● Covered by 2 sources · 2 reportsMedium impact2 neutral
Cloudflare's 1.1.1.1 DNS service optimizes cache, saving 100TB memory
Key points
- Cloudflare optimized DNS cache entry storage for 1.1.1.1.
- Five changes reduced memory footprint by over 50%.
- 100 terabytes of memory were saved across the fleet.
- Cache insert throughput increased 43%, lookup latency dropped 19%.
- Each DNS cache entry was reduced from 953 bytes to 420 bytes.
- The platform behind 1.1.1.1 is named Big Pineapple.
- The changes were made at the Rust-level.
- The memory savings are equivalent to 130 of Cloudflare's 768GB Gen 13 servers.
- Cloudflare's 1.1.1.1 resolver holds over 250 billion cached DNS entries.
- Replacing Vec and String containers with fixed-size boxed slices saved over 15TB.
DNS Cache Optimization
Cloudflare's Big Pineapple platform, responsible for services like 1.1.1.1 DNS, implemented five successive changes to how it stores DNS cache entries in memory. This platform manages over 250 billion DNS cache entries at any given time, making memory efficiency critical.
Memory and Performance Gains
The optimizations resulted in a reduction of the per-entry memory footprint by over 50%. This translated to approximately 100 terabytes of memory freed across Cloudflare's server fleet, equivalent to the RAM in 130 Gen 13 servers. Beyond memory savings, the cache also saw performance improvements, with insert throughput increasing by 43% and lookup latency decreasing by 19%.
Caching Mechanism
The cache starts empty and fills as DNS queries arrive, evicting older or less popular items when it reaches its maximum capacity. Cache size varies by data center, especially with EDNS Client Subnet (ECS) in use, which can increase the number of entries and memory consumption due to caching multiple versions of the same query. Each cache item is a key-value pair, where the key identifies the query (qname, qtype, authenticated, tag) and the value stores the DNS response along with metadata like timestamp, TTL, and hit count.
Benchmarking Methodology
To measure the impact of each change, Cloudflare benchmarked memory usage by filling the cache with randomly generated entries. These entries mimicked production traffic distribution, consisting of 56% A records, 25% AAAA records, and 19% TXT records, with each entry containing between one and four records. TXT records served as a stand-in for other variable-length record types, with sizes randomized between 64 and 224 bytes.
Updates
- Each DNS cache entry was reduced from 953 bytes to 420 bytes.
- The platform behind 1.1.1.1 is named Big Pineapple.
- The changes were made at the Rust-level.
- The memory savings are equivalent to 130 of Cloudflare's 768GB Gen 13 servers.
- Cloudflare's 1.1.1.1 resolver holds over 250 billion cached DNS entries.
- Replacing Vec and String containers with fixed-size boxed slices saved over 15TB.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Primary sources
CVE CVE-2023-503877.5 HIGHHow outlets covered it
Cloudflare optimized the memory layout of DNS cache entries in its 1.1.1.1 resolver, reducing each entry from 953 bytes to 420 bytes. This change freed up approximately 100TB of RAM across its global server fleet and improved cache performance, with insert throughput up 43% and lookup latency down 19%. The memory reclamation addresses rising DDR5 prices and allows Cloudflare to scale its DNS service more efficiently.
Cloudflare's Big Pineapple platform, which powers the 1.1.1.1 DNS service, implemented five changes to its DNS cache entry storage. These optimizations reduced the per-entry memory footprint by over 50%, freeing up approximately 100 terabytes of memory across its fleet and improving cache performance.