"Graph Abstraction enables millisecond level queries across roughly 650 TB of graph data, allowing engineers to analyze complex relationships quickly and reliably, which is crucial for operational workloads."
"To meet the demands for fast responses at high throughput, Netflix's system restricts traversal depth and often requires a defined starting node, trading some query flexibility for consistent low latency."
"The platform supports several internal use cases, including a real-time distributed graph for service interactions, a social graph for Netflix Gaming, and a service topology graph for incident analysis."
"Graph Abstraction is implemented as a layer on top of existing data infrastructure, with the latest graph state stored in a Key Value abstraction and historical changes recorded through a TimeSeries abstraction."
Graph Abstraction is a high throughput system created by Netflix engineers to handle large-scale graph data in real time. It powers internal services like social graphs for Netflix Gaming and service topology graphs for operational monitoring. The system allows millisecond-level queries across approximately 650 TB of data by separating edge connections from properties and globally replicating data. While it sacrifices some query flexibility for consistent low latency, it supports real-time distributed graphs, historical state maintenance, and integrates with existing data infrastructure to enhance performance.
Read at InfoQ
Unable to calculate read time
Collection
[
|
...
]