Translating into Русский. The English version is shown until it is ready.
Yandex open-sources YTsaurus Flow for real-time data processing
Коротко
Yandex has released YTsaurus Flow, a technology for continuously processing large streams of data, into open source, according to TASS, CNews and vc.ru. It handles clicks, likes and other online events as they appear, which the company says lets algorithms react faster to changing user interests and detect fraud sooner. Yandex says the technology is already used in several of its services, including Market, antifraud systems and Advertising.
Читать полностью 2 мин чтения
Yandex has released YTsaurus Flow, a technology for continuously processing large streams of data, into open source, according to TASS, CNews and vc.ru. The technology handles clicks, likes and other internet events as they appear, the outlets reported. CNews said the announcement came from Yandex representatives. [ 1 , 2 , 3 ]
CNews described the problem the technology addresses: user interests change constantly, so algorithms need fresh data on what a person bought, listened to or asked about. It said such data is usually accumulated first and processed later, like parcels waiting for a train, leaving delays that can reach hours, while YTsaurus Flow reduces that delay almost to zero by processing data as it arrives. [ 2 ]
According to the same report, the technology is designed for high loads and processes each event exactly once, without losing or duplicating data. It keeps working during failures by redistributing load among available servers. CNews said Yandex presents it as useful wherever data freshness matters, including preparing data for machine-learning model training, updating information for sites and applications, and counting views, clicks and orders, as well as in advertising and recommendation services. [ 2 ]
Yandex said YTsaurus Flow is already used in several of its services: at Market it helps collect statistics faster so sellers can react to changes in demand, in antifraud systems it helps detect malicious activity, and in Advertising it helps select relevant offers, according to CNews. The report also said the technology lets algorithms recognise the actions of fraudsters more quickly and pick advertising that is useful. vc.ru likewise noted that Yandex Market uses the technology for collecting statistics. [ 2 , 3 ]
TASS reported the release in a brief item, describing YTsaurus Flow as technology that processes clicks, likes and other internet events as they appear. Neither it nor vc.ru gave additional detail on the open-source terms or the scope of the released code. [ 1 , 3 ]
Почему это важно
Yandex is making part of its data infrastructure openly available, which could let outside developers and companies build real-time processing into their own products. The company ties the technology to faster recommendations, advertising and fraud detection inside its own services, indicating that speed of data processing is a focus for it.
Ключевые факты
- Yandex released YTsaurus Flow into open source, according to TASS, CNews and vc.ru. [ 1 , 2 , 3 ]
- The technology processes clicks, likes and other online events continuously, as they appear. [ 1 , 2 ]
- CNews, citing Yandex representatives, said conventional processing accumulates data first, causing delays that can reach hours. [ 2 ]
- Yandex said the technology processes each event exactly once, without losing or duplicating data. [ 2 ]
- Yandex said the technology keeps working during failures by redistributing load among available servers. [ 2 ]
- Yandex said YTsaurus Flow is already used in services including Market, antifraud systems and Advertising. [ 2 ]
- vc.ru noted that Yandex Market uses the technology for collecting statistics. [ 3 ]
Подтверждено несколькими источниками
Пока неясно
- The licence terms and scope of the open-source release are not specified. All three documents say the technology was released openly, but none describes the licence or what code is included.
- The full list of Yandex services using YTsaurus Flow beyond the named examples. Only Market, antifraud systems and Advertising are mentioned, in a single tech outlet report.
- Yandex’s statements about how the technology performs in its own services. These claims come from the company and are reported via its representatives rather than independently verified.
Что пишут местные СМИ
Хронология, местное время
- TASS reports that Yandex has released real-time data processing technology YTsaurus Flow into open access. [ 1 ]
- CNews reports that Yandex has open-sourced YTsaurus Flow, citing company representatives on its functions and existing uses. [ 2 ]
- vc.ru reports the open-source release, noting the technology is used at Yandex Market for collecting statistics. [ 3 ]