CacheCrow is a simple key-value, LFU, time-based cache. It is thread safe and lightweight.

Now a SecondaryCacheProvider is avaiable and using that a custom cache can be used to replace the default filesystem. Create a new class (let's call it RedisCustomCache) and inherit ISecondaryCache<K, V>. There are 2 ways to register this class.

1. Add a configSection to app.config like below

<configSections>
  <section name="cacheCrow" type="System.Configuration.NameValueSectionHandler" />
</configSections>
<cacheCrow>
  <add key="SecondaryCache" value="TestWebsite.Cache.RedisCustomCache`2, TestWebsite" />
  <add key="SecondaryCacheExpireTime" value="500000" />
</cacheCrow>

OR

2. While initializing the cache

var redisCustomCache = new RedisCustomCache(CacheExpireInMilliseconds);
ICacheCrow<string, string> cache = CacheCrow<string, string>.Initialize(redisCustomCache, 500, 100000, 200000);

Refer to DefaultSecondaryCache on how to implement ISecondaryCache.

CacheCrow internally deploys 2 cache storage to maintain data. This helps maintain the performance of the cache. The 2 caches are:

• Active CacheCrow
• Dormant CacheCrow

As the name suggests, it is the currently active cache. It uses in-memory to store data entries. It has a limited size that is user defined. At initilization, the entries in dormant cache, having most lookup hits are loaded in active cache. As the time goes, data entries are expired, removed from active cache and new entries having most lookup hits are loaded in-memory from Dormant CacheCrow.

Data entries that have least lookup hits resides in dormant cache. All data is stored on disk by default (can be changed with SecondaryCacheProvider), in a file '_crow\CacheCrow'. It has no limit on size and can grow upto total disk size. As data entries are in dormant cache, they cannot expire on their own. For removing expired entries 'Cleaner' is employed that is called every x milli-seconds and its value is only configurable at initialization time.

Data addition can takes place in 3 conditions:

• If user is adding new data to CacheCrow, it is first inserted to active cache i.e. if it can accomodate as it has limited size. Else it is inserted into dormant cache.
• If new space has been accomodated in active cache due to data expiry/removal of data. In this situation, data entry in dormant cache having most lookup hits is loaded into active cache.
• Auto swapped when key is searched/value is fetched.

Data removal takes place in 3 conditions:

• If user is removing data from CacheCrow, then corresponding entry is removed from Active or Dormant cache.
• Data expiry also leads to removal.
• Auto swapped when key is searched/value is fetched.

When a data entry in dormant cache is looked up for or its value is fetched then LFU is performed, the data entry with least hits/frequency in active cache is compared with data entry with most hits in dormant cache. If the former has less hits then the two entries are swapped between active and dormant cache. Auto data swapping also works while adding data entries to active cache, however for new entries it works as stated above in section Add Remove data.

Internally CacheCrow relies on Dictionary<> and it takes a down hit in performance when the size increases. To maintain its performance 2 different caches have been deployed, one with a fixed size(Active) having frequently used data and the other with a variable size.

Use Visual Studio Powershell command

$ Install-Package CacheCrow

OR Download Package

// initialization of singleton class
ICacheCrow<string, string> cache = CacheCrow<string, string>.Initialize(1000);

// adding value to cache
cache.Add("#12","Jack");

// searching value in cache
var flag = cache.LookUp("#12");
if(flag)
{
    Console.WriteLine("Found");
}

// removing value
var value = cache.Remove("#12");

Any small to medium .Net desktop/Web application can make use of CacheCrow for performance boost.

Please raise an issue or drop me a message at [email protected]