Data is a set of words, patterns, numbers etc..... It is raw in nature. Information is extracted from data by following some standard procedure based on the business knowledge is about taking decisions based on the extracted information.
Companies like google, Facebook, twitter, linked in have started a new business model of acquiring customers by providing completely free service. But they have to pay charges for their hardware and pay salaries for their employees. This forces these companies to make money from sources like advertisement.
The problem of understanding and categorizing customers or audience will come into picture because of this.
Eg :- If a person likes 3000 times in a day on an average then total number of likes will be 1 million approximately per year. To analyse these many likes taking an average of 1mb of data per like we have to process so much data.
In computer we have CPU, RAM and HardDisk in regular computer we can extend RAM upto 64gb. So,to process 1tb of data we need machine with 1tb of RAM, but the motherboard should replaced with 1tb supported RAM.
To improve this vertically computer scientist came up with super computer likes mainframes. But there will be a limit on the hardware based on advanced electronics and also super computer are very expensive. Hence analyzing big data with traditional hardware is a problem.
Big data analogy :- The ballet of US president election has to be conducted. If a single person has taken up the task of counting all the ballet papers it will be time consuming and very tired some that is also a big data problem. To solve this problem google came up with idea of scaling horizantally. In the give example of presidental elections if a team of people divide ballet into equal parts and count them and came up with individual intermediate number then they can count the ballet number to get the result.
To solve the problem of big data google published two white papers out of which one describes GFS (google file system) and another GMR (google map reduce). In this solution data will be distributed across several machines and analyse individually and come up with intermediary solution. These results will later be crunched again to come up with final result. The implementation of this solution is known as Hadoop.
Hadoop :- Hadoop is the implementation of two white papers published by google and it is the solution of bigdata. GFS is implemented as hadoop distributed file system(HDFS). Where as google map reduce was implemented as MapReduce.
Cluster is as set of multiple nodes (computers, servers etc.....).
Virtual Machine :- To share physical resources with out any conflict by multiple machines virtually virtual machine software will be helpful.
Eg :- VMware, oracle-virtual box.
Companies like google, Facebook, twitter, linked in have started a new business model of acquiring customers by providing completely free service. But they have to pay charges for their hardware and pay salaries for their employees. This forces these companies to make money from sources like advertisement.
The problem of understanding and categorizing customers or audience will come into picture because of this.
Eg :- If a person likes 3000 times in a day on an average then total number of likes will be 1 million approximately per year. To analyse these many likes taking an average of 1mb of data per like we have to process so much data.
In computer we have CPU, RAM and HardDisk in regular computer we can extend RAM upto 64gb. So,to process 1tb of data we need machine with 1tb of RAM, but the motherboard should replaced with 1tb supported RAM.
To improve this vertically computer scientist came up with super computer likes mainframes. But there will be a limit on the hardware based on advanced electronics and also super computer are very expensive. Hence analyzing big data with traditional hardware is a problem.
Big data analogy :- The ballet of US president election has to be conducted. If a single person has taken up the task of counting all the ballet papers it will be time consuming and very tired some that is also a big data problem. To solve this problem google came up with idea of scaling horizantally. In the give example of presidental elections if a team of people divide ballet into equal parts and count them and came up with individual intermediate number then they can count the ballet number to get the result.
To solve the problem of big data google published two white papers out of which one describes GFS (google file system) and another GMR (google map reduce). In this solution data will be distributed across several machines and analyse individually and come up with intermediary solution. These results will later be crunched again to come up with final result. The implementation of this solution is known as Hadoop.
Hadoop :- Hadoop is the implementation of two white papers published by google and it is the solution of bigdata. GFS is implemented as hadoop distributed file system(HDFS). Where as google map reduce was implemented as MapReduce.
Cluster is as set of multiple nodes (computers, servers etc.....).
Virtual Machine :- To share physical resources with out any conflict by multiple machines virtually virtual machine software will be helpful.
Eg :- VMware, oracle-virtual box.