Current Schemes

We shall look at two different schemes that are currently used for activation of the alternate path. The first one is implemented by extensions to the RSVP protocol [6] while the second one can be implemented using plain IP or other standard signaling schemes such as GMPLS [4]. The following subsections discuss each scheme and derive average time estimates to complete each phase of the restoration process from detection to traffic-switching.

Table 1: Terms and values used in analysis and experiments as suggested in [4].

Term	Description
$f$	hop distance from ingress to node upstream to failed link, (nodes in OTN)
$l_w$	length of working path, (nodes in OTN)
$l_a$	length of alternate path, (nodes in OTN)
$h$	diameter of DCN, (nodes)
$b_{len}$	length of fiber link between adjacent OTN nodes, (miles)
$c$	speed of optical signal propagation through the OTN fiber links, (miles/second)
$d$	time by which the data burst must lag control header, (seconds)
$t_{LOL}$	time to decide Loss Of Light, (5 ms)
$t_{prop}$	propagation delay per link in DCN, (8 $\mu$s/mile)
$t_{proc}$	processing delay per node in DCN, (20 ms)
$t_{detect}$	time to decide link failure, (seconds)
$t_{restore}$	time to restore traffic, (seconds)
$t_{FIS}$	fault indication signaling delay, (seconds)
$t_{setup}$	time to setup alternate path, (seconds)
$t_{switch}$	time to switch traffic from working path to alternate path, (10 ms)
$t_{\times connect}$	time to perform a cross-connect operation (10 ms)