DataTerminology.comNEW
Time delay between initiating a request and receiving a response, influenced by network, compute, and storage factors.
DESCRIPTION
Latency measures responsiveness of systems and networks, impacting user experience and real-time analytics. It includes propagation, processing, queuing, and transfer delays. Engineers optimize latency via caching, parallelism, batching, and efficient protocols. Tail latency management and SLAs are critical for reliability. Observability tools track latency distributions to detect regressions and bottlenecks.
ALTERNATIVE DEFINITIONS
Cloudflare
Latency is the time it takes for data to travel from one point on a network to another.
Focuses on network travel time as the core metric.
AWS
Latency is the time it takes for a request to travel from the sender to the receiver and for the receiver to process that request.
Adds processing time to network travel definition.
Microsoft Learn
Latency is the amount of time it takes for a request to travel through the system and for the response to be returned.
System-wide perspective on request/response path.
RELATED TERMS
Response Time Jitter Tail Latency SLA Service Level Objective Network Delay Real-time Streaming Performance Tuning Caching
COMPETING TERMS
RELATIONS
- Narrower: Low Latency Tail Latency
KNOWLEDGE AREAS
Real-time & Streaming Data Observability & Telemetry Infrastructure Software Engineering Data Storage & Operations
TARGET AUDIENCE
DevOps Engineer Software Developer Data Engineer Cloud Architect MLOps Engineer
EXTERNAL SOURCES
- Wikipedia - Latency (engineering)
- Cloudflare - Latency overview and factors
- AWS - Latency definition and mitigation
- Google SRE - Tail latency and overload handling
TOP QUESTIONS
Why does Latency matter in data systems?
High Latency degrades user experience, breaks real-time analytics, and increases operational risk, especially in tail latency scenarios.
Category: Relevancy
What contributes to Latency in distributed systems?
Network propagation, serialization, queuing, storage IO, CPU contention, and synchronization costs contribute to Latency.
Category: Definition
How do you measure Latency effectively?
Measure percentiles (p50, p95, p99), instrument critical paths with traces, and correlate with throughput and error rates.
Category: HowTo
What are common mistakes addressing Latency?
Focusing only on averages, ignoring tail latency, and overusing synchronous calls. Missing caching and batching opportunities is frequent.
Category: Mistakes
What is a typical strategy to reduce Latency?
Use caching, parallel requests, connection pooling, efficient serialization, and deploy services closer to users (CDNs, edge).
Category: Requirements
What risks exist when optimizing Latency?
Aggressive caching can cause staleness; parallelism increases resource usage; premature optimization complicates systems.
Category: Risks
How does Latency relate to Throughput?
Throughput and Latency interact; optimizing one can impact the other. Balanced capacity planning addresses both.
Category: Definition
What is tail Latency and why is it important?
Tail Latency reflects worst-case performance; user experience is governed by high percentiles, not averages.
Category: Topic
How to set Latency SLAs?
Define percentile-based targets per critical endpoint, align with business impact, and monitor continuously with alerts.
Category: Management