Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

Recent Flow Analysis Widget

 

The Recent Flow Analysis dashboard widget enables you to view the results of all analyses of application flows that have been initiated by Network Director in your network. It also enables you simulate and analyze a flow between virtual machines (VMs) and between bare metal servers (BMSs) to determine the best placement for a new application in your data center.

Network Director uses Cloud Analytics Engine to perform flow analysis. The Compute Agent component of Cloud Analytics Engine creates a probe, or synthetic packet, that traces the path of the application flow through the network. When a device detects the probe, it collects various metrics that are sent to the Compute Agent, which then sends the metrics to the Data Learning Engine (DLE) component of the Cloud Analytics Engine. Network Director, in turn, obtains this information from DLE.

Flow analysis provides you with the number of hops, latency per hop, and end-to-end latency. For each hop, you can view information about the device—for example, CPU utilization, traffic statistics, and ingress and egress ports used. The information provided enables you to determine congestion points in the network that might be affecting application performance.

In addition to using the Recent Flow Analysis widget, you can perform flow analysis or view the results of flow analysis using the following widgets or tasks:

  • Virtual Machine & Bare Metal Servers dashboard widget—This widget enables you to initiate flow analysis on existing application flows and to view the results on a per VM or BMS basis. It also enables you to put a VM or BMS on a watchlist, which automatically starts flow analysis on all flows.

  • Device and Port Utilization dashboard widget—This widget provides a graphical view of the level of port utilization on a device. You can mouse over a representation of a port and click View active flows through this link to display the results of flow analysis on flows on that port.

  • View Connectivity Task in Build mode in the Datacenter view—You can view the flows on a BMS or VM and start analysis on selected flows by mousing over the BMS or VM and clicking Show active flows.

This topic describes:

Requirements for Flow Analysis

To perform flow analysis with Network Director, you must:

  • Ensure that the components of Cloud Analytics Engine are installed on your network devices and that the Compute Agent discovery file has been created and uploaded to the Data Learning Engine (DLE).

    See Understanding Cloud Analytics Engine and Network Director for more information.

  • Specify the DLE server IP address under Preferences > Monitoring> Data Learning Engine Settings.

  • Enable LLDP on the servers hosting the VMs, on the BMSs, and on the connecting switches. In addition, the switches must be discovered by using the SNMP option in Network Director.

Refer to the Cloud Analytics Engine documentation in the QFX Series switch documentation for more information about the Cloud Analytics Engine.

Network Director performs flow analysis between VMs and BMSs in the following topologies:

  • BMSs connected by a Layer 3 IP fabric

  • VMs connected by VMware NSX or OpenStack VXLAN tunnels overlaid on a Layer 3 IP fabric

Recent Flow Analysis Main View

The main view of the Recent Flow Analysis widget lists the 10 most recent flow analyses performed by Network Director, as determined by their start time. You can view all flow analyses by clicking the expand icon at the top of the widget or View all at the bottom of the widget.

Each flow analysis consists of one or more traces—a trace being a period during which flow analysis was active. Multiple traces can exist because a flow can become inactive and then active again during the duration of a flow analysis or because multiple analyses have been performed on the same flow over a period of time. By default, the summary and expanded views of the Recent Flow Analysis widget show the results of the most recent trace for a flow.

You can take the following actions:

  • Click View Details to view in-depth analysis of the trace results.

  • Click All Traces to view the results of all traces performed on a flow.

Table 1 describes the fields in the Recent Flow Analysis expanded view.

Table 1: Recent Flow Analysis Expanded View

Field

Description

Source IP

Source Port

Source IP address and source TCP/UDP port for the flow.

If the port is associated with a well-known service, the service name is also shown.

Destination IP

Destination Port

Destination IP address and destination TCP/UDP port for the flow.

If the port is associated with a well-known service, the service name is also shown.

Status

Status of the most recent trace:

  • Active (rotating circle)—Trace has started and is ongoing.

  • Completed (check mark)—Trace completed at the end time shown.

  • Failed (red triangle)—Trace was unable to complete.

  • Scheduled (clock)—Trace is scheduled to start and finish at the start and end times shown.

Start Time

End Time

Start and end time for the most recent trace.

Min Latency

Lowest end-to-end latency experienced by a probe during the trace, in milliseconds.

Note: Time drift on the network devices between NTP synchronizations can result in negative values.

Max Latency

Highest end-to-end latency experienced by a probe during the trace, in milliseconds.

Avg Latency

Average end-to-end latency experienced by the probes in the trace, in milliseconds.

Actions

Click View Details to open the Flow Analysis Details window, which provides in-depth path analysis for this trace.

Click All Traces to display all previously completed traces for this flow. For each trace, the information described in this table is provided.

Flow Analysis Details Window

The Flow Analysis Details window provides detailed information about a flow trace.

The Flow Analysis Details window is divided into three sections:

  • The flow path diagram—This diagram shows the path taken by a probe through the network. By default, the path shown is the path taken by the probe that experienced the highest per-hop latency in the trace. You can change this diagram to reflect the path taken by a different probe by selecting the probe from the top Latency Trend chart.

  • Latency Trend charts—These charts show the change in latency experienced by the probes during the trace. The bars in the top chart are grouped by completed probes, with each bar in a probe group representing the latency experienced by the probe at a hop. By clicking a probe group, you can change the flow path diagram and the Analysis Results section to reflect the results of that probe. For traces of long duration, the bar chart shows only a portion of the trace results.

    The bottom area chart graphs the highest latency experienced by each probe over the entire duration of the trace. You can use the provided controls to focus on a portion of the trace—the portion you choose is reflected in the top bar chart. By default, the focus is on the portion of the trace that had the highest latency. If the trace is ongoing, a rotating circle appears at the end of the plotted area and the chart is periodically refreshed to show new results.

    Both charts display a path change icon () when the path a probe takes through the network differs from the path taken by the previous probe.

  • Analysis Results—This section provides details about the overall trace results and about the selected probe:

    • The Latency table provides overall latency information for the trace: the highest and lowest latency experienced at a single hop and the average latency of all hops.

    • The Latency for Selected Path table shows the latency experienced by the selected probe at each hop.

You can perform the following actions in the Flow Analysis Details window.

General actions:

  • For bidirectional traces, you can select the direction for which you want results by clicking one of the arrows at the top of the window (these arrows do not appear for unidirectional traces).

  • To stop an active flow analysis, click Stop Flow Analysis at the bottom of the window. When you stop an active flow analysis, the results up to the time you stopped the flow analysis are retained and the previously active trace is marked as complete.

On the flow path diagram, you can:

  • Reposition the topology diagram by dragging it or reposition devices by dragging them.

  • Zoom in or out by clicking the plus or minus signs on the left.

  • Mouse over the link connecting two devices to get the connecting port names. The names are displayed in green if the link is up and in red if the link is down.

  • Mouse over a device to view details about the device, such as name, connection state, and IP address. The details shown depends on the device type.

    If a device in the flow path does not support Cloud Analytics Engine, it is shown in the diagram in light grey color and minimal details, such as IP address, are available.

  • Display the traffic statistics for switches by mousing over the device to display the device details and clicking the Show Traffic Data link. If you selected the Capture Bandwidth option when you started the flow analysis, the flow bandwidth is also displayed along with the traffic statistics.

  • Display the active flows associated with a VM, BMS, or virtualized host by mousing over the device and clicking Show Active Flows in the details box.

On the Latency Trend charts, you can:

  • Mouse over a bar group in the top bar chart. A pop-up box displays the latency figures for each hop taken by the probe.

  • Click a bar group in the top bar chart. The flow path diagram and the Analysis Results change to reflect the information for the probe.

  • Mouse over a path change icon in the top bar chart. Information about the old and new paths is displayed.

  • Change the span and position of the focus indicator on the bottom area chart:

    • To increase or decrease the time span of the focus—in other words, to zoom in or zoom out on a portion of the trace—click one of the handle controls () and move it in either direction.

    • To change the focus to another time period, click on the arrows at either end of the slider bar.

Simulate Flow Analysis Window

The Simulate Flow Analysis window enables you to create a simulated application flow between VMs or between BMs in a data center and to analyze the flow results. By performing analysis of simulated flows, you can determine the best placement for new applications in the data center.

You can simulate flows between:

  • Two existing BMSs in a non-overlay network.

  • Between two VMs in an overlay (VXLAN) network. The VMs do not need to exist to simulate a flow between them.

To begin simulated flow analysis, you must specify which type of flow you are simulating:

  • Select Non Overlay Network to simulate flows between two BMSs. Table 2 describes the settings for this option.

  • Select Overlay Network to simulate flows between two VMs. Table 3 describes the settings for this options.

Table 2: Non Overlay Network Settings

Field

Action

Select Datacenter

Select a data center from the list.

IP Address

Select the source and destination IP addresses for the simulated flows.

The IP addresses listed are existing IP addresses on the BMSs in the data center.

Port

Source and destination TCP/UDP ports.

Protocol

Select protocol—TCP or UDP.

Table 3: Overlay Network Settings

Field

Action

Select Datacenter

Select an existing data center from the list.

VNI ID

Select the VNI (VXLAN Network Identifier) for the overlay network.

VTEP IP Address

Select the source and destination VTEP (VXLAN Tunnel Endpoint) IP addresses.

MAC Address

If the source and destination VMs exist, enter their MAC addresses.

If the source and destination VMs do not exist and the tunnel is from hypervisor to hypervisor without a VXLAN gateway between the hypervisors, enter the MAC addresses for possible virtual Ethernet ports on the servers where these VM might reside. If you do not know the VM location or the VM location is not decided, enter made-up MAC addresses.

If the source and destination VMs do not exist and there is a VXLAN gateway on the tunnel between the hypervisors, enter MAC addresses from the MAC pools on the servers or hosts.

Port

Source and destination TCP/UDP ports.

Protocol

Select protocol—TCP or UDP.

After you have defined the basic settings for the type of flow you are simulating, click Simulate Flow & Analyze to begin the simulated flow. You can also click Advanced Settings to configure the advanced settings described in Table 4 before you start the flow analysis.

Table 4: Simulate Flow Analysis Advanced Settings

Field

Action

Schedule

Run Now—Select to start the flow analysis as soon as you click Simulate Flow & Analyze.

Schedule Later—Select to schedule the analysis to run at a later time and enter the date and time in the fields provided. When you click Simulate Flow & Analyze, the analysis is scheduled to run at the date and time you specified.

Scheduled flow analyses are shown only on the Recent Flow Analysis widget and cannot be managed as scheduled jobs under System > Manage Jobs.

Duration

Specify how long the flow analysis runs.

Frequency

Specify how often a probe is sent on the flow path during the duration in which flow analysis is active.

Timeout

Specify how long Network Director waits before timing out the flow analysis after a probe fails to respond.

Max. number of hops

Specify the maximum number of hops on which flow analysis is performed.

Select Direction

Click an arrow to select the direction of the flow you want to analyze: the right arrow for source to destination, the left arrow for destination to source, and the double-headed arrow for both directions.

Capture Flow Bandwidth

Select to have Network Director collect and report information about the flow bandwidth.

If you select this option, you can view the captured flow bandwidth by:

  • Displaying the traffic statistics by mousing over a device in the Flow Analysis Details window.

  • Viewing active flows through a port in the Device & Port Utilization dashboard wizard.

Mirror Flows

Select to copy the packets in the flow to the IP address you specify.