Article URL: https://jdan.github.io/98.css/#status-bar Comments URL: https://news.ycombinator.com/item?id=49028927 Points: 4 # Comments: 1
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Read-through and write-through are two Redis caching patterns that shift data-loading and write logic from the application into the cache layer itself. In read-through caching, the cache automatically fetches missing data from the database on a miss, centralising the loading logic rather than repeating it across the application. Write-through caching ensures that every write updates both the cache and the database synchronously, keeping them consistently in lockstep at the cost of higher write latency. Together, the two patterns let applications treat the cache as their primary data interface, eliminating scattered invalidation logic. However, the approach is best suited to read-heavy workloads where consistency is critical, and may be overkill for simpler use cases where cache-aside suffices.
Microsoft Azure allows users to create Linux virtual machines through its cloud Infrastructure as a Service (IaaS) platform, eliminating the need for physical hardware. Users can set up a VM via the Azure Portal by configuring a resource group, selecting Ubuntu as the operating system, and enabling SSH and HTTP inbound ports. After deployment, it is recommended to extend the public IP idle timeout from the default 4 minutes to 30 minutes to avoid connectivity disruptions. The VM can then be accessed remotely from a local terminal using an SSH command with the assigned public IP address. This setup is particularly suited for beginners using a free-trial Azure account for learning or testing purposes.
Redis sets are unordered collections of unique strings that enable instant membership checks and server-side set operations such as intersection, union, and difference. Commands like SISMEMBER run in constant time, making sets ideal for use cases including permission checks, deduplication, follower relationships, and unique visitor tracking. Unlike lists, sets automatically ignore duplicate entries, eliminating the need for application-side deduplication logic. Operations like SINTER and SUNION allow Redis to compute relationships between sets atomically, replacing loops in application code with single fast commands. For cases requiring only approximate unique counts at scale, Redis offers HyperLogLog as a memory-efficient alternative to full sets.
Redis lists are ordered string sequences that support push and pop operations from either end, making them well-suited for basic job queues and recent-item feeds. Producers can add jobs using LPUSH while workers retrieve them via RPOP or the blocking BRPOP, which eliminates inefficient polling by making workers wait until a job is available. A key limitation is that plain list-based queues offer at-most-once delivery, meaning a job is lost if a worker crashes after popping it but before completing the task. The LMOVE command partially addresses this by atomically moving a job to a separate processing list, enabling crash recovery through manual bookkeeping. For use cases requiring strict delivery guarantees and acknowledgements, Redis Streams are the recommended alternative, while lists remain a practical choice for best-effort queues and bounded recent-activity feeds.
Discussion (0)
Log in to join the discussion and vote.
Log in