Glossary of Terms
Overview
This glossary defines key terms used throughout Provider Intelligence documentation. For detailed explanations, see the linked articles at the end of each description.
A
APNIC (Asia-Pacific Network Information Centre)
A database of Autonomous System Numbers (ASNs) used by Provider Intelligence to identify commercial internet service providers. One of the third-party data sources. See Data Sources and Coverage.
Application
A specific software service or business tool (such as Webex, Microsoft 365, or Salesforce) that serves as the target endpoint for provider traffic evaluation. Applications are one of two types of destination used in Provider Intelligence, alongside cloud regions. See Choose Destinations.
ASN (Autonomous System Number)
A unique identifier assigned to a provider (for example, Verizon AS 701). See Results Overview.
Average
The arithmetic mean of all measurements for a metric over your selected time frame. One of the summary statistics options. See Summary Statistics.
C
Cloud Agent
A ThousandEyes virtual machine deployed in a data center that generates synthetic tests. See Data Sources and Coverage.
Cloud Region
A specific geographic location where a major cloud provider (such as AWS, Azure, or GCP) operates data centers and hosted services. Cloud regions are one of two types of destination used in Provider Intelligence, alongside applications. See Choose Destinations.
Combined Score
See Performance & Stability Score.
Coverage
The extent of available test and third-party data in Provider Intelligence. See Data Sources and Coverage.
D
Destination
The target endpoint for provider evaluation, either an application (for example, Webex) or a cloud region (for example, AWS US-East-1). See Choose Destinations.
E
Evaluation Metrics
The metrics that are used to calculate the overall score of a provider. They include latency, loss, jitter, and, when applications are selected, Time to First Byte. Each metric’s score carries a different weight depending on your priorities. Evaluation metrics are always visible by default on the results table. See Prioritize Metrics.
H
Harmonic Mean
A method for combining performance scores and stability scores that penalizes score imbalance between the two more than a simple average. See Performance & Stability Scoring.
HTTP Server Timings
A breakdown of the components of an HTTP request/response cycle, beyond Time to First Byte (TTFB), including:
DNS Time: Domain name resolution.
Connect Time: TCP connection establishment.
SSL Time: SSL/TLS handshake.
Send Time: Request transmission.
Wait Time: Server processing time.
Receive Time: Response download.
Redirect Time: HTTP redirect overhead.
Total Time: Complete request/response cycle.
These metrics are available only when you have selected an application as a destination, and must be selected to make them visible in results. See Prioritize Metrics.
J
Jitter
The variance in round-trip time (latency) over time, measured in milliseconds. One of the evaluation metrics that is used to calculate the overall score. Lower is better. See Prioritize Metrics.
L
Latency (RTT)
Round-trip time — the time for a packet to travel from source to destination and back, measured in milliseconds. One of the evaluation metrics that is used to calculate the overall score. Lower is better. See Prioritize Metrics.
Location
The geographic origin (city) where ThousandEyes agents are deployed, serving as a proxy for end-user location. See Select User Locations.
Loss
The percentage of packets that fail to reach the destination. One of the evaluation metrics that is used to calculate the overall score. Lower is better. See Prioritize Metrics.
M
Metric
A specific aspect of network or application performance being measured. Provider Intelligence tracks:
Other metrics: outages, Total Unique AS Paths, HTTP server timings, page load metrics.
See Prioritize Metrics.
O
Outages
Provider disruptions detected by Internet Insights, counted over your selected time frame. Visible in results by default. Fewer is better. See Prioritize Metrics.
Overall Score
A weighted average of evaluation metrics scores that determines provider comparison results. See Overall Scoring.
P
Packet Loss
See Loss.
Page Load Metrics
Measurements of complete web page loading performance, including:
Page Load Time: Time from start to fully rendered page.
DOM Load Time: Time until HTML structure is loaded.
Throughput: Effective data transfer rate (page size / load time).
These metrics are available only when you have selected an application as a destination, and must be selected to make them visible in results. See Prioritize Metrics.
PeeringDB
A database of Internet exchange points and peering relationships used to validate provider data. One of the third-party data sources. See Data Sources and Coverage.
Percentile
A statistical measure indicating the value below which a given percentage of observations fall. For example, “P95 latency = 50 ms” means 95% of latency measurements were ≤ 50 ms. Each percentile focuses on a different segment of the data distribution, helping you understand performance across different scenarios. Provider Intelligence offers the following percentiles:
P95 (95th percentile) - Worst-case performance.
P75 (75th percentile) - Between typical and worst.
P50 (Median) - Typical performance.
P25 (25th percentile) - Better than average.
P5 (5th percentile) - Best-case performance.
See Summary Statistics.
Performance & Stability Mode
A results calculation option that blends the individual performance score and stability score, recommended for most production environment decisions where predictable performance matters. See Specify Results Calculation.
Performance & Stability Score
A score that incorporates both performance and stability using a harmonic mean. See Performance & Stability (Combined).
Performance Mode
A results calculation option that scores providers based purely on performance scores without considering stability. Emphasizes raw averages over consistency. See Specify Results Calculation.
Performance Score
A normalized score (0–100) quantifying how well a provider performs on average per metric relative to others in the location over your chosen time frame. See Performance Scoring.
Prioritized Metrics
Allows you to assign different weights to the evaluation metrics (latency, loss, jitter, Time to First Byte (TTFB)) in the query builder, so the overall score reflects what matters most to you. See Prioritize Metrics.
Provider
In Provider Intelligence, a provider refers to an Internet Service Provider (ISP) or Autonomous System identified by name and AS Number. Providers are the entities being compared and scored in your analysis.
Q
Query Builder
A tool that allows users to create and customize provider comparisons by specifying parameters (time frame, results calculation method, locations, destinations, and evaluation metrics) to generate tailored provider performance results. See Build A Query.
R
Report
A saved snapshot of a provider comparison analysis available from the Provider Intelligence homepage; the report includes selected providers and locations, query parameters, and scores. See Generate Reports.
Results Calculation
The method used to determine whether overall scoring uses each metric’s performance score alone, or also incorporates stability scores (consistency over time) when computing that metric’s contribution to the overall score. You select which method to use within the query builder. See Specify Results Calculation.
S
Scores
Calculated to normalize provider comparisons. Scores are calculated for metric performance, metric stability, metric performance & stability (combined), and overall. See Understanding Scores.
Single-Destination View
A results view showing provider comparison data for one specific application or cloud region. Single destination views display raw metric values (milliseconds, percentages) rather than scores because the metrics are comparable within a single destination context. See Destination Views.
Stability Score
A measure of how consistent a provider’s metric performs over time, expressed as a score (0 = different measurement across all hours, 100 = same measurement across all hours). See Stability Scoring.
Summary Report
See Report.
Summary Statistics
Statistical measures used to represent the distribution of performance data. Options include:
ThousandEyes Recommended: Uses different percentiles per metric based on best practices.
Percentiles: P95, P75, P50, P25, P5.
Average: Arithmetic mean.
Your selection affects which aspect of performance is emphasized (best-case, typical, worst-case). See Summary Statistics.
Summary View
A results view that aggregates performance across only your selected destinations (applications and/or cloud regions) into relative metric scores (0-100). Summary View is the default results view for all queries except those where no destination is selected.
Results always include at least one of your selected destinations, though not always all, since tests from the user location to the specific destinations might not always be available through every provider’s network. Therefore, results might not always provide a perfect match, but instead provide the closest match. It tries to answer the question, “Which provider best meets my specific needs here?” See Destination Views.
T
Third-Party Data Sources
ThousandEyes uses two public third-party data sources to help verify and fill data gaps to remove selection bias and improve data confidence. The sources are APNIC and PeeringDB. See Data Sources and Coverage.
ThousandEyes Recommended
A summary statistic mode that uses domain expertise to select the most appropriate percentile per metric. This approach recognizes that the same percentile isn't optimal for all metrics. Recommended default for most use cases.
ThousandEyes Recommended Percentiles by Metric
Latency
75th
Balances typical and peak performance
Loss
5th
Packet loss should be near-zero
Jitter
Average
Smooths out momentary spikes
TTFB
75th
Balances typical and peak application response
See Summary Statistics.
Time Frame
The historical period analyzed in your provider comparison query. Options are last 30 days, last 3 months, or last 6 months. Time frames are rolling: "last 3 months" always means the 90 days immediately preceding today, not calendar quarters. See Set the Time Frame.
Time Series
A chronological view of metric values showing hourly performance data over your selected time period. Time series graphs display each provider as a separate line, with time on the x-axis and metric values on the y-axis. Essential for identifying patterns, outages, trends, and anomalies that aggregate statistics might obscure. See Analyze Time Series.
Time to First Byte (TTFB)
The time from sending an HTTP request to receiving the first byte of the response, measured in milliseconds. One of the evaluation metrics that is used to calculate the overall score; only applicable when you select applications as destinations. Visible in results by default. Lower is better. See Prioritize Metrics.
Total Unique AS Paths
The number of Autonomous Systems (provider networks) that packets traverse from source to destination. Visible by default. See Prioritize Metrics.
Trend
The directionality of raw metric measurements over time, classified as up, down, or flat. Since most metrics perform better at lower values, up means worsening (throughput is the exception where up means improving). Trends are only visible in single-destination views. See Trend Analysis.
U
Universal Application Summary
A results view that aggregates performance data across all applications with test coverage from each provider in your selected locations. Available when you select one or more applications, or as the only view when you select no destinations.
Ensures fair comparison when test coverage varies across providers.
Provides results even when specific application coverage is limited.
Shows each provider's performance across a broader application set than just your selections.
Avoids bias from uneven test coverage.
See Destination Views.
User Location
See Location.
W
Weighted Average
The method used to combine the scores of evaluation metrics into a single overall score, where different metrics receive different weights based on your specified priorities.
4 Metrics, including TTFB
1st (highest)
40%
2nd
30%
3rd
20%
4th
10%
3 Metrics, excluding TTFB
1st (highest)
50%
2nd
33.3%
3rd
16.7%
See Overall Scoring.
Last updated