Nominum’s Asia Pacific channel partner, Acom Networks, recently conducted performance testing for a Tier 1 client in Taiwan to better understand Nominum’s Vantio CacheServe performance versus BIND9. Here’s their findings in their own words:
Vantio CacheServe outperforms BIND9 using just 2 cores. See the test results and find out how you can save OPEX using Vantio CacheServe.
We set up a comparison using the following hardware:
- HP DL380 G8
- Single Intel Xeon E5-2667-V2 (3.3GHz) 8core processor
- 64GB Ram
- 300GB SAS 10Krpm drives x 4 in a RAID10 configuration
- CentOS 6.5
Both Nominum CacheServe and BIND9 enabled full logging capabilities. CacheServe has this function by default, but BIND9 had to be configured to log all DNS packets.
We stress tested both DNS services using another server running queryperf. Nominum was licensed to utilize only 2 CPU cores while BIND9 was using all 8 CPU cores.
And now, the results.
Maximum QPS (Query Per Second) Performance:
- Full packet logging was enabled for both Nominum’s CacheServe and BIND9
- Nominum’s Vantio CacheServe was testing using 2~4 cores, BIND9 was using all 8 cores
- 10% cold cache and 90% hot cache
DNS Response Latency:
From the above figures, we can conclude that Nominum’s Vantio CacheServe brings faster performance per server than using BIND9.