⏱️ Network Latency Tools
km
Distance between source and destination
Enter a distance between 1 and 50,000 km.
Signal speed varies by physical medium
One-Way Latency
—
⚠️ Disclaimer: These are theoretical propagation delay calculations. Real-world latency is typically 1.5 to 3x higher due to routing overhead, router hops, and queuing delays.
Was this calculator helpful?
km
Approximate distance to server
Enter a distance between 1 and 50,000 km.
hops
Typical: local 5–8, cross-country 10–15, international 15–25
Enter hops between 1 and 64.
Estimated RTT
—
ms
Enter the ping value you measured (e.g. from ping command or speed test)
Enter a valid ping value between 1 and 5000 ms.
Ping Rating
—
Sources & Methodology
Latency calculations based on ITU-T G.114 one-way delay standard, RFC 2681 round-trip delay metric, and published fiber optic signal propagation constants. Application benchmarks sourced from Cisco QoS guidelines and gaming industry standards.
ITU-T G.114 — One-Way Transmission Time (International Telecommunication Union)
International standard defining acceptable one-way latency thresholds for voice communication. Establishes 150ms as the maximum one-way delay for acceptable speech quality, the benchmark used by all VoIP and video conferencing systems.
RFC 2681 — A Round-Trip Delay Metric for IPPM (IETF)
IETF standard defining the methodology for measuring round-trip time, including the relationship between RTT, one-way latency, and the ICMP ping measurement technique used by network diagnostic tools.
AWS — What is RTT in Networking?
Amazon Web Services documentation on RTT measurement, factors affecting latency in cloud environments, and CDN-based latency reduction strategies used in production AWS deployments.
Methodology:
Propagation delay (one-way) = Distance / Signal speed in medium RTT estimate = (2 x propagation delay) + (hops x 0.5ms processing) Real-world RTT = theoretical RTT x 1.5 to 3.0 overhead factor Fiber optic signal speed: 200,000 km/s (67% of vacuum speed of light, refractive index ~1.47). Wireless/5G: 300,000 km/s (near speed of light in air). Per-hop processing delay: 0.1 to 1ms typical, 0.5ms used as conservative estimate. Application benchmarks based on ITU-T G.114, Cisco QoS design guides, and published gaming industry latency standards.
Propagation delay (one-way) = Distance / Signal speed in medium RTT estimate = (2 x propagation delay) + (hops x 0.5ms processing) Real-world RTT = theoretical RTT x 1.5 to 3.0 overhead factor Fiber optic signal speed: 200,000 km/s (67% of vacuum speed of light, refractive index ~1.47). Wireless/5G: 300,000 km/s (near speed of light in air). Per-hop processing delay: 0.1 to 1ms typical, 0.5ms used as conservative estimate. Application benchmarks based on ITU-T G.114, Cisco QoS design guides, and published gaming industry latency standards.
Last reviewed: April 2026
Network Latency, Ping & RTT: The Complete 2026 Guide
Network latency determines how responsive every online experience feels — from gaming and video calls to stock trading and cloud applications. Understanding what latency is, how to calculate it, what causes high ping, and what values are acceptable for different applications is essential for network engineers, gamers, IT admins, and anyone who works online. This guide covers every aspect of network latency with practical tables, formulas, and actionable advice to improve your connection.
Latency, Ping, and RTT: What Each Term Means
These three terms are often used interchangeably but have precise technical meanings. Latency is the one-way time for a packet to travel from source to destination. Ping (RTT) is the round-trip time for a packet to reach a destination and for the reply to return. RTT is approximately twice the one-way latency assuming symmetric routing. In practice, the path from A to B and from B to A is often different on the internet (asymmetric routing), meaning RTT divided by 2 is only an estimate of one-way latency.
One-way latency = Distance / Signal speed in medium
RTT (theoretical) = 2 x one-way propagation delay
RTT (real-world) = propagation + transmission + processing + queuing delays
Example: London to New York is ~5,570 km. Over fiber (200,000 km/s): one-way = 5570/200000 = 0.02785 sec = 27.85 ms. Theoretical RTT = 55.7 ms. Real-world typically 70 to 90 ms due to routing overhead.
Latency by Connection Type: Benchmark Reference
Different connection technologies have fundamentally different latency characteristics driven by their physical properties. This table shows typical real-world latency ranges for each major connection type.
| Connection Type | Typical RTT | Theoretical Minimum | Main Latency Factor |
|---|---|---|---|
| Fiber Optic (local) | 1 – 5 ms | ~0.1 ms | Propagation (short distance) |
| Fiber Optic (cross-country) | 20 – 60 ms | ~15 ms | Propagation (distance) |
| Cable Broadband | 10 – 40 ms | ~5 ms | DOCSIS protocol overhead |
| DSL (ADSL/VDSL) | 20 – 80 ms | ~10 ms | ATM overhead + copper distance |
| 5G Mobile | 10 – 30 ms | ~1 ms (5G SA) | Radio protocol + core routing |
| 4G LTE | 30 – 60 ms | ~10 ms | LTE air interface + core |
| Starlink LEO | 20 – 60 ms | ~3.7 ms | LEO orbit + ground routing |
| GEO Satellite | 480 – 650 ms | ~480 ms | 35,786 km orbit distance |
| Wi-Fi (5 GHz) | +2 – 5 ms added | ~1 ms | Wireless contention + retransmit |
| Wi-Fi (2.4 GHz) | +5 – 20 ms added | ~2 ms | Interference + congestion |
Latency Benchmarks by Application Type
Different applications have vastly different latency requirements. What is perfectly acceptable for email is completely unusable for high-frequency trading. Use this table to understand whether your current ping is suitable for your use case.
| Application | Excellent | Good | Acceptable | Poor |
|---|---|---|---|---|
| FPS / Fighting Games | < 20 ms | 20 – 50 ms | 50 – 100 ms | > 100 ms |
| Strategy / RPG Games | < 50 ms | 50 – 100 ms | 100 – 150 ms | > 150 ms |
| VoIP / Voice Calls | < 100 ms | 100 – 150 ms | 150 – 300 ms | > 300 ms |
| Video Conferencing | < 100 ms | 100 – 150 ms | 150 – 250 ms | > 250 ms |
| Video Streaming | < 100 ms | 100 – 200 ms | 200 – 500 ms | > 500 ms |
| Web Browsing | < 100 ms | 100 – 200 ms | 200 – 500 ms | > 500 ms |
| File Transfers | Any | Any | Any | Very high impacts TCP throughput |
| Algo Trading | < 1 ms | 1 – 10 ms | 10 – 50 ms | > 50 ms |
What Causes High Ping? All 8 Root Causes
High latency has multiple causes, each requiring a different fix. Understanding the root cause before trying to fix ping is critical — the solution depends entirely on where the delay originates.
- Physical distance: The single biggest factor. Every 1,000 km adds approximately 5 ms of one-way propagation delay over fiber. You cannot overcome physics — the only solution is using servers closer to you or a CDN.
- Too many router hops: Each router adds 0.1 to 2 ms of processing delay. Intercontinental connections can have 15 to 25 hops. Poor ISP peering agreements cause inefficient routing with extra hops.
- Network congestion: Overloaded routers and ISP backbones queue packets, adding variable delay. This causes both high average ping and high jitter. Worst during peak hours (evenings, weekends).
- Wi-Fi interference: Wireless contention, interference from neighboring networks (especially on 2.4 GHz), and retransmissions add 5 to 30 ms of extra latency. Ethernet always gives lower, more stable ping.
- Server load: Overloaded game servers and web servers take longer to process requests. Even if your network latency is 10 ms, a server under heavy load can add 50 to 200 ms of processing delay.
- ISP routing inefficiency: Some ISPs route traffic through distant peering points instead of direct paths. A ping to a server in the same city can sometimes travel internationally if ISP peering is poor.
- Bufferbloat: Consumer routers with large packet buffers introduce latency under load. When your connection is busy, the buffer fills up and incoming packets wait in the queue, dramatically increasing latency. QoS/SQM settings fix this.
- VPN overhead: VPNs encrypt and re-route traffic through an additional server, typically adding 5 to 50 ms depending on the VPN server location and encryption method.
Jitter: The Latency Problem Nobody Talks About
Jitter is the variation in packet arrival time. If packets arrive at 20ms, 45ms, 35ms, 80ms intervals, the average is 45ms but the jitter is 60ms (difference between minimum and maximum). For applications that process packets in real time — VoIP, gaming, video calls — jitter causes audio distortion, rubber-banding, and screen freezing. A stable 80ms ping is often better than a variable ping that swings between 20ms and 150ms.
💡 Jitter thresholds: For VoIP (per ITU-T G.114), jitter must be below 30ms. For competitive gaming, below 15ms is ideal. Jitter above 50ms makes voice calls difficult and gaming nearly unplayable. Jitter can be measured by running ping for 30 to 60 seconds and comparing minimum and maximum values.
Bandwidth-Delay Product: Why High Latency Kills TCP Throughput
The bandwidth-delay product (BDP) is the amount of data that can be in transit simultaneously across a network path. BDP = Bandwidth x RTT. For a 1 Gbps link with 100ms RTT, BDP = 12.5 MB. TCP must fill this entire pipe to achieve maximum throughput. If the TCP window size is smaller than the BDP, the pipe is never full and throughput is reduced. This is why satellite internet with 500ms RTT achieves far less than its nominal bandwidth on large file transfers despite having adequate bandwidth capacity.
How CDNs and Edge Computing Reduce Latency
Content Delivery Networks (CDNs) solve the distance problem by caching content on servers close to end users. Instead of a user in Dubai downloading from a server in New York (adding ~100ms of propagation delay), they download from a CDN edge node in the same city (adding just 5ms). Major CDNs like Cloudflare, AWS CloudFront, and Akamai have hundreds of edge locations worldwide. Edge computing takes this further by running compute workloads at the CDN edge, reducing latency for dynamic requests that cannot be cached.
Frequently Asked Questions
Network latency is the time delay for a data packet to travel from source to destination, measured in milliseconds. It has four components: propagation delay (limited by speed of light in the medium), transmission delay (time to put bits on the wire), processing delay (router and switch processing time), and queuing delay (waiting time in network buffers). Propagation delay dominates over long distances and cannot be reduced without moving the endpoints closer together.
Latency technically means one-way delay. Ping (RTT) measures the round-trip time for a packet to reach a destination and the reply to return. RTT is approximately twice the one-way latency assuming symmetric routing. In practice, most people use latency and ping interchangeably. True one-way latency is difficult to measure accurately without synchronized clocks at both endpoints. The ping command measures RTT, not one-way latency.
Under 20ms is excellent for competitive gaming. 20 to 50ms is good for casual and competitive play. 50 to 100ms is acceptable for casual games but noticeable in fast-paced games. Above 100ms causes visible lag in FPS and fighting games. Above 150ms makes competitive gaming impractical. Jitter matters as much as average ping: a stable 60ms is better than a variable ping swinging between 20ms and 150ms. Server tick rate also affects perceived smoothness independently of ping.
Traditional geostationary (GEO) satellites orbit 35,786 km above Earth. A data packet travels 35,786 km up to the satellite and 35,786 km back down to the ground station, adding ~240ms per one-way trip and ~480ms round-trip from physics alone. Starlink and other LEO satellites orbit at only 550 km, reducing the round-trip propagation to ~3.7ms, similar to terrestrial internet. Total Starlink RTT is 20 to 60ms once routing overhead is included.
Propagation delay equals distance divided by signal speed in the transmission medium. Fiber optic carries light at 200,000 km per second (67 percent of vacuum speed of light). For 1,000 km of fiber: 1000 divided by 200,000 equals 0.005 seconds equals 5 ms one-way. RTT is approximately double: 10 ms. Use the Propagation Delay tab above to calculate any distance and medium combination instantly.
The main causes are: distance to server (physics, partially fixed by choosing closer servers), too many routing hops (ISP peering issue), network congestion (peak hours), Wi-Fi interference (fix with ethernet), bufferbloat (fix with QoS/SQM on your router), and VPN overhead (turn off VPN for gaming). Connect via ethernet, choose geographically close servers, disable background downloads, and enable QoS on your router to prioritize gaming traffic.
Jitter is the variation in packet arrival time. If consecutive pings are 20ms, 50ms, 35ms, 90ms instead of a steady 45ms, that inconsistency is jitter. For real-time applications like VoIP, gaming, and video calls, jitter causes choppy audio, rubber-banding, and screen freezing even when average latency is acceptable. Per ITU-T G.114, jitter must be below 30ms for VoIP. For gaming, below 15ms is recommended. Measure jitter with a 60-second ping test and compare minimum to maximum values.
ITU-T G.114 recommends maximum one-way latency of 150ms for acceptable voice quality and below 100ms for good quality. RTT should be below 300ms. Jitter must be below 30ms. Packet loss should be below 1 percent. Above 150ms one-way, callers begin to talk over each other. Enterprise UC systems target under 80ms one-way. Consumer VoIP (Zoom, Teams, WhatsApp calls) works acceptably up to 150ms one-way with modern jitter buffers.
Bandwidth-delay product (BDP) equals bandwidth multiplied by RTT. It represents how much data is in flight across the network at any moment. For a 1 Gbps link with 100ms RTT: BDP = 1,000,000,000 x 0.1 = 100,000,000 bits = 12.5 MB. TCP must keep this much data in flight continuously to fully utilize the link. If the TCP receive window is smaller than BDP, throughput is capped. High-latency high-bandwidth links (long fat networks) require TCP window scaling to achieve full speed.
Real-world RTT is always higher than theoretical propagation delay because it includes router processing delays (0.1 to 1ms per hop), queuing delays in congested equipment, routing inefficiency (data rarely travels in straight lines), Wi-Fi contention, and ICMP de-prioritization (routers process ping with lower priority than normal traffic, artificially inflating measured RTT). Real-world latency is typically 1.5 to 3 times the theoretical minimum propagation delay over long distances.
ICMP de-prioritization is a network practice where routers process ICMP packets (used by ping) with lower priority than regular TCP and UDP traffic. This means a router under load will delay responding to ICMP echo requests even if regular application traffic is passing through without delay. The result is that your measured ping may be higher than the actual latency experienced by your applications. For this reason, network professionals use application-layer latency measurements alongside ping for accurate network assessment.
High-frequency trading (HFT) requires sub-millisecond latency. HFT firms co-locate their servers physically inside exchange data centers (NYSE, NASDAQ, CME) to achieve microsecond-level latency. Even 1ms of extra latency represents a significant competitive disadvantage worth millions of dollars in trading opportunities. For retail algo trading, under 10ms is excellent. Retail online trading tolerates 50 to 200ms without significant impact on order execution quality.
CDNs (Content Delivery Networks) reduce latency by caching content at edge servers geographically close to users. Instead of data traveling from a server in the US to a user in India (adding ~150ms propagation delay), the content is served from a CDN edge node in Mumbai (adding just ~5ms). Major CDNs including Cloudflare, AWS CloudFront, and Akamai operate hundreds of edge locations. For dynamic content, edge computing runs application logic at these edge nodes, reducing latency for requests that cannot be served from cache.
100ms RTT is acceptable for video conferencing. Zoom, Microsoft Teams, and Google Meet work well below 150ms RTT. Below 100ms RTT is good. Above 150ms RTT causes noticeable conversation delays where participants start talking over each other. Above 300ms RTT makes natural conversation very difficult. In addition to latency, jitter below 30ms and packet loss below 1 percent are required for clear audio and stable video in conferencing applications.
Related Networking Calculators
Popular Calculators
Stripe Fee Calculator
Finance
Pain & Suffering
Legal
TDEE Calculator
Health
Roofing Calculator
Construction
GPA Calculator
Education
Mortgage Refinance
Finance
CPM Calculator
Marketing
Markup Calculator
Finance
5g Lte Throughput Calculator
Networking
Nat Pat Session Calculator
Networking
Bgp As Path Calculator
Networking
Dns Lookup Calculator
Networking