Hello ZO,
"I have question about submarine optical network vs. in land optical network. Generally speaking, what is the major difference from the following points of view: a) technically
b) major obstacles c) what is the key to have competitive advantage"
Transoceanic submarine cables, as opposed to shorter underwater crossings and those traveling over land, tend to be static resources in the sense that there are no mid-span intersections with other systems, which means no adding and dropping (yet) of channels, as takes place on overland systems where nodes serve as interconnects for converging / diverging routes. It's a multilane highway which remains consistent throughout, with no on- and off- ramps along the way, in other words.
There are numerous challenges in dealing with underwater systems, to put it mildly, that barely show up on normal overland systems. Electrolytics, and the increased atmospheric pressures which exist thousands of feet below the surface, have a detrimental effect on cable and deep sea amplifiers. These make material selection and engineering far more difficult than for those used at sea level. Where the latter, in contrast, are usually housed in environmentally controlled spaces.
And the powering of those deep sea amplifiers is a huge nut to crack, because all electrical power must come from the cable stations on land, or special electrical power generating stations near the landing points.
I've seen some system specs where power being supplied to subsea systems were rated at thousands of volts d.c., which is sent down both ends of the system via the cable's center metallic members.
In contrast, overland amplifier power supplies are fed by nearby-generated power at normal commercial levels, and are usually backed up by UPS, emergency generation and battery.
Due to the powering challenges, the number of amplification elements in deep sea enclosures are kept to a minimum, which is a primary, but not the only, reason why there are so few strands used in very long haul submarine cables.
Typical systems have only a half dozen or so strand pairs -- eight strands <four pairs> for four independent bi-directional multigigabit flows, and I think sometimes another one or two pairs for standby (sparing) and/or operations maintenance, telemetry and provisioning purposes. And in contrast to this seemingly paltry count, overland cable constructions now boast hundreds of strands.
Components, too, must be rated at much-longer mean-time-between-failures criteria, or MTBFs, because components are not easily accessed when they fail at the bottom of the ocean and need to be more reliable (and well worth the premium paid for ensuring same). And, of course, the main concern here is this: If the components are not accessible when they fail, then outages would last much longer and occur more frequently.
For this reason, testing of individual components and systems are far more stringent than their overland counterparts, which, consequently, leads to a much lower yield, which in turn drives the costs of manufacturing up further.
Power levels of pumps used in erbium doped fiber amplifiers, or EDFAs, are also notably higher than their topside brethren, due to greater spacing intervals used at the bottom of the ocean than over land.
Other challenges exist from the very early stages of system design, when performing initial surveys, as well as the methods employed in protecting cables on the continental shelf (by plowing), and making repairs at sea when fishing trawlers disregard the charts.
The other stuff you asked me about, I think would be answered better by someone else. |
