ICT OCR A Level - Distributed Databases

ICT OCR A Level - Distributed Databases

1. What are they?

   • A single logical databse
   • All parts are physically in different locations
   • Parts are accessible over a network.
   • A standard Database looks like:
   • A typical database looks like:
   • Each site has a Database Management System (DBMS)
     • Allows queries to be made locally
     • Also has a Distributed Database Management System
       • Allows queries to be made globally for the whole system

2. Why use them?

   • Examples of uses:
     • Travel company with branches across the globe
     • International car manufacturer
     • Supermarket stocks with multiple branches
   • Some companies may need information up to date all of the time
     • Access the data globally once a user has updated the files.
   • Distributed databases reduce the risk of losing work
     • Central database is updated overnight
       • Batch processing
     • If the network goes down, then work is not lost and employees can work as normal until the issues are fixed.
       • Mainly independant; if one branch goes down, the others aren't affected

3. Duplicated Databases

   • Local copy of each part of the entire database is kept at each location
   • Not very practical for a number of reasons.
   • Disadvantages.
     • Very heavy use of the network
     • Whole database may be large and require a lot of storage space
     • Difficult to keep all copies up to date simultaneously
     • Does not scale well if the company grows
     • Wasteful if only certain parts are needed at certain branches
   • Advantage
     • More practical if the company is fairly small

4. Partitioned Databases

   • Database is 'split' or partitioned into different datasets depending on the specific needs of the organisation
   • Advantages
     • Each data store is independant
     • Network load is much less
     • Higher performance since there is no network bottleneck
     • Easier to solve problems due to each store being independant
     • Scales up very easily
   • Disadvantages
     • May not be the best solution if each branch needs constant updates of data
     • Needs to be carefully partitioned to keep each section as local as possible
   • Partitioned+Index
     • A modification of the partitioned database is to include an index of all remote database records as well
     • Each store has an index pointing to all records in the entire database
     • Advantages
       • Still a high performance as most queries remain local
       • Can efficiently access remote records by using the index
     • Disadvantages
       • System must keep all indexes up to date
       • More complicated to create and manage

5. Overall

   • Security
     • Multiple entry points, this can make the database less secure. Each 'node' needs to be kept secure
     • Data encryption keys are more likely to be intercepted if they are spread around the organisation constantly
     • If one node is hacked then the rest of the organisation is vulnerable as well
   • Advantages
     • Resilient, if one part of the database stops working, there rest are not affected
     • Security: each employee is able to access one portion of the data
     • Network traffic is minimised so there is less bandwidth expenses
     • Easier to keep any errors local
     • High performance due to less network bottleneck
   • Disadvantages
     • More complex to create and maintain
     • Many entry points to the whole database is less secure
     • Data integrity is harder to maintain to ensure the indexes are not corrupted
     • Not as effective if there is heavy interaction between branches

Imagens e arquivos do mapa