One of the areas in Knowillence that we have built a good level of expertise in data warehouse migration to cloud. In this post I will share some details on what triggered the clients to initiate this data warehouse migration projects apart from financial reasons.

Transactions timeout: The transactions executed from variety of applications started to put a big dent in the performance of transactions database. It is because the data getting into the database has a higher volume, velocity and variety compared to the past. This is often not tracked by the team and they upgrade the hardware and optimize the servers continuously until the law of diminishing returns kicks in. If you are getting timeout in transactions then it is high time you evaluate how to migrate your data warehouse cloud platform that provides auto-scaling (both upscaling and downscaling) and elasticity in processing power, memory usage and storage.

Reports timeout: The reports used by the business users and IT users starts to timeout. When a new project is launched with a new database then this database is used for both transactions and reporting. Over the years when the data grows exponentially then both transaction processing and report generation takes a big hit in performance. The main reason for this problem is that data in transaction database must be highly normalized for performance where are data in reporting database must be denormalized for performance. Hence using transaction database for reporting and reporting database for transaction will not yield good performance or good benefits.

Schedule driven processes: People don’t realize this as a problem until they start to think about real-time analytics. All the jobs in the database are driven by schedule and not by events. Companies start with schedule driven jobs because the are so few jobs and dependencies between then can be documented and easily tracked. When the number of jobs grow in the database server that’s when people lose track of dependencies and when they shuffle the schedule then they break the dependency between the jobs and it leads to chaos. It is better to build a decoupled system which is event driven using a powerful middleware or messaging system.

Replication of data to other sources: The biggest blocker of some applications from being event driven is that the enterprise architecture used replication of data to other sources as one of the primary means to transfer data from one system in the company to another. This affects the workflow of other departments and becomes the biggest blocker when developing analytics solutions for the company. Using data lake and other services you can push data in parallel to various system at the same time.

If you have faced any of these issues and need help in migrating your data warehouse to cloud then reach out to us via the live chat or enter a comment and we will discuss possible solutions with you.

H.Thirukkumaran is the Founder and CEO of Knowillence Pvt Ltd. He is an expert in emerging technologies like cloud, big data and analytics, blockchain and artificial intelligence. He is the author of the book Learning Google BigQuery.