Alibaba 2024 Backend Salary Ranges and Essential Interview Preparation Guide

Alibaba's 2024 campus recruitment backend salaries are reported as follows: "白菜" 26K, "小 SP" 27‑28K, "大 SP" 29‑30K, "SSP" 31‑32K. The Alibaba group offers 16 months of salary, Ant Group 15 months, and Cainiao 14 months, with some departments providing a 2 k housing subsidy.

Compared with other companies such as Xiaohongshu, JD.com, and ByteDance, these salaries are slightly higher, though the exact figures may vary when the hiring scales up.

The technical interview typically consists of two or three rounds (rarely four); a strong technical performance leads to an HR interview, which is crucial and should not be considered a guarantee of success.

Project presentation tips for candidates include: (1) summarise the project in one sentence covering core business and pain points; (2) highlight key features and high‑value components; (3) describe the overall architecture and justify technology choices; (4) clarify personal role, responsibilities, and concrete contributions using the STAR method; (5) discuss challenging technical problems, analysis, solution options, and results; (6) demonstrate deep understanding of critical technologies such as Seata configuration and transaction modes.

For high‑concurrency inventory deduction, a Redis Lua script is used to pack the read‑check‑write steps into a single atomic command, preventing overselling. However, the script cannot roll back partial writes on error and loses atomicity in Redis Cluster when keys span multiple hash slots.

Cache‑database consistency is addressed with the Cache‑Aside pattern: reads first check the cache, fall back to the database on miss, and writes update the database then delete the cache entry. If cache deletion fails, two mitigation strategies are suggested: shortening the TTL (not a fundamental fix) or implementing a retry mechanism, preferably via a message queue for asynchronous retries.

Redis provides a built‑in slow‑log feature to record commands exceeding a configurable latency threshold. Configuration parameters slowlog-log-slower-than and slowlog-max-len can be set in redis.conf or via CONFIG SET. Each log entry includes a unique ID, timestamp, duration (µs), command with arguments, client IP:port, and optional client name.

MySQL slow‑query logging is enabled by setting slow_query_log = ON, specifying the log file, and adjusting thresholds such as long_query_time and min_examined_row_limit. An example slow query on a million‑row table without an index demonstrates high execution time, and the resulting log entry is explained field by field. Tools like mysqldumpslow and EXPLAIN help analyse and optimise such queries.

Java reflection allows runtime inspection and manipulation of classes, methods, and fields, enabling frameworks like Spring, MyBatis, and Hibernate to perform dependency injection, annotation processing, dynamic proxies for AOP, and ORM mapping. A sample dynamic proxy implementation using InvocationHandler illustrates how method calls are intercepted and delegated.

The differences between short polling, long polling, and WebSocket are compared: short polling sends periodic HTTP requests (simple but high latency and resource waste), long polling holds the request until data arrives (better latency, still resource‑intensive), and WebSocket establishes a persistent full‑duplex TCP connection after an HTTP Upgrade (lowest latency, efficient, but more complex to implement).

Server‑Sent Events (SSE) and WebSocket are contrasted: SSE offers unidirectional server‑to‑client communication over standard HTTP/HTTPS with automatic reconnection, suitable for text streams; WebSocket provides bidirectional communication over a dedicated ws:// / wss:// protocol, supporting both text and binary data but requiring explicit connection management.

A comparison table outlines TCP vs UDP characteristics, including connection orientation, reliability, statefulness, efficiency, header overhead, communication mode, and typical applications.

Finally, HTTP and HTTPS are compared: HTTP uses port 80, plain text transmission, and no built‑in authentication, while HTTPS (port 443) runs over TLS/SSL, encrypts data, authenticates servers, incurs higher resource consumption, and offers SEO advantages. An illustrative diagram highlights the differences.