redundancy Archives — LIRNEasia


Given melting ice in the Arctic, cable laying ships are moving in. In the first phase northern communities are being connected. The plan is to extend to Asia, providing an additional layer of redundancy to international backhaul. In later phases of the project, Quintillion Networks will extend the network internationally to Europe and Asia, representing the first time the two continents will be directly linked by cable. In addition to cutting latency, these arctic cables add a layer of redundancy to our global communication systems.
The problem of sabotage of undersea cables was brought to the attention of UN ESCAP and the senior government officials who attended the ICT and DRR Committee meeting by LIRNEasia as far back as November 2010 (see slide 20). Now it’s headline news. Russian submarines and spy ships are aggressively operating near the vital undersea cables that carry almost all global Internet communications, raising concerns among some American military and intelligence officials that the Russians might be planning to attack those lines in times of tension or conflict. The issue goes beyond old worries during the Cold War that the Russians would tap into the cables — a task American intelligence agencies also mastered decades ago. The alarm today is deeper: The ultimate Russian hack on the United States could involve severing the fiber-optic cables at some of their hardest-to-access locations to halt the instant communications on which the West’s governments, economies and citizens have grown dependent.
We’ve been working with UNESCAP since 2010 on addressing a key condition for affordable and reliable broadband in Asia. Today, I was impressed by how far UNESCAP has advanced the process. I am a strong believer in the power of image. They had, together with ITU, commissioned a map of the existing fiber optic cables. This will, after all the necessary approvals have been received (good luck on presenting a map of India that won’t upset somebody!
In 2010, Senior Policy Fellow Abu Saeed Khan proposed that we address the problem of expensive and unreliable international backhaul in Asia. UN Under Secretary General Noeleen Hayzer of the Economic and Social Commission for Asia and the Pacific (ESCAP) understood the importance of the issue and we began to work with ESCAP. By 2013, the initiative has quite a bit of wind behind its sails. Two studies have been done on South East and Central Asia. The link between the lack of redundancy in backhaul and disaster management is being explored.
LIRNEasia and its people have been intimately involved in the spurts and starts of the policy discussion on Bangladesh’s international connectivity. We were early in pointing to the need for an additional cable, pointing to the multiple vulnerabilities created by the single undersea cable controlled by the government-owned BTCL and the non-ring architecture of the dry link from Dhaka to Cox’s Bazar. Now, with 3-4 of the terrestrial cables coming online, we have a natural experiment running in what addressing redundancy means. Renesys has shown the results for those with backup and those without. Neat.
Bangladesh is a big country and a coastal country. Yet it was very late in getting connected to a cable. SEA-ME-WE 4 became operational only in 2006. Then it was the largest country without redundancy,having to rely on SEA-ME-WE 4 even to communicate with India, the country that surrounds it. Now all that is over.
O3B is a new satellite company that is offering low latency (120-150 ms) satellite connectivity that should be of great interest to small countries without access to submarine cables; for those wanting redundancy for the existing cables, and even cruise lines. Simple physics does not allow geostationary satellites to give low latency. O3B is using medium orbit satellites. Worth a look.

Cloud requires redundancy

Posted on June 30, 2012  /  1 Comments

Al Jazeera reports that a lot of server farms got knocked out by bad weather in the US. Instead of making us rethink the cloud, this suggests that we need more redundancy, preferably with server farms on the other side of the world. Sean Ludwig, from VentureBeat, wrote in a blog post, “The outage underscores the vulnerabilities of depending on the public cloud versus using your own data centers.” The outages on Amazon’s cloud server come two weeks after a similar incident when a number of popular websites hosted by Amazon went down. A report into the incident by Amazon found that a configuration error was made during a routine upgrade.
Led by Senior Policy Fellow Abu Saeed Khan we’ve been saying that Asia needs a terrestrial cable system to back up the submarine cables. By the time international government organizations get organized, the private workarounds will be fully operational. Like traders plying the ancient Silk Road, telecommunications operators routing bits and bytes from Asia to Europe and back have to pass through the Middle East, whose tricky geography and even more challenging geopolitics have sometimes made the region just as much of a bottleneck in the digital realm as in the physical world. When things go wrong, the consequences can be serious and far-reaching. In January 2008, for example, several underwater cables off the Mediterranean coast of Egypt were inexplicably severed.
We first explored the idea of embedding sensors in dams so there would be better information about potential failures back in 2005 in the course of our dam safety research project. We were talking about relatively unproven RFID or electronic dust systems back then. Today it’s a proven technology, according to the NYT. Traditionally, most systems that monitor structures’ responses to earthquakes or strong winds have been wired ones. But wireless alerts may one day be an alternative.
Undersea cable operators have a nasty habit of laying cables close to each other. When they get cut, they tend to go in sets. The first question I have is why Maldives would lose 100% of traffic when it is connected by two undersea cables, one to Colombo and the other to India. That’s serious redundancy, especially for a tiny country of 300,000+ people. I can understand the traffic on Reliance’s Flag system going down because it was Atlantic focused.