About eight years ago, Illinois computer science professor Brighten Godfrey and his collaborators formed a research team to dramatically increase Internet speeds.
Initially, Godfrey and this team thought they could develop an Internet with latency near the speed of light. Their idea represented not only an improvement in the performance of the Internet, but an overhaul of how it works and is built.
Most know that their Internet Service Provider (ISP) provides some level of bandwidth, but the speed of a networked application depends on more than that. Even the smallest message takes a while to reach its destination and for a response to come back.
“It’s called latency, and it turns out that many interactive applications – like web browsing, cloud gaming, and sometimes even video conferencing – experience lag that often stems more from latency than bandwidth,” Godfrey said.
Supporting an energy and belief that Internet latency could be dramatically improved, Godfrey and the team recently published a paper – titled “cISP: A Speed-of-Light Internet Service Provider” – that proved their theory was possible.
Introducing Godfrey’s Fun Side Project – an online escape room based on CS principles
In early March, Brighten Godfrey also unveiled one of his side projects – something he described as “a rather atypical escape game” which incorporates CS principles.
The missing links came from Godfrey and co-creator Renata Cáceres — a professor at Illinois Wesleyan University School of Music — during the peak months of the pandemic, when opportunities for camaraderie and fun were hard to come by. Each enjoyed the typical escape room experience, but felt that the online versions left a little to be desired.
“In-person escape rooms offer a whole world of possibilities,” Godfrey said. “You can interact with anything in the room, and there’s enormous freedom and creativity in the design of the puzzle. One of our ideas was to recreate some of that magic in an online version.
The story revolves around a team of players who go to visit a friend in her laboratory, until they discover that she has disappeared. While in the lab, an AI agent asks them to help fix the network, in which case they must discover the missing network links – hence the name of the game.
“The puzzles themselves can be solved by anyone, but many of them incorporate computer concepts,” Godfrey said. “We drew inspiration from areas such as graph algorithms, networks, logic, and NP completeness.”
Players came from all over the world, with the fastest “escape” time records currently held by a team here at the University of Illinois at Urbana-Champaign and a team of engineers from Google.
Godfrey and Cáceres have made collaborative play free for educators in hopes that more students can understand the creativity possible through CS.
“We hope it’s a great way for people to get together and that it can serve as an activity associated with computer education programs,” Godfrey said. “A lot of software engineers have played with friends, but it’s also great for students, even high school kids, to play. is CS and it can be fun.
The Missing Links isn’t the only science-themed game associated with the University of Illinois Urbana-Champaign. In the in-person escape room LabEscape, players solve puzzles based on the principles of physics.
Presented at the 19th USENIX Symposium on Designing and Implementing Networked Systems on April 6, the paper showed that “cISP instantiations in the United States and Europe would achieve average latencies less than 5% of those achievable using orthodromic trajectories at the speed of light. , over medium and long distances.
Godfrey noted that this is a very forward looking project.
“The IT industry develops amazing technologies, many of which address more immediate needs. What I believe academic research shines on – and what I love to pursue – are fundamental innovations in architecture. And this project is about as fundamental and architectural as it gets,” Godfrey said. “He says, ‘Let’s build a whole new Internet. “”
To design this parallel Internet, researchers from Illinois teamed up with collaborators from Yale University, Duke University and ETH Zürich.
They believe the resulting service could be rolled out to 120 of the largest US cities, giving 85% of Americans a much faster internet connection.
Initial skepticism stemmed from concerns about feasibility and cost-effectiveness. The group persevered and answered questions.
“The important thing our paper does is show, yes, it’s doable as a service and reasonably profitable,” Godfrey said. “It now allows us to focus on more exciting things like future applications of a low latency internet. There may be great opportunities for applications like distributed augmented reality or extended reality.
To make such a concept a reality, Godfrey and the team focused on creating a new methodology that addresses both bandwidth and latency.
Bandwidth, the professor said, has improved dramatically on the Internet, but it’s latency that’s much more difficult to improve.
“Computing and communications hardware has improved bandwidth, so millions of people can watch 4K UHD video at home whenever they want. But latency depends on physical distance, which means you can’t not improve latency in the same way,” Godfrey said. “Getting closer to the physical limits of the speed of light and ultimately producing an internet that is within 5% of the speed of light was intentionally a very difficult.
“It’s exciting in retrospect, because you don’t necessarily know where a project like this is going to take you when you first start it.”
Where do you start when you have a project with an uncertain outcome?
To begin, the group measured the current Internet. One of the key metrics is something called round-trip time (RTT), which is the time it takes for a device on the internet to send a small packet of information and receive a response from a machine. distant. They found that the RTT was typically 3-4 times longer than if the messages were traveling at the speed of light, and was often 80 or 100 times slower than the speed of light.
Next, the group then identified three main reasons for this delay.
First, the network does not follow a straight line as the crow flies, but rather the fiber laid underground, which is usually diverted. Second, ISPs don’t necessarily optimize for latency when routing packets across the network, so packets don’t even follow the shortest path through available fiber. And third, the physics of the material through which the information is transmitted is important: the speed of light in the glass – of which the fiber is made – is about 1.5 times slower than the speed of light in the air or vacuum.
These factors led to the choice of a different physical medium for the Internet at the speed of light: wireless directional microwave transmissions between cell towers. This sends information through the air rather than through the glass and makes it easier to build straight line paths from tower to tower. But such technology – which has recently been boosted by high-frequency commercial use – has much lower bandwidth than fiber.
Their next idea was to use two networks in parallel.
“Today’s Internet does a great job of providing bandwidth; it’s the right way to stream video on demand, for example. Then a second parallel network will be really good at delivering low latency, which is what our new design does,” Godfrey said. “This allows for a system that can intelligently send latency-sensitive traffic to the latency-sensitive network, and the rest to the high-bandwidth network. The document includes an algorithm for deciding how to split traffic between the two networks.
“And the cool thing is that by accelerating a small fraction of overall traffic for applications like web browsing – around 10% or less – it can result in a huge benefit for the application.”
Godfrey explained that the success of this project became possible thanks to a diverse group in their specializations and deep in terms of student involvement.
Student co-authors include Debopam Bhattachejee, ETH Zürich; Waqar Aqeel, Duke University; Iker Nadi Bozkurt, Duke University; William Sentosa, CS of Illinois; and Muhammad Tirmazi, Harvard University.
In addition to Godfrey, faculty contributors included Sangeetha Abdu Jyothi, UC Irvine; Anthony Aguirre, UC Santa Cruz; Balakrishnan Chandrasekaran, VU Amsterdam; Gregory Laughlin, Yale University; Bruce Maggs, Duke University; and Ankit Singla, formerly of ETH Zurich and currently at Google. Abdu Jyothi and Singla are also Illinois CS alumni.
“We knew this was going to require several different types of expertise, including areas outside of IT,” Godfrey said. “It took a sustained effort from the team over several years.”
The project was funded by the National Science Foundation and a donation from Google.