Oracle database for linear scalability and data sovereignty / localization

August 18, 2022 | 3 minute read
Ram Posham
Strategic Client Executive Architect, Oracle
Text Size 100%:

For over two decades Oracle has been a huge proponent of database consolidation, and to support the strategy, Oracle created required technology solutions such as RAC (Real Application Clusters) in 2001, and Exadata (an engineered system that comes with storage, network, and compute, pre-integrated and tuned to run Oracle RAC databases) in 2009 for scalability and availability/resiliency, Data Guard and GoldenGate for disaster recovery and active-warm/active deployments, respectively.

Oracle in early 2000 consolidated many of its business application instances that were spread across the world into few and saved billions of dollars.  Here is an OCT 2, 2000 article, titled “Oracle shoots for $2 billion in savings through e-business”, in ComputerWorld that provides more details of the story.  Here is an Oracle OpenWorld presentation, dated Sep 20, 2010, titled “How to Reduce: 10 Benefits of Instance Consolidation in Oracle E-Business Suite”, which describes the benefits of consolidation. As hardware and software systems became more reliable, applications and databases consolidation has been a popular solution for lowering overall IT costs for many years. But …

As countries become more nationalistic and/or like to monitor money transactions for fraudulent activities such as money laundering, they would like their citizens’ data to be stored locally in their countries. For instance, in April 2018, Reserve Bank of India (RBI) issued a directive asking all payment system operators (PSOs) to store entire payment data in India so it can conduct fraud, money laundering, and other investigations locally. This article provides more details on how credit card companies such as Amex and Master Card were forced to comply with the regulation. 

One solution to meet the above regulatory requirement is to create a replica of database and application infrastructure in the country, however it is a not cost-effective and complete solution. The cost of infrastructure and operations doubles, complexity increases significantly, and slows down innovation.

An elegant and efficient solution would be to have databases deployed in the countries where data localization / sovereignty is a requirement, manage the countries-specific data in the databases, and have the application, with minimal/no-changes, continue to run where it was originally deployed. Oracle database offers this solution! The following diagram describes the future state architecture to meet data sovereignty/localization requirement of one country, and the same architecture can be extended to multiple countries; each country requires its own shard (local database). If the application knows the shard key based on say, user credentials/profile, it can identify the country to which the data belongs to, make connection to the shard and perform database operation using direct routing of SQL requests. If the application does not know the shard key, it can connect to the shard coordinator (the catalog database) and the coordinator will in turn proxy-route the SQL requests. The important thing to note here is, this requires minimal/no changes to the application. This type of sharding is known as user-defined sharding.

Oracle Database Sharding for Localization


Oracle database can not only be used to meet data sovereignty/localization (aka geo-distribution) requirements but also to meet scalability requirements where the data is distributed across a set of shards based on consistent hash value of shard key. This type of sharding is known as system-managed sharding.

Oracle also supports a third type of sharding, called composite sharding where geo-distribution is done based on shard key 1, known as super_sharding_key, and consistent-hash-value based sharding within the geography based on shard key 2, known as sharding_key.

To learn more about Oracle database sharding, you can watch my following presentations and demos on youtube.

  1. Oracle database sharding overview – here. Copy of the presentation referred in the video is posted here.
  2. Oracle database sharding for scalability - here
  3. Oracle database sharding – user-defined sharding for data sovereignty/localization demo - here


You can find more technical details on Oracle database sharding on and youtube presentation here.

Ram Posham

Strategic Client Executive Architect, Oracle


Ram Posham is currently a Strategic Client Executive Architect with Oracle with a
passion to help Oracle’s strategic enterprise customers modernize their IT
infrastructure and innovate faster with Oracle Cloud Infrastructure (OCI). Ram has
been working for Oracle for over 25 years, and during this period he helped large
enterprise customers such as BofA, State Street, Lowe’s, LabCorp, AT&T, Comcast,
etc., maximize their investments and innovate with Oracle technology products and
solutions, which are available both onprem and in OCI, by leveraging the core
architectural and foundational capabilities of the products and solutions. Ram is
Togaf 9, Oracle Cloud and AWS certified architect.

Previous Post

Recent New Graph Features in Autonomous Database

Rahul Tasker | 3 min read

Next Post

Microsoft Azure and Oracle Database on OCI: A Big Win for Users

Guest Author | 5 min read