Shulou(Shulou.com)11/24 Report--

CTOnews.com, October 27 (Xinhua) it is well known that Musk has not only renamed the platform X, but also made drastic changes to the development business and laid off a large number of staff since his acquisition of Twitter.

Today, platform X officially released its annual development summary, showing the improvement in the background over the past year and boasting as "a year of engineering excellence". The attached contents of CTOnews.com are as follows:

Integrates the technology stacks of For you, Following, Search, Profiles, Lists, Communities and Explore around a single product framework.

The "For you" and ranking system has been completely rebuilt from scratch, reducing lines of code by 90 per cent from 700K to 70K, computing footprint by 50 per cent, and post throughput per request by 80 per cent.

For you and video personalization and ranking model are unified, and the quality of video recommendation is significantly improved.

Refactored the API middleware layer in the technology stack and simplified the architecture by removing more than 100,000 lines of code, thousands of unused internal endpoints, and clearing unselected client services.

The latency after metadata acquisition is reduced by 50%, and the global API timeout error is reduced by 90%.

The ratio of intercepted robots to content crawlers is 37% higher than in 2022. On average, the X platform blocks more than 1 million robot registration attacks every day, and DM spam is reduced by 95 per cent.

Shut down the Sacramento data center and reconfigure 5200 racks and 148000 servers, saving more than $100 million a year. A total of 48 MW of capacity was released and 60,000 pounds of network racks were removed before reconfiguring it to other data centers.

Optimized use of cloud service providers and started to do more preprocessing. This shift reduces monthly cloud costs by 60%. Officials have moved all media / blob artifacts out of the cloud, reducing the overall cloud data storage size by 60%. In addition, officials have successfully reduced cloud data processing costs by 75%.

Based on the prem GPU supercomputing cluster, a new network architecture of 43.2Tbps is designed, developed and delivered to support the cluster.

Expanded backbone network capacity and redundancy, saving $13.9 million a year (CTOnews.com Note: currently about 102 million yuan).

Initiate automated peak traffic failover testing to continuously verify the scalability and availability of the entire platform.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.