Source: Tech Times
A lot of people are so dependent on the Internet that a few hiccups can lead to widespread panic. While it's easy enough to curb the rage when online speeds slow down, it's a different matter altogether when the Internet stops working. That's exactly what happened Tuesday when Internet connection slowing down in various parts of the world and others shutting down entirely for a considerable portion of the day.
While North America was most heavily hit, the entire world felt the effects of the internet basically struggling under its own weight. The problem stemmed from what is known as the Border Gateway Protocol (BGP), a routing table that essentially contains all the pathways in the internet that data uses to travel to get to one point from another.
"BGP is what tier-one ISPs, your last-mile ISP and various large networks use to route data from their own machines to others, and vice versa. When you visit a website, that data bounces all over the world, through machines belonging to all manner of companies and organizations. To make this work, machines called routers (large commercial versions of what you have at home) keep a table of known, trusted routes through the tangled web," explained Matthew Sparkes, The Telegraph's Deputy Head of Technology.
As the Internet grows, the BGP grows as well to accommodate all possible pathways that data can use, allowing routers to pick the best one to facilitate connection in the fastest way possible. Unfortunately, there's a limit to how big the BGP can go and some of the routers can no longer keep up. This limit is 512K.
The 512,000 limit translates to 512 rows in the routing table each with 1,000 ports. Essentially, that translates to 512,000 possible routes data can use to travel around the Internet. In the same way that regular roads can get congested when too many vehicles are using it, the BGP bogs down when routers reach their maximum capacity. The result? Internet access slows down or completely shuts off until "traffic" eases.
While the incident on Tuesday did take Internet users by surprise, the problem itself was not unprecedented. It has been anticipated for years but the solution is expensive and difficult to implement so very little has been done about it. Not that anyone can be blamed--it is hard to overhaul the Internet. Many routers have already been replaced but mostly only those that have already reached their limit. As more eventually meet theirs, the problem that took out the Internet a few days ago will happen again.
Some of those affected by the 512K issue include LastPass, Liquid Web and eBay.
Slow Internet connection? The Web might have a bigger problem
Source: Tech Times
Slow Internet connection frustrated users on Tuesday, but the slowdown wasn't caused by cable problems or data center issues.
Researchers reported that the slow Internet speeds were caused by problems in the structure of one of the core technologies of the Internet.
The technology in question is the Border Gate Protocol routing, which is responsible for bridging smaller, interconnected works so that data traffic can travel all over the world.
"It's kind of like the telephone directory for every Internet Service Provider, showing how it sends traffic to every other part of the Internet," said Renesys senior analyst Doug Madory. "We have a vast, global Internet here, and it can be described in about 500,000 lines, which is kind of amazing."
Renesys released a report on the possible Internet outages that could happen next week, as the number of IPv4 networks in the world reached the number 512K, equivalent to 524,288 or 2 to the 19th power.
This number is the maximum number of routes that is supported by the default TCAM configuration on several older hardware platforms.
Renesys analyst Jim Cowie, however, is quick to point out that the situation is "more of an annoyance than a real Internet-wide threat," as local connectivity problems can be quickly fixed by Internet service providers through upgrades.
The problem, Cowie adds, stems from the growth of the Internet, which just keeps on expanding. Shortages in the space within IPv4 networks continue as new systems from bodies such as companies, schools and network providers connect to them.
However, the shortage of IPv4 space may actually be increasing the growth of the Internet, as companies and service providers pull out all the tricks to be able to operate within the diminishing space within IPv4 networks.
As the first providers traverse the 512K mark for routes, there seems to be no serious signs of higher instability on the Internet. Users who are using affected equipment, either by themselves or by Internet providers, may have noticed early signs of the problem.
As the routing table size continues to increase and maintains a number above 512K, more and more vulnerable gear will be exposed, which will be upgraded and fixed over time.
However, with the massive growth of the Internet, it might not be too long before the number of routes reaches the 1024K mark, or 2 to the 20th power.
Cowie suggests that users and providers can reclaim more TCAM for the IPv4 networks by giving up support for the IPv6, which is the next-generation version of the IPv4 that has not been widely adopted yet.
Browsing speeds may slow as net hardware bug bites
Source: BBC News
Browsing speeds could slow over the next week as old hardware is upgraded to handle the net's growth, says networking giant Cisco.
Some older kit has hit an upper memory limit in the number of routes it can use to despatch data around the world.
In the last few days the number of possible routes breached that upper limit which might mean those routers start to struggle.
Already some firms are reporting users are having problems.
Global issue
The problem has emerged as the number of connections between the different networks that make up the internet has continued to grow, wrote Omar Santos from Cisco in a blogpost.
Routers, which send data around the net, keep track of all the ways data can travel via an internal log known as a routing table. This list is constantly updated according to the ebbs and flow of internet traffic.
This week the number of entries on that global routing table went past more than 524,000. That represents a growing problem, said Mr Santos, because five separate devices Cisco makes can only handle a routing table of 524,288 entries.
As more and more routers around the world have to support 512K entries and beyond, the potential problems will grow, said Jim Cowie from internet monitoring firm Renesys.
"512K is right around the corner for everyone on Earth, as early as next week," he wrote in a blog post, adding, "this situation is more of an annoyance than a real internet-wide threat."
Routers that hit the memory limit could slow down, lose data or crash.
So far, said Mr Cowie, there was little evidence that the 512K problem was bringing about any more disruption than Renesys normally sees.
However, hosting firm LiquidWeb blamed the 512K bug for service disruption that hit it on Tuesday and it is also thought to be instrumental in causing problems for eBay, Comcast and Time-Warner.
Paul Lettington, network architect at UK ISP Andrews and Arnold, said workarounds did exist for the bug that should help older kit cope. Cisco has also published advice for owners of vulnerable hardware.
Andrews and Arnold had only seen indirect evidence of the 512K bug starting to bite, said Mr Lettington.
"We have seen anomalies with other networks on the internet which could have been caused by it, and these may have had an effect on our customers accessing those other networks," he said.
He added; "It is unlikely that any network operators will step forward and say that they were affected by it, as it would require admitting that they are running older, less capable hardware and are not on top of managing the maintenance of it."
Roland Dobbins, a senior analyst at Arbor Networks, said the too-many-routes problem had emerged twice before when the number of routes surpassed 128K and then 256K.
At both times unskilled and smaller organisations were caught out as hardware struggled to cope. A similar situation may recur this time, he said, because skilled engineers who know about the limitations of ageing network hardware were still rare.
"This may come as a surprise to non-specialists who view the internet as a high-tech affair comparable to the bridge of the USS Enterprise of Star Trek fame," he said. "In actuality, the internet is more akin to an 18th century Royal Navy frigate, with a lot of running about, climbing, shouting, and tugging on ropes required to maintain the desired course and speed."
Anyone experiencing problems with connectivity and speed yet?
.jpg)
