Payments are experiencing issues due to temporary restrictions in Russia. If your payment does not go through, please submit a support request.Our support team is available 24/7 — we are always here to help with hosting and server issues.We are now accepting requests for dedicated server rental and colocation services in our data center.Reminder: we recommend enabling backups for additional data protection.A new VPS/VDS lineup with NVMe storage and improved performance is now available.Maintenance work on some servers has been completed. All services are operating normally.
Article3 min read

Cloudflare Tests Cache Compression: When Disk Savings Help Hosting and When They Hurt

In a Cloudflare experiment, cache data took up less space thanks to Zstandard. I analyze why compression must be evaluated alongside CPU load and latency.

Server drives and a spare disk on the work surface
In this article

On September 1, 2026, Cloudflare announced an experiment compressing its own cache content. The approach uses Zstandard: suitable data is stored more compactly and decompressed before delivery. Since the company describes this as a prototype, it should not be treated as a ready-to-use setting for any website.

For a hosting provider or large online store, the key principle here is that disk space can be exchanged for additional computational work. Whether this trade-off is beneficial depends on the nature of the data and where the system is already hitting a bottleneck.

Not all files compress equally

Text data and pre-processed photos behave differently. Re-compressing a format that is already well-packaged may yield little benefit, while the CPU still spends time processing it.

Therefore, before enabling a new scheme, analyze what occupies the storage: original images, derived thumbnails, logs, temporary files, or page caches. Sometimes the main issue is not the storage format, but the uncontrolled accumulation of old copies.

Start with a small, representative sample. Compare file size, read/write times, and resource usage under multiple concurrent requests. A good result for a single file does not describe the entire catalog.

Cache compression and page acceleration are different tasks.

Internally storing compact data can reduce disk usage and inter-node transfer. However, visitors judge the time until the page is ready. If decompression adds latency in an overloaded system, local savings may not improve their experience.

For online stores, the difference between a cache hit and a first request after cache clearance is critical. In the latter case, the server may need to fetch data, generate the page, and prepare it for storage. This mode should be tested before a sale or a major catalog update.

Personal data should not be cached blindly. Shopping carts, individual prices, and user accounts require rules that prevent mixing content from different users. A new packing algorithm for these rules does not replace the need for such safeguards.

Evaluate the effect on the entire system.

Include disk usage, CPU load, response time, and operational costs in your final comparison. Reducing one metric is only beneficial if it does not create a more expensive problem elsewhere.

The Cloudflare experiment indicates the direction of engineering efforts, not a universal recipe. For a specific site, I would first look for measurable overspending and test a limited hypothesis. This approach helps improve infrastructure without promising that a single setting will speed everything up immediately.

Discussion0

Share your experience and ask questions. Comments without links appear after editorial review.

No comments yet. Start the discussion.