OCI Network Monitoring adopts Autonomous Database for performance at scale
Oracle Cloud Infrastructure Network Monitoring pushes the boundary of Autonomous Database, reaching extraordinary scale with Exadata hardware.
“We had two main requirements for our Network Monitoring service: high performance at massive scale, and reduced administration time and costs. We needed to manage the performance of our network across 38 full-scale cloud regions worldwide. Only Autonomous Database on Exadata hardware could meet our needs.”
Business challenges
Network is one of the key pillars of Oracle Cloud Infrastructure (OCI). It touches literally everything. If there are any problems in the network, business-critical services would be impacted. Keeping tabs on what’s going on is important to keeping OCI running smoothly, and that’s the job of OCI Network Monitoring.
OCI Network Monitoring watches all aspects of network traffic throughout OCI, which includes internet Ingress/egress, traffic across regions, and within the data centers. Using this data collected, operators can troubleshoot and identify when there is degradation of performance in the network due to latency, congestion, and packet drop issues. OCI Network Monitoring assists operators to troubleshoot DDOS attacks and determine the source. OCI Network Monitoring facilitates the understanding of traffic patterns and growth rates enabling OCI network planners to perform capacity planning and service provider selection to maintain the highest quality of service for our OCI customers.
All these capabilities require not only capturing the data at scale but being able to retrieve this vast amount of data quickly to give the operator experience of real-time analysis. OCI Network Monitoring needed a high-performance database, capable of ingesting a massive amount of data, horizontally scalable, and self-managed.
Why Oracle Chose Autonomous Database
The OCI Network Monitoring leaders had two main requirements. First, they were looking for a managed service. As networking experts, they wanted to focus on the network and its data, not managing databases.
Second, they needed an elastic system that could scale to meet their current and future needs. OCI Network Monitoring staffers conducted extensive profiling based on mixed workloads and transaction processing time using various databases such as Apache Druid, Elastic Search, and finally selected Oracle Autonomous Database on Exadata. It was really the only option that met the requirements for scale and reduction in management complexity.
Results
Today OCI Network Monitoring is deployed on a production system of Autonomous Database on dedicated infrastructure, provisioned with 400 OCPUs and 598TB of storage to handle up to 120 million records per minute. For analysis and correlating the data in real time, Oracle Autonomous Database retrieves millions of transactional records per minute. This enables OCI Network Monitoring to feed this data to future northbound applications to manage traffic engineering, predictive analysis, and AI/ML analysis.
OCI Network Monitoring is scaled across all OCI regions and data centers around the world and presents a unified dataset to the network operations team. This reduced the complexity of operational workflows and improved mean time to detection and resolution of production issues.
The OCI Network Monitoring system is built entirely on OCI, using several OCI managed services. In addition to Autonomous Database, it uses OCI's Compute, Storage, and OCI Load Balancer.
Given the rapid growth of OCI cloud regions and the requirements of a self-managed database that can scale and sustain performance at level of 120 million records per minute and beyond, the obvious choice was Oracle Autonomous Database on Exadata hardware.