RaptorDB is ServiceNow's high-performance database engine, based on HTAP (Hybrid Transactional/Analytical Processing) architecture, which allows you to run transactional and analytical queries simultaneously on the same live dataset—eliminating the need for ETL pipelines and data replication. Released with the Xanadu release in September 2024 and available in Standard and Pro versions, RaptorDB replaces MariaDB as the primary engine of the Now Platform.
Why did ServiceNow replace MariaDB?
For over a decade, the Now Platform operated on MariaDB—a robust fork of MySQL, but designed for an era when data volumes and analytical workloads were substantially smaller. With the growth of enterprise implementations, predictable limitations began to appear: analytical queries running concurrently with transactions created contention at the database level; complex joins on large tables like the CMDB progressively degraded; and Performance Analytics reports and dashboards would expire during peak times.
In 2021, ServiceNow acquired Swarm64—a Berlin-based startup specializing in high-performance extensions for PostgreSQL—and spent three years integrating this technology into the core of its platform. The result is RaptorDB, built on PostgreSQL with proprietary extensions optimized specifically for ServiceNow workloads.
What is HTAP and why does it matter for ITAM and CMDB?
HTAP (Hybrid Transactional/Analytical Processing) is the architecture that allows a single database engine to process both transactional operations (incident creation, CI updates, asset registrations) and analytical queries (inventory reports, license compliance dashboards, trend analysis) on the same data simultaneously—without requiring separate systems or synchronization via ETL.
For ITAM and CMDB managers, the impact is direct:
- Filters in large CMDB tables They now return results in seconds, not minutes.
- SAM compliance reports They run on real-time data, not nightly snapshots.
- Performance Analytics Dashboards They load with up-to-date data, eliminating the lag that compromises operational decisions.
- Hardware reconciliations (HAM) Lifecycle analyses can be performed ad hoc without impacting ongoing transactions.
Technical architecture: how RaptorDB is faster
RaptorDB internally maintains two representations of the same data: one row-store for transactional operations (traditional format, optimized for access to individual records) and a column-store For analytical operations (columns stored together, optimized for aggregations and scans). The engine routes queries to the correct internal representation automatically—without the administrator or developer needing to modify existing GlideRecord queries or workflows.
Parallel processing breaks down complex queries into smaller tasks executed simultaneously, instead of processing them row by row sequentially. This combination—columnar indexing + parallelism—is what delivers the performance gains documented by ServiceNow with early adopters.
| Metric | Result |
|---|---|
| Improved transaction time | Up to 53% faster |
| Reports, analytics, and list views | Up to 27x faster |
| Transactional throughput | Up to 3x larger |
Source: ServiceNow Data Sheet — RaptorDB Professional (2024). Benchmarks performed on production instances of enterprise clients.
RaptorDB Standard vs RaptorDB Pro
RaptorDB is available in two versions with distinct scopes and licensing models:
| Feature | Standard | Pro |
|---|---|---|
| Availability | Included for new customers; gradual migration for existing customer base. | Available to new and existing customers (additional SKU) |
| HTAP Engine | Yes | Yes |
| Column-store indexing | Basic | Advanced (ultra-scale) |
| Parallel processing | Yes | Yes, on a larger scale. |
| Live Archive | No | Yes — fully accessible historical data. |
| Live Connect (SQL access for BI) | No | Yes — direct access to external BI tools. |
| Agency AI workloads | Basic support | Optimized for machine-scale and Now Assist |
| Cost | No additional cost. | Additional license |
For organizations that operate large volumes of CMDB, run intensive SAM/HAM workloads, or are expanding their use of generative AI with NowAssist, RaptorDB Pro delivers the necessary scalability without compromising the performance of day-to-day operations.
How to check if your instance has already migrated to RaptorDB
Migrating to RaptorDB Standard requires no administrator action—ServiceNow performs the transition at the platform level during the release upgrade process. To check if your instance is already on RaptorDB, go to [link to RaptorDB Standard page]. replication.do In your ServiceNow instance, check the JDBC driver string field:
jdbc:mysql://— instance still in MariaDBjdbc:postgresql://— instance already in RaptorDB
Operational impact for ITAM and CMDB teams
For ITAM and CMDB managers, RaptorDB represents a paradigm shift in how the platform handles data at scale. Environments with dense CMDBs—hundreds of thousands of CIs, complex relationships, integrations with Discovery and Service Mapping—were historically the most affected by performance degradation. With RaptorDB, this structural bottleneck is addressed at the architecture level.
This opens up practical possibilities that previously required complex workarounds: real-time asset reconciliations without maintenance windows, on-demand configuration drift analyses, and license compliance reports based on the current state of inventory—not yesterday's data. For organizations on their ITAM maturity journey with ServiceNow, migrating to RaptorDB removes one of the main technical obstacles to adopting advanced analytical dashboards and preparing the database for agency AI workloads.
The 4MATT, ServiceNow Elite Partner in Brazil and winner of the Technology Excellence Partner Award 2024–2025, monitors implementations of CMDB ITAM in enterprise environments and supports technical teams in preparing instances to take advantage of RaptorDB's performance gains — including reviewing queries, data structures, and migration strategy to RaptorDB Pro when volume justifies the investment.