THE MAPGRID PROJECT : CACHING POWER
--------------------------------------

AT PRESENT CACHING SUCH AS EHCACHE/MEMCACHE/REDISS CACHING OF DATA HAS NOT BEEN IMPLEMENTED, 
IN THE MAPGRID SESARCH API [ REST API SERVICE ].

HOWEVER IN THE FUTURE, WE HAVE AN OBJECTIVE/PLAN TO IMPLEMENT HIGH SPEED DATA CACHING AT THE 
REST API level.


---------------------------------------------------------------------------------------------

The MAP SEARCH and TEXT SEARCH ENGINES - ALLOW SEARCH WITH 4 PARAMETERS.

[ GREEN DROP DOWN + BLACK DROP DOWN + BLUE DROP DOWN + PAGE NUMBER ]


THUS THAT ALLOWS US TO MAKE A SPECIAL CACHE PATTERN :
 
# [ Green category + Black category + Blue drop down [ usa, japan, germany..] + Page number ]

E.g [ Events + Tech/ENGG Event + USA + PAGE NUMBER 1 ] = A Cache pattern

The cache object will store the THIS CACHE PATTERN in the LIVE MEMORY OF THE CACHING SERVER,
along with the search data/value.

WHEN A SEARCH QUERY IS INITIATED : THE CACHE PATTERN STORED in the MEMORY is checked First, for 
the search result.

IF The search pattern exists in the cache, a Mongo Database Search Query CALL is not required.

If, HOWEVER a Mongo call is required, the mongo db - data fetch is immediately cached to make the 
system, a High speed memory based caching system, ALLOWING all subsequent search queries to work, 
directly from the cache server, rather than hitting the backend mongo db database for each query .


----------------------------------------------------------------------------------------------

THIS makes The MAP GRID a extremely High speed/ HYPER SPEED Search System.


If the application scales : We can easily store the Entire planets information/projections in the
High Memory Cache server.

TO SUPPORT THIS caching : There are many many high memory systems in the Market:

AIX/POWER SYSTEMS - POWER 775 
BLUE GENE Q - [ 24 Terabytes per Rack ].
ORACLE/FUJITSU : M32 [ 32 TERABYTE Memory in  One rack ]

With this LARGE AMOUNT of MEMORY, IN A  single system image LEADS TO THE ARCHITECTURE WHERE,  The entire 
VIASATF PLATFORM DATA CAN BE CACHED in MEMORY - LEADING TO A HYPER SPEED / SUB MILLI SECOND SEARCH SERVICE.


FINALLY: We do understand that  SUPER COMPUTING SYSTEMS Are expensive. If only small comodity
servers are available due to BUDGET CONSTRAINTS - There are memory virtualization systems, to aggregate 
main meory into a SINGLE LARGE SHARED MEMORY.[ The firm that does that : is called SCALE MP ]

----------------------------------------------------------------------------------------------