In fact, the algorithm delivered at least 1.5x better throughput and 4x lower latency than the best rule-based technique. Reinforcement learning (purple) outperformed all rule-based congestion control algorithms in NVIDIA’s tests. “What was especially gratifying was we trained the model on just 32 network flows, and it nicely generalized what it learned to manage more than 8,000 flows with all sorts of intricate situations, so the machine was doing a much better job than preset rules,” he added. thesis on reinforcement learning at Technion, Israel’s prestigious technical university. “We were like, wow, ok, it works very nicely,” said Dalal, who wrote his Ph.D. A Wow Factorĭalal recalls the meeting where a fellow Nvidian, Chen Tessler, showed the first chart plotting the model’s results on a simulated InfiniBand data center network. The algorithm enabled the team to create, train and run an AI model on their NVIDIA DGX system. Part of their breakthrough, described in a 2021 paper, was coming up with an algorithm and a corresponding reward function for a balanced network based only on local information available to individual network streams. To smooth traffic, the NVIDIA team created new reinforcement learning techniques inspired by state-of-the-art computer game AI and adapted them to the networking problem. To be effective, networks need to respond to situations in about a microsecond, that’s one-millionth of a second.
DIGITAL CLOCK 3D MODEL DRIVERS
That’s a tough balancing act when no one driver on the digital road can see the entire, ever-changing map of other drivers and their intended destinations.Īnd it’s a race against the clock. Networks need to be both fast and fair so no request gets left behind. That’s why Dalal is among many researchers around the world looking for ways to make networks smarter with reinforcement learning, a type of AI that rewards models when they find good solutions.īut until now, no one’s come up with a practical approach for several reasons. It’s basically a set of rules embedded into network adapters and switches, but as the number of users on networks grows their conflicts can become too complex to anticipate.ĪI promises to be a better traffic cop because it can see and respond to patterns as they develop. Networks use congestion control to manage digital traffic. Like rush hour, it’s caused by a flood of travelers angling to get somewhere fast, crowding and sometimes colliding on the way. The senior research scientist at NVIDIA, who is part of a 10-person lab in Israel, is using AI to reduce congestion on computer networks.įor laptop jockeys, a spinning circle of death - or worse, a frozen cursor - is as bad as a sea of red lights on the highway. Gal Dalal wants to ease the commute for those who work from home - or the office.
0 kommentar(er)
