API Latency Benchmark

Learn about Reltio’s baseline API performance and the key factors that affect it.

Overview

The Reltio platform powers complex implementations. As you plan your implementation, it’s important to predict the impact of your implementation on performance. This article helps you predict performance by providing:
  • Baseline latency for four API types.
  • Factors that affect latency for each API type.

Baseline API Latency and Parameters

We calculated the latency API types with the following assumptions:
  • No Lifecycle Actions (LCA) are used in the tenant
  • Simple address cleanser is used
  • No ordering and scoring is used in GET requests
  • No queries to the data history storage as part of the GET calls
  • Transitive matching isn’t enabled (parameter deep = 1)
  • URI only flag is set to true for the POST API
The following table summarizes the baseline (average) latency you can expect:
Table 1. Latency Details
Latency Type API Call Baseline Latency, ms
Save/Update Operation Load data into tenant 130
Find Operation Get entities using a filter 47
Find Operation Find entity by ID 51
Find Operation Find all matches for an entity 97
Find Operation Discover tree 260
Find Operation Get Hops 85
Find Operation Find Connections 227

We calculated the latency based on the following baseline parameters. Changes in these parameters will change the API latency.

Table 2. Baseline values of Parameters
Parameter Baseline value (no more than)
Batch size for POST entities 10 entities per batch
Entity size 10 KB
Number of simple attributes per entity 100
Number of nested attributes per entity 10
Number of reference attributes per entity 0
Number of RDM lookups per entity 5
Overall API response size 50 KB
Filter conditions for GET entities 10
Match rules for an entity type 5