The documents are indexed as text indexes.The data is split across multiple storage devices and in terms of performance disc I/O time is major factor.

Google is working with software development companies to keep the database as in memory database..in this scenerio, whenever there is a hit, it is loaded in as in memory database and cached so that, the search and processing is within the memory that way the access is faster and reduces the disc I/O s...
 

o final - declare constant
o finally - handles exception
o finalize - helps in garbage collection,knlk n;lk nl; l;kn ;lk jl m; m;lml; m;l,

Variables defined in an interface are implicitly final. A final class can't be extended i.e., final class may not be subclassed. This is done for security reasons with basic classes like String and Integer. It also allows the compiler to make some optimizations, and makes thread safety a little easier to achieve. A final method can't be overridden when its class is inherited. You can't change value of a final variable (is a constant). finalize() method is used just before an object is destroyed and garbage collected. finally, a key word used in exception handling and will be executed whether or not an exception is thrown. For example, closing of open connections is done in the finally method.

domains are set up in a hierarchical structure, each segment (between the dots) represents one level of the hierarchy. The right-most is the highest (most significant) segment. Each segment directly to the left of a given segment is said to 'belong to' or to 'be a part of' that segment. For example, with www.networkactiv.com, the most significant domain segment is 'com' (at the right), the next most significant is 'networkactiv' (in the middle), and the least significant is 'www' (at the left). I know the 'www' thing being the least significant may seem counter-intuitive, and this will be explained in a moment. This all means that in the example of www.networkactiv.com, 'networkactiv' belongs to (is a part of) 'com' and 'www' belongs to (is a part of) 'networkactiv'. Whoever controls a given segment also controls the segments to its left. If you own 'something.com', you also own 'www.something.com' and 'anythingelse.something.com'. This is because the DNS server (the computer running the DNS server software) responsible for a given 'something.com' is what handles providing the information of what IP address anything to the left of 'something.com' has. I know I hadn't yet mentioned it, but each composition of a domain name, that is, 'something.something.something' (however many segments) represents one (or sometimes more) IP addresses. This means that 'www.google.com' and 'google.com' could actually represent different IP addresses. What does this mean in practice?... It means that each could have a different Web site because each could represent a completely different computer located in a completely different place. Even if they have the same IP address, it's still possible for them to have different Web sites, but likely that they will be on the same computer, but that will be explained later in this tutorial. The DNS system was designed for more than just Web pages... The 'www' at the left of most domain compositions is supposed to mean, "The IP address of whatever.com that represents the computer hosting the Web site for whatever.com.". Basically, by putting the 'www.' at the left, you are saying that you want the Web site (hence, WWW (World-Wide-Web)) for that domain, and not something else.

ref:  http://www.networkactiv.com/WebSiteAndWebHostingBasicsTutorial.html