Both solutions will be analyzed based on the requirements and criteria listed above. During the analysis, each solution will be given a score from 1 to 10 according to how well it meets each criterion. In the end, a overall score will be calculated base on the score and the Criteria Weighting Scale listed in Section 4 on page 2 for each solution. The solution that has the higher score will be the more feasible solution. From here, the solution that preloads the local cache will be called solution 1, and the solution that uses a shared cache will be called solution 2.
Criteria 1 - Read and write operations:
Both solutions can support read and write operations. Solution 2 always allows write operation. However, solution 1 only allows one write operation which is populating the local cache during the start up of the application. Therefore, the scores of solutions 1 and solution 2 are 8 and 10.
Criteria 2 - Data Consistency: …show more content…
But data consistency between the cache and the database becomes an issue when an broker catalog entry is updated, because the entry was updated in the database but not in the cache. When a user requests to get the broker catalog which is just updated and existing the cache, the user will get the older version of the data from the local cache. Therefore, data consistency is hard to achieve by using solution 1 in a multi-instance application. Solution 1 scores a 5 in data consistency. Whereas, in the solution 2, data is always consistent, because all the instances are using one shared cache and shared cache will be updated after updating the database. Solution 2 scores a 10 in data