Distributed database design using CAP theorem


CAP is a theorem/standard which can be applied while designing distributed systems like distributed databases. A distributed database is a set of databases(nodes) which are spread across the network. 
These set of databases(nodes) in the distributed database can be either designed as either a master/slave or a sharded model.

Consistency - Database should return the latest write on the DB.
Availability - Database should return data in a decent amount of time. But, the data might not be latest.
Partition tolerant - In spite of n/w failure, the database should always be available.

A distributed database can be designed either as a CP, AP, CA but not with all the 3 capabilities. 

AP - Available and Partition tolerant  -  Sharded model (Data is distributed, No Master)  - 
    Availability - Because, the data is distributed, any of the nodes can return the data. So, always available
    Partition tolerant - If any node goes down, the other nodes can step up and returns the data since the data is distributed.
    Consistency  - Since there is no master and all the nodes are working collaboratively work to store the data.  Your DB might not return the latest write when requested. Because the consistency arrives once all of you nodes come to sync.  Though it takes less amount of time for all the nodes to be in sync, consistency is sacrificed during that time.
MORE READs, FEWER WRITEs databases can use this model



CP - Consistency and Partition tolerant - (Shraded/Distributed model but one node will be master)
    Consistency - Since, you have a master, your writes always goes through the master. So, you always have the latest read. 
    Partition tolerant -  If any node goes down, the other nodes can step up and returns the data since the data is distributed.                        
    Availability  - If the master goes down while writing, from the remaining nodes, one will be elected as a master, until that time the DB availability will be poor means delay in returning the data. 
MORE WRITEs, FEWER READs databases can use this model


CA - Consistency and available (Master and slave (slaves syncs data from master), not sharded)
    Consistency - Since, you have a master, So, you always have the latest read.
    Availability -  As long as the network is available, the DB is also available.
    Partition tolerant  -If the master  DB n/w is down, the system is down. (Until your slave is ready and in sync with master DB).
NON-PARTITIONED(single node) databases can use this model.


So, every model has a tradeoff. To met 2 aspects, you will have to relax the other aspect.

Comments

Post a Comment

Popular posts from this blog

SQL Analytical Functions - Partition by (to split resultset into groups)

Partition by:  Splits the resultset into groups.  Employees -Table EMPLOYEE_ID FIRST_NAME SALARY ----------- -------------------- ---------- 100 Steven 24000 101 Neena 17000 102 Lex 17000 103 Alexander 9000 104 Bruce 6000 105 David 4800 106 Valli 4800 107 Diana 4200 select e.employee_id, e.department_id, salary, row_number() over (partition by department_id order by salary desc) as rownumber from employees e; The above query splits the result into small subgroups for every department.  Syntax :  Function(arg1,..., argn) OVER ( [PARTITION BY <..>] [ORDER BY <....>]  [<window_clause>] ) partition by  - create  subgroups for e...
Continue Reading »

Web Security

CSRF(Cross site request forgery) Attacker crafts a request to a website that the victim has access to. Victim is tricked to submit the request which he doesn't intend to. Victim tricked to changing  his personal information with attackers mobile/email. How To Fix: This attack can be protected by including csrf token in web forms which keeps changing for every request and every user. Referrer header is another option. Session fixation Attacker creates a session by accessing a malicious site. Persuades victim to login to the site with the same session id.  Attacker uses the same session and impersonate victim How To Fix: This attack can be prevented by creating a new session whenever user logs in. HSTS(HTTP Strict transport security) If the https protocol is omitted in the url, victim will be allowed to access the un-secured site. Since, the communication is not secured via http, victim will potentially be vulnerable to man in the middle attacks. Attacker can view the netw...
Continue Reading »