Usually, when you go on a trip, one of your first thoughts might be, “I hope I have a good time.” From the moment you step foot outside your door until you reach your final destination, you’re really hoping that the traffic on the roads does not create a delay. Maybe, you’re looking at the map application to find the best route, scanning for any option that lacks the dreaded red line of waiting. Especially on holidays, you know that a lot of cars are going to be on the roads going to a lot of different places, just like you.
In IT, modern networks are very similar to crowded roads. They generate volumes of network traffic as users, applications, devices, and systems continuously communicate with one another. To increase complexity, cloud apps, cloud computing, remote users, and Internet of Things (IoT) devices now expand the traditional network perimeter, creating new cybersecurity risks.
Network traffic analysis (NTA) collects and analyzes network telemetry, turning it into usable information. Network traffic analysis provides context around network activity so that teams can understand what is communicating, how it behaves, and when to investigate abnormal behavior.
What Is Network Traffic Analysis?
Network traffic analysis (NTA) is the process of collecting and analyzing data about network traffic to understand communications between devices and systems. It gives network administrators and security professionals visibility into information that helps them with network monitoring, performance troubleshooting, and potential malicious behavior, including:
- Source and destination addresses
- Protocols
- Connection patterns
- Data volumes
While some tools record whether an event occurred, NTA examines patterns across communications to establish baselines for normal activity so that teams can identify anomalies like network congestion or potential security incidents. Often, anomaly detection surfaces activity that traditional signature-based controls may not recognise, improving threat detection for previously unknown or changing attack techniques.
What Is The Difference Between Network Traffic Analysis vs. Network Detection and Response?
Network traffic analysis and Network Detection and Response (NDR) are closely related, but they serve different functions within a security program. Some key differences include:
- Purpose: While NDR focuses on detecting and responding to potential security threats, NTA focuses on understanding patterns and behaviors within network traffic.
- Scope: While NDR is primarily a cybersecurity capability, NTA can support both security and operational use cases, including performance monitoring and troubleshooting.
- Analysis: While NDR uses network data to identify activity associated with potential threats, NTA analyzes communications to establish normal behavior, identify anomalies, and provide context about network activity.
- Response: While NDR typically incorporates investigation and response capabilities, NTA primarily provides the visibility and analysis that teams can use to determine whether further investigation or action is necessary.
How Does Network Traffic Analysis Work?
Network traffic analysis turns raw network traffic data into information that security and networking teams can use by collecting data, establishing baselines, identifying anomalies, and investigating potentially risky activity.
Collecting Network Traffic Data
Before doing anything else, teams need to collect network telemetry from relevant sources across the environment. Data can come from any of the following:
- Flow data: Communications between systems, including source and destination IP addresses, ports, protocols, connection duration, and traffic volume.
- Packet capture: Information contained in the packet headers and payload as individual packets move across the network for context into where communications originated and their destinations while deep packet inspection also looks into packet content.
- Network and security logs: Data from sources such as firewall logs, DNS, and a web proxy can provide additional context about connections and network behavior.
Establish Normal Network Behavior
After collecting data, teams use historic and real-time data to define baselines for normal network activity that might include:
- Traffic volumes
- Protocols
- Connection frequencies
- Destinations
- Communication patterns between systems
Identify Anomalies and Suspicious Activity
Anomaly detection tools compare new network activity against baselines to identify potential performance or security issues by looking at:
- Unexpected connection between systems
- Unusual data transfers
- Traffic spikes
- Communications with external destinations
Some network traffic analysis tools also use machine learning to analyze large volumes of traffic to strengthen threat detection by surfacing malicious network activity even when it fails to match a known attack signature.
Investigate the Activity
Once teams have an alert, they need context to investigate it and determine whether it represents a potential threat, a legitimate change, a configuration problem, or a performance or network configuration problem. Often, teams correlate network findings with the rest of their data, including data from:
- Endpoints
- Identity and access management tools
- Asset inventories
- Vulnerability management tools
- Threat intelligence dataTota
- Application logs
- Device information
- Configuration changes
- Infrastructure metrics
For example, when looking at increased network traffic, they need to know whether the change arises from a newly deployed application or a malicious actor purposefully overloading a resource as part of a Distributed Denial of Service (DDoS) attack.
By enriching data, teams can rapidly move from identifying a change to understanding why the changes occurred. Then the organization knows whether the security or network team needs to remediate the issue.
What Can Network Traffic Analysis Detect?
Network traffic analysis defines normal for an organization’s network and identifies deviations that can help teams detect security, performance, and operational issues. Depending on the network traffic and contextual data collected, teams can detect:
- Potential security threats: Unexpected connections, unusual data transfers, command-and-control communications, lateral movement, or other malicious network activity can indicate compromised systems or an active threat actor.
- Abnormal network behavior: Changes in connection frequency, protocols, destinations, or traffic volumes can surface activity that warrants further investigation, even when it does not match a known threat signature.
- Performance issues: Traffic spikes, bandwidth utilization, latency, and network congestion can help teams identify bottlenecks and troubleshoot degraded application or system performance.
- Configuration problems: Unexpected communication paths or changes in traffic patterns can reveal routing errors, misconfigured devices, or other network configuration issues.
- Unusual device activity: Monitoring communications from endpoints, Internet of Things devices, and other connected systems can help teams identify unexpected connections or changes in device behavior.
How to Analyze Network Traffic: Best Practices
While effective network traffic analysis includes collecting data, organizations need to identify the traffic that matters, provide context for it, and make it accessible to the security and network teams who work with it.
Prioritize Relevant Network Traffic
Collecting all the data from every relevant source can increase storage and processing requirements and costs without improving visibility. Teams should start by identifying the systems, network segments, and communications that are most important to their security and operational needs.
When prioritizing network traffic, organizations should consider:
- Support for collecting flow records such as NetFlow and IPFIX from routers, firewalls, and other network infrastructure.
- Flexible data ingestion that can bring together information from network devices, security tools, servers, applications, and other relevant sources.
- Routing and filtering capabilities that help organize incoming data according to its value and intended use.
Enrich Network Data with Context
Raw network telemetry shows that systems communicated. To determine what that communication means, security and networking teams need information related to assets, threats, identity, locations, or other information that makes investigation easier.
When enriching network data, organizations should consider:
- Data normalization that creates consistent fields across different log and network data sources, making information easier to search and compare.
- Lookup and enrichment capabilities that can add contextual information to incoming data, such as mapping an IP address to a hostname or other relevant attributes.
- Asset context that can connect activity to information about the affected system, including vulnerability information where available.
Correlate Data Across the Environment
To fully understand an alert’s meaning, network and security teams need to connect the network activity data with information from other systems. This process helps them determine whether the anomaly is a security incident, performance problem, configuration change, or expected behavior.
When correlating network data, organizations should consider:
- Centralized search capabilities that allow analysts to query normalized data across relevant sources.
- Event correlation that can connect related activity rather than requiring analysts to investigate isolated events individually.
- Dashboards and visualizations that make it easier to identify trends, anomalies, and relationships within collected data.
Maintain Historic Data for Investigation
While organizations need real-time visibility into their networks, historic data provides the context necessary for comparing current activity against previous behavior. Retaining relevant data can support forensic investigation, troubleshooting, baselining, and longer-term analysis.
When maintaining network data, organizations should consider:
- Tiered storage that keeps frequently searched network traffic data in a high-performance hot tier, moves less frequently accessed data to a lower-cost but still searchable warm tier, and retains long-term historic or compliance data in lower-cost archive storage.
- Fast historical search that enables analysts to move backward from an anomaly or alert and reconstruct related activity.
- Anomaly detection that uses historic information to establish expected behavior and identify meaningful deviations over time with machine learning models to evaluate incoming activity.
Graylog: Turning Network Traffic into Actionable Intelligence
Graylog enables security and IT teams to bring network traffic data together with logs, security events, asset information, and other telemetry to provide the context needed for effective network traffic analysis. Centralized search, normalization, enrichment, correlation, visualization, and anomaly detection help teams move beyond identifying unusual activity to investigating what happened and why.
With flexible data ingestion and tiered storage, teams can maintain access to the historic and real-time data needed for threat detection, troubleshooting, and forensic investigation without treating every piece of network data the same. Graylog helps organizations turn distributed network signals into contextualized information that analysts can use to identify security threats, investigate performance problems, and make faster, better-informed decisions.