This is a fork of https://github.com/golang/groupcache with adding some extra new features. I recognize the effort of https://github.com/mailgun/groupcache but here provides a slightly different kind of features:
- support of gRPC instead of HTTP
- upgrade proto file syntax to proto3
- cleanup some small dependencies -- for example, use golang.org/x/sync/singleflight instead of its own implementation.
- support of kubernetes deployment, where
- the list of peer names are not known a priori
- the peer set can change dynamically
groupcache is a distributed caching and cache-filling library, intended as a replacement for a pool of memcached nodes in many cases.
For API docs and examples, see http://pkg.go.dev/github.com/jmuk/groupcache
- shards by key to select which peer is responsible for that key
-
does not require running a separate set of servers, thus massively reducing deployment/configuration pain. groupcache is a client library as well as a server. It connects to its own peers, forming a distributed cache.
-
comes with a cache filling mechanism. Whereas memcached just says "Sorry, cache miss", often resulting in a thundering herd of database (or whatever) loads from an unbounded number of clients (which has resulted in several fun outages), groupcache coordinates cache fills such that only one load in one process of an entire replicated set of processes populates the cache, then multiplexes the loaded value to all callers.
-
does not support versioned values. If key "foo" is value "bar", key "foo" must always be "bar". There are neither cache expiration times, nor explicit cache evictions. Thus there is also no CAS, nor Increment/Decrement. This also means that groupcache....
-
... supports automatic mirroring of super-hot items to multiple processes. This prevents memcached hot spotting where a machine's CPU and/or NIC are overloaded by very popular keys/values.
-
is currently only available for Go. It's very unlikely that I (bradfitz@) will port the code to any other language.
In a nutshell, a groupcache lookup of Get("foo") looks like:
(On machine #5 of a set of N machines running the same code)
-
Is the value of "foo" in local memory because it's super hot? If so, use it.
-
Is the value of "foo" in local memory because peer #5 (the current peer) is the owner of it? If so, use it.
-
Amongst all the peers in my set of N, am I the owner of the key "foo"? (e.g. does it consistent hash to 5?) If so, load it. If other callers come in, via the same process or via RPC requests from peers, they block waiting for the load to finish and get the same answer. If not, RPC to the peer that's the owner and get the answer. If the RPC fails, just load it locally (still with local dup suppression).
groupcache is in production use by dl.google.com (its original user), parts of Blogger, parts of Google Code, parts of Google Fiber, parts of Google production monitoring systems, etc.
See http://talks.golang.org/2013/oscon-dl.slide
Use the golang-nuts mailing list for any discussion or questions.