Uber Unveils its Realtime Market Platform
Matt Ranney, Chief Systems Architect at Uber, gave an overview of their dispatch system, responsible for matching Uber's partners, i.e. drivers, and riders. Ranney explained the driving forces that led to a complete rewrite of this system. He described the architectural principles that underpin it, driven by availability and performance, several of the algorithms implemented and why Uber decided to design and implement their own RPC protocol.
The old dispatch system was designed for private transportation (1 driver, 1 vehicle, 1 rider) and built around the concept of moving people, but Uber wants to enter new markets, such as goods transportation. Uber also sharded its data by city. As Uber grew and expanded to ever more cities, some of them quite big, it became difficult to manage its data. Finally, the dispatch system had multiple points of failure, a consequence of Uber's hectic growth and the system's scramble to keep up with that growth.
The new dispatch system has two major services: supply, the drivers, and demand, the riders. These services track all the capabilities and the state machines of supply and demand. For instance, the supply service knows how many seats a vehicle has or if it can fit a wheelchair. The dispatch system has a third service, called Disco (Dispatch Optimization), whose main function is to match supply and demand. Disco enables Uber to "look into the future" and to use information as it comes in. For instance, the old dispatch system only looked to current available supply. As most partners are usually busy, this approach allowed Uber to maintain a global index. The new dispatch system is more efficient, but it requires much more data. Uber wants this new system to handle one million writes a second and a much higher read rate, so it needed to shard its data.
Read full article from Uber Unveils its Realtime Market Platform
No comments:
Post a Comment