The recent earthquakes and tsunamis have brought wide-spread devastation to Japan, including to the domestic and international telecommunications infrastructure that companies doing business in Japan rely on. (See the article, “In Japan, Many Undersea Cables are Damaged”). This impact extends to Cisco where Japan is home to numerous field offices with Tokyo the site of Cisco’s North Asia network backbone hub. This hub provides an aggregation point for regional WAN and Internet connectivity in North Asia as well as direct connectivity to four other regional CAPNet hub locations in Asia and the US.
In the wake of these tragic events multiple submarine cable and other telecommunications infrastructure was damaged. Three of the failed submarine cables took out three of our backbone circuits which were down for several days. But Japan was never isolated and the impact to Cisco’s business was limited and well contained due to a Cisco IT strategy for managing capacity, diversity and redundancy on the backbone network. A strategy of provisioning multiple (preferably three or more), redundant circuits using a strategic selection of diverse access providers and submarine cables has increased the likelihood that the network (and by extension, Cisco’s business) can withstand an event such as this where the impact to the regional telecommunications infrastructure is wide-spread and severe. To minimize impact during these events the strategy not only needs to account for on-going connectivity during outages but also must be sure that adequate capacity (bandwidth) exists on circuits still in up to service the load. Traffic from the three backbone links that went down was re-routed over other nearby WAN links, without loss of connectivity and only a moderate amount of additional latency.
Cisco hasn’t escaped this disaster completely unscathed. Several Japanese offices, particularly in the areas where damage is the greatest, have been impacted due to facilities damage. Cisco has closed some offices near Tokyo and is offering relocation support for employees and families who live near Tokyo. Some network links were impacted but those that weren’t ensured no overall business impact, as the rest of the Asia Pacific network sites remain connected into Cisco IT’s global WAN.
How long it will take for life to return to normal, especially in the hardest hit locations, is still to be determined. For telecommunications infrastructure alone repairs in the aftermath of such events can take days/weeks to restore basic service and a year or more to fully restore normal service. As such, without adequate capacity, diversity and redundancy on our backbone the business impact might not only have been significant but could also have been long-lasting. At the same time a bit of luck doesn’t hurt either. Cisco’s regional networks can be affected or even completely isolated by events that severely impact facilities, power or connectivity at one of our hub locations. But even then impact would be isolated to that single region and the rest of Cisco’s global network would continue to operate normally.
The entire Asia-Pac region, from the Bering Sea in the north all the way down to Australia, is highly susceptible to extreme weather and seismic events. While this particular series of earthquakes and tsunamis may be unprecedented in terms of impact to people and property the impact to telecommunications infrastructure is not. While not routine, these types of events and their impact to telecommunications infrastructure are not uncommon and will continue to be factored into on-going discussions around balancing business risk and the cost of our network. Today, the incremental cost for redundant backbone links is easily justified, as Cisco continues to work in Asia.