ANFIMAU | Data intelligence system migration

Overview

Our customer, a financial & risks solutions provider, offers its clients a Data Intelligence System that utilizes industry available merge & acquisitions, equity, debt and syndicated loans transactions, fee and rankings enabling preparation of strategic business planning information, benchmarking of competitors as well as analysis of a market financial state and trends identification across assorted product types, geographic areas and sectors.
Due to company technical strategy changes there is an immediate need to migrate the existing system from on-prem to Azure public cloud and modernize the system leveraging PaaS offerings to benefit from native scalability and reduce maintenance efforts.

Challenge

Existing solution has high dependency on legacy technologies like SSIS and SQL Server Analysis Service that creates risks and prevents from smooth migration & modernization using Azure PaaS offerings.
There is an intent to uplift from OLAP cubes to Tabular models as part of migration & modernization which is associated with potential technical challenges and additional efforts.
Migration should run smoothly for existing clients without disruption thus requires intelligent approach for data migration and preparation for failover scenarios in case of any issues.

Approach

The delivered solution is based on Azure PaaS technologies and leverages following approaches:

OLAP to tabular model transition

Migration from OLAP cubes to Tabular models hosted on Azure Analysis Service.

SSIS package execution with Azure data factory

Leveraging Azure Data Factory with SSIS Integration Runtime to execute existing SSIS packages.

Web service modernization and .NET Core upgrade

Migration of web services from VMs to Azure App Services and upgrade to .NET Core.

Optimizing data Mart management with Azure SQL managed instances

Usage of Azure SQL Managed Instance as execution engine for SSIS packages and hosting of existing client Data Marts that has tight coupling to some of the features that are not available in SQL Azure.

PoC execution results did not identify any technical risks and as next step a detailed modernization & migration architecture was prepared to be used for uplifting of the current system.

Modernization approach included PaaS usage, switch to Tabular models, data migration on per client basis with A/B testing, failover strategies in case of incidents as well as decommissioning of on-prem hosted system.

Proving modernization feasibility with Azure PaaS

Feasibility of modernizing current system leveraging Azure PaaS offerings and switch to Tabular models from OLAP cubes was proved by execution PoCs in the beginning of engagement.

Architecting and executing a modernization roadmap

Prepared detailed modernization architecture using PaaS services and Tabular models, executed modernization with existing clients migration and decommissioning of on-prem system in 8 month.