High-performance, document-oriented database

BSON

 {
	"_id" : ObjectId("51267154e4b05bd0461ed426"),
	"timestamp" : NumberLong("1361473875951"),
	"payload" : [
		{
			"oid" : "1.3.6.1.2.1.1.3.0",
			"value" : "0:00:11.00",
			"snmpType" : "TimeTicks"
		},
		{
			"oid" : "1.3.6.1.6.3.1.1.4.1.0",
			"value" : "1.3.6.1.6.3.1.1.5.1",
			"snmpType" : "OID"
		}
	]
}

Are indexable aggregates of Documents

Groups Collections together

(usually used by the same application)

Are like SQL tables

But they have no fixed schema!

They can contain documents with different shapes

JavaScript command line

    > db.getCollectionNames()
[ "system.indexes", "traps" ]
> db.traps.findOne()
{"_id" : ObjectId("51267154e4b05bd0461ed426"),
 "timestamp" : NumberLong("1361473875951"),
 "payload" : [
  {
   "oid" : "1.3.6.1.2.1.1.3.0",
   "value" : "0:00:11.00",
   "snmpType" : "TimeTicks"
  },{
   "oid" : "1.3.6.1.6.3.1.1.4.1.0",
   "value" : "1.3.6.1.6.3.1.1.5.1",
   "snmpType" : "OID"
  }]}
> 

Query by similarity

 > db.messages.find({from:"mario"})
{
          "_id" : ObjectId("5187b06ba504d801ed2c70f6"),
          "from" : "mario",
          "to" : "luigi",
          "text" : "hello mario"
        }
>

Updates: atomic, create if not exists, create or update.

 > db.getCollectionNames()
[ "system.indexes" ]
> db.messages.save({ from: "mario", to:"luigi", text:"hello mario"})
> db.getCollectionNames()
[ "messages", "system.indexes" ]
>

Drivers for almost every language.

Replica-Set: one Primary, many Secondary

Drivers makes connection to the Primary transparent.

When the Primary node goes offline, the remaining Secondaries will automatically start an Election for a new Primary.

Easy setup

 >rs.initiate()
 >rs.add("hostname2:27017")
 >rs.add("hostname3:27017", {arbiterOnly:true})

Support for Google's Map-Reduce

An example:

> var mapFunction = function() {
    emit(this.from, 1);
  }
> var reduceFunction = function(keySender, occurrences) {
    return Array.sum(occurrences);
  }
> db.messages.mapReduce(
    mapFunction,
    reduceFunction,
    { out: "map_reduce_example" }
  )