Click here to Skip to main content
15,885,366 members
Please Sign up or sign in to vote.
1.00/5 (1 vote)
See more:
Hi and thank you for clicking my question.

I've been studying web development recently and have an important project in mind. It would basically be a social-network( I know what you might think: borring we have facebook) but I plan to give this a twist that could change our meaning of 'like' and 'friend'.

I've decided to put all the not so dynamic data into a MySQL relational database that could be able to handle relationships like followers/following, general relationships with the website structure and personal data.

Since I plan the site to be able to have feeds like twitter's (fan-out on write) and knew MySQL wasn't the best option to handle that type of data I decided to incorporate Hadoop-Hbase to handle it.

So I plan having tables in the MySQL database that only have indexes not actual data, about who will receive who's content and handle the actual data on HBase making use of the indexes already existing in MySQL; that way I could handle relationships in a nice manner and not overwhelm MySQL with BigData.

My question: Is this a good approach? What would you do?

Take into account that when I say content I refer to a post type of object that may include comments,likes and other details.
Posted
Updated 4-Mar-14 16:19pm
v2
Comments
Sergey Alexandrovich Kryukov 4-Mar-14 22:23pm    
I will be good or bad approach if you describe it. So far, this is a post about nothing.
Well, the "indexes not actual data" sounds like a really bad idea. How about integrity?
—SA
Member 8437747 5-Mar-14 21:55pm    
Sorry if the question isn't so clear. My system has a many to many relationship with a 'groups' table(a user might be subscribed to many groups and a group might have many subscribers), since I know managing all the content(post with lots of comments, likes, dislikes, etc) in MySQL is more complicated, I thought HBase would be a good alternative to save this data because of its BigTable design that suits the twitter-style feed I want to have. The approach I think of is having my intremediate many-to-many relationship table to serve as a list of unique keys I could use in HBase, this way I can extract data like 'how many subscribes in a group' more easely in MySQL and HBase would only handle the delivery of data. Since I have no experience on BigData or HBase, I would like to know what would be your approach. Thanks for answering by the way.
PIEBALDconsult 4-Mar-14 22:45pm    
Probably not so good. "Bid gata" doesn't just mean "a lot of data".
Member 8437747 5-Mar-14 21:57pm    
Sorry if the question isn't so clear. My system has a many to many relationship with a 'groups' table(a user might be subscribed to many groups and a group might have many subscribers), since I know managing all the content(post with lots of comments, likes, dislikes, etc) in MySQL is more complicated, I thought HBase would be a good alternative to save this data because of its BigTable design that suits the twitter-style feed I want to have. The approach I think of is having my intremediate many-to-many relationship table to serve as a list of unique keys I could use in HBase, this way I can extract data like 'how many subscribes in a group' more easely in MySQL and HBase would only handle the delivery of data. Since I have no experience on BigData or HBase, I would like to know what would be your approach. Thanks for answering by the way.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)



CodeProject, 20 Bay Street, 11th Floor Toronto, Ontario, Canada M5J 2N8 +1 (416) 849-8900