 everyone. Welcome to the session on big data technology landscape. In this one we will be studying NoSQL in detail. At the end of the session you will be able to list types of NoSQL databases. You will be able to compare and differentiate between what exactly SQL and NoSQL are there, what is the difference in those one and where to use which part. Basically this big data technology landscape it is supporting under the two important technologies. One is the NoSQL technology and another one is Hadoop. And in this session we will be studying more on NoSQL and Hadoop session I will be conducting later in the next video. Let us see in detail this NoSQL. See what is this NoSQL? NoSQL meaning is not only SQL. It is not exactly NoSQL. It is not only SQL means it is supporting the features more than SQLs. Talking about the history of it actually this is introduced in 1998 from Carl Stauge and even but it is later on widely used by or reintroduced in 2009 by Eric Evans. What this NoSQL is there? Actually this NoSQL is supporting the lightweight thing. It is open source and it is a non-relational database. Very important is it is not a relational database. It is not a traditional RDBMS. So the standard SQL interface are not used in this NoSQL. It is popular because it is supporting scale out property and it is dealing with rich variety of data. The data may be structured, semi-structured, unstructured, whatever the variety of data is there that is supported by this NoSQL. Now where it is used? Where to use this NoSQL? We are using this for big data analytics but basically for the real-time applications you can say real-time any applications are there social media data is there web applications are there to analyze them this NoSQL is using. Why we are using NoSQL for this one? Because whenever the real-time data is there whenever the social media data is there we know that that data is not in a structured format that is in the unstructured format. We want to deal with that one using this NoSQL and then other usage is it may be used for log data like stock analytics is there or whether analytics is there whatever so in that analytics part also it is used. It can be used for time-based data also where the time is very important login logouts the times are there the check-ins are there. Time-based data analytics there also this NoSQL is used because usually the RDPMS is not doing analyzing such part of thing. It is not having the supportive functions for these one. Now what is NoSQL? Let us see few features in this one additional features the basic feature is non-relational it is not relational database it is totally a non-relational database. What kind of database it is supporting? It is supporting the database in the form of key and value where one key is provided and the value associated that is given according to that one that is in the pair form Next comes a document-oriented data where the data can be treated as one complete set where the document itself the data is talking about a metadata also means a data and a metadata both are combined in this one. Next comes a column-oriented data then a graph-based database also we will see those one how exactly these are there in detail later. It is distributed next comes it is distributed okay why because it is supporting several nodes in a clusters are working with this one so it is work totally the distributed database you can see. Another thing it is not supporting asset properties okay means all the asset properties like atomicity, consistency, idempotent and then even durability all these are not supported by this one it is supporting a brevers theorem cap theorem and the video for this one already I have done you can refer that video what exactly cap theorem is then no fixed table schema is provided okay then no fixed table schema is provided by this one flexibility is there okay the schema flexibility is there the data where data types are also flexible so no fixed data types are provided by this one so these are the additional features. Now what exactly no SQL is going towards more detail in this one as we have seen that non-relational data storage system is there no fixed table schema is there as no fixed table schema is there it is not easily handled the join operations and all that so no join operations are there here no multi document transactions are provided and it is totally relaxing the asset properties. Let us talk about the types of no SQL databases basically no SQL databases is converted into the two things that is key value pair or you can say the big hash table also and a schema less database this key value pair or big hash table is supported by Amazon S3 scalars scalar is also and schema less it is of column based document based graph based many more let us see one these one in detail more detail before that one just pause the video and think about what exactly schema less means why we can say schema less so these are the schema less types of the databases let us see one by one these one the first one is talking about a key value the volatile and key value persistent both the types of database are supported so here you can see these are the keys and the values are support given by this one so there is a link that a key is provided okay like one two three is a key and the value here is something different one two six is there and the value provided is different why because the key is provided here but the data for the value may be vary therefore no schema is required for this value next comes wide column or a column based data here what what happens that every column is having its own identification you can see here okay and then the every column is supported by few data which are having along with the metadata with that one okay so this particular thing this is called as a column here so this column is containing the data as well as the metadata required for that even this also you can see and you can see the variations between them next comes the document based data in this one what happens it is provided using this opening and closing parenthesis we have given here so here also the data and the metadata is also given by this one so every data is supported by its attribute name also so that is one kind of database no SQL database next comes a graph based database where in a graphical format the interconnections the relations are provided so these are few no SQL databases so the key value format is used by dynamo db voltamot or scalaris the document based database is supported by MongoDB couch DB column based data is supported by big table Cassandra H base graph based data is used by Neo4j and infogrid few more are given here you can see here that graph based is supported by FlockDB or column based is supported by big table and Cassandra our document based is by MongoDB and couch DB and key value is supported by react ready servers, scalaris and all these so a few examples are provided here that key and value more their details so these are the keys and the values are different here so this is one example of key value pair database next comes the document based data here the document is having this as the personal data so with the tag as person then the first name is provided by this one so the first name is the attribute name and the value is given as beta next comes attribute as last name and then the value is provided so automatically this this is a structured format where addresses are there in that street is also provided another address is also provided here multi-value data you can say so like this the document document based data is provided next comes column based data and graph based data so in column based data you can see here that these are called as the column family so these one and this one is a row ID you can see one all these are there provided title is provided for the column and the details are there and as well as it is having a combination of time and value also like this the column based data is to similarly for the graph based data you can see that how these are interconnected and it is represented in this way these are the types of no SQL databases now why no SQL because it is supporting scale out architecture and it is storing large volume of data with the variety like structured semi-structured unstructured and everything supporting a dynamic schema so it is inserting it is allowing to insert any kind of data with the pre even we can have a predefined schema or no schema also okay and even if it is supporting a faster development easy code integration is there it is requiring less database administration therefore no SQL is required auto shorting is also provided by this what is auto sharding here it is breaking the data into small blocks and automatically spreading working and automatically again combined okay if a server is going down it is immediately replaced by another one okay and it is supporting replication so it is guaranteeing high availability it is guarantee fault tolerance as well as a disaster recovery these are few advantages of no SQL you can see cheap and easy to implement easy to distribute can easily scale up and scale down relax the data consistency doesn't require a predefined schema data can be replicated on multiple loads and can be partitioned these are few references thank you