Hi all,

It took me much longer than I expected to deliver the first version of
my project that I am happy with. But finally version v0.1 is here: 

https://github.com/ademakov/MainMemory/releases

I expect next versions to come out much faster.

Regards,
Aleksey


On Tuesday, February 23, 2016 at 10:44:38 PM UTC+6, Aleksey Demakov wrote:
>
> Hi all,
>
> I've been silently working on my own memcached clone for quite some
> time. Apart from implementing the memcached protocol I wanted to get
> a reusable framework for other kinds of servers, so in a way memcached
> was a testbed for a network library competing e.g. with libevent and libuv.
>
> The project is located here: https://github.com/ademakov/MainMemory
>
> I am about to make the very first release of the project fixing the last
> known issues. It is not yet ready for real life use, but just to play 
> around
> with it.
>
> For those who cannot find a better use of their time I welcome if they try
> it and report any problems, issues, suggestions.
>
> The key features of the framework:
>
> * co-operatively scheduled fibers running on top of OS-level threads;
> * concurrent event loop with lock-free thread notification mechanism;
> * memory allocation mechanisms that try to avoid locks as much as
>   possible.
>
> The memcached-related features are nothing special, just regular text
> and binary protocol via TCP. The internals are a bit different however:
>
> * no slabs, all items are allocated in a single dlmalloc space;
> * use clock eviction algorithm instead of LRU;
> * there is experimental code that uses ring-buffer-based combiners
>   instead of locks for synchronization.
>
> Generally, IMHO, my version is written in a more modular fashion than
> the original memcached. So it should be more open for those who are
> interested to modify it for e.g. their research purposes and such.
>
> Regards,
> Aleksey
>
>

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"memcached" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to