Part 1: 150-180 word, critical reply to this forum discussion with references.
As explained by Fitzpatrick (2003), "memcached is a high-performance, distributed memory object caching system, generic in nature, but originally intended for use in speeding up dynamic web applications by alleviating database load."
The diagram above shows a simple example of how separate memory sitting on separate web servers can be optimized by Memcached in order to better allocate memory as is needed. The two separate web servers in the first picture each has 64megabyte of memory to spare that is just sitting there. In the second picture it is seen that they are combined and there is now 128 (64+ 64) megabyte memory left to spare.
Memcached provides the ability to shift memory around to where memory is needed instead of just having the memory sit there. The idea behind memcached is brilliant because it provides another layer when it comes to accomplishing high scalability and performance.
Memcached controls reliability because it allows resources to be shared and when it comes to data replication, it is very beneficial because the idea of replication is to have data sitting in two or more databases in a distributed system. For example, "For disaster recovery (DR) purposes, replication typically occurs between a primary storage location and a secondary offsite location" (Rouse, 2015).
Using memcached in distributed system which has replication sitting on it would easily manage memory in a way between the connecting servers. There is a difference between Memcachedb and memcached. Memcachedb has its own key-value database system based on Berkeley DB, which is a software library used to provide high-performance embedded database for key/value data, while memcached is just the memory object caching system.
Part 2: 150-180 word, critical reply to this discussion with references.
Memcached was created by Brad Fitzpatrick in 2003 to be used by LiveJournal. It is an open-source and free platform that bases its architecture in an in-memory key-value storage system. MemcacheDB is a distributed key-value storage system. It is a variant of Memcached and does follow Memcached protocol, so all Memcached clients can connect to MemcacheDB.
This design offers great performance, especially for web applications by lessen the database load. It does use distributed memory object caching system, it supports replication, and it is available in most of the popular languages. Even though this is a very powerful system mainly used in Big Data systems, it is also very simple to deploy and maintain.
Simply explained, Memcached determines in what areas of your system memory is not used or you have more than needed and makes that memory available to other areas where memory is running low or there is less memory available than what the system is actually needing. It is said to be a short-term memory for applications.
The management program can be changed in order to save data in the MemcacheDB and be certain that it will remain there. The biggest difference between MemcacheDB and Memcached is that the former has its own key-value database system. Even though this system seems like a great system to implement in a large scale in order to take advantage of the overall system's memory in a smart and dynamic way, development of this system seems to have come to a halt around 2009.
Some of the Memcached's main users are Craiglist, Wikipedia, Flickr, LiveJournal, and WordPres.com.