Re:A new fast hash table in response to Google’s new fast ha...
by josefx on Tuesday May 29, 2018 @02:01PM
(#17181515)
I think they implemented that change to deal with a denial of service attack in http servers. A malicious client could send a large list of parameters with an identical String::hashCode() value, resulting in a single large bucket. Other workarounds included a size limit on the parameter list or a hash function with some parameters randomized at program start.
[ Parent
| on HN ]
There are no comments on this comment yet.