more

 


       Figure 2.
 Typical implementation

Latency and failover speed are additional considerations in a dual pathed link. Voice packets are particularly sensitive. Ideally, the failover should occur after the first packet is dropped, and before the next one is transmitted to eliminate errors on the link. This is achievable, but requires planning.

Failover delay is usually expressed as milliseconds, microseconds or nanoseconds. The decimal expression of each is:

0.000 000 001 [ billionth ]

nanosecond [ ns ]

0.000 001 [ millionth ]

microsecond [ µs]

0.001 [ thousandth ]

millisecond [ ms ]

The range for typical failover solutions runs all the way from about 50 milliseconds down to 700 nanoseconds. Suffice it to say, the faster the better.

So what's the "top of the class" hybrid configuration?

For the auto-failover, we really like a Layer one, dedicated hardware based solution because it is the quickest (usually in the 600-700 nanosecond range), most reliable, simplest to implement, and often the lowest cost as well. >>

For the Free-Space Optics link, our hot buttons include integrated redundancy support such as:

1. Configurable "receive light intensity" LAN port shutdown (used as a positive layer one or layer two failover trigger)

2. Configurable return-to-primary delay (eliminates Flapping)

For the RF redundant path, use a solid-but-slow 5.8GHz link just because it is cheap; or go right up to a matched FastE or GigE 60GHz link so the backup is as fast as the primary link (just costs more).

The picture at the top of this page is a dual GigE hybrid utilizing an active aligned LightPointe FlightStrata FSO link + a BridgeWave GE60 60GHz GigE RF backup.

__________________________

 

 

 

What's HOT, what's not...

    Auto tracking? >>

    Single vs. multiple beams? >>

    Hybrid FSO + RF? >>

    850nm vs. 1550nm? >>

    Passive vs. Active? >>

    Wikipedia on FSO >>

    FSO vs. RF? >>

    Security? >>

    Eye Safety >>

    Tell me more! >>
 

Where does Free-space Optics fit best? >> 

Where does Free-space Optics not fit? >> 

Competitive comparison of the major Free-Space Optics solutions. >>

Free-Space Optics White Papers. >>

Who are we, and what’s our motivation for supporting this web site? >> 

And last but not least…here’s our Free-Space Optics commercial website for specific product and/or service inquiries.  >>

...and our refurbished FSO equipment site. >>

 

 

 

 

 

 

Copyright-free site. Feel free to reproduce any items on this site unless protected or owned by others. Questions or comments? Email us.