The “cloud” era has led to an increase in usage of web applications that are delivered by network service providers to their customers over the network. Because these web applications are revenue generating, it is important for service providers to understand how they are performing and identify any potential service degradation before users begin to complain. One method for analyzing IP application or HTTP service performance is to use the IP SLA HTTP operations supported on Cisco routers. Note, Juniper, Huawei and Nokia provide similar tests for their devices.
IP SLA is a feature on Cisco devices that can be used to measure network performance across a network path between different nodes. Cisco IP SLA simulates network data across a network path and collects performance information such as jitter, latency, delay, round-trip time (RTT) and packet loss. This information can then be used by network engineers to understand how a path is handling specific traffic. While there are many types of IP SLA operations, the one that is used to monitor service performance is the IP SLA HTTP test.
The IP SLA HTTP operation measures the RTT between a Cisco device and an HTTP server when retrieving a web page. There are three types of measurements that are collected by the HTTP test:
DNS Delay – This is the RTT taken to perform a domain name lookup.
TCP Delay – This is the RTT taken to establish a TCP connection to the HTTP server.
Transaction Delay – This refers to the RTT taken to send a request and retrieve the home HTML page from the HTTP server.
Using this information, the total HTTP RTT, which is a sum of the three delays above, is calculated.
For information on configuring IPSLA HTTP tests on your Cisco routers, check the Cisco configuration guide:
The Performance Explorer module in the Packet Design Explorer Suite collects SNMP based performance metrics from the devices it monitors. Performance Explorer supports IP SLA HTTP operations, as well as those for other manufacturers, making it a useful tool to monitor service performance.
While measuring RTT for web services helps identify potential issues, it alone cannot help in identifying the root cause of many issues. This is because IP SLA HTTP operations do not include information about the path taken by the HTTP traffic from source to destination. And this means that while the network engineer knows that there is an issue with HTTP services, they do not know the “where” and “why” of the issue.
So, how can service providers find out what caused an increase in RTT or delay for a specific web service? How do they pinpoint if the cause of the issue is their network or if it has something to do with the remote network? This is where correlation of HTTP service performance metrics with routing path (control plane) information can help.
The Route Explorer module of Packet Design’s Explorer Suite collects information about IGP and BGP routing paths in real time, displaying the exact path taken by traffic from a source to destination. By correlating this with the measurements from IP SLA HTTP tests, the network operator sees if the traffic was indeed taking the shortest path to a web server’s prefix as intended or if a routing issue caused traffic to be redirected over a longer path resulting in an increase in HTTP delay. This correlation of HTTP service performance metrics with routing data helps network teams identify service performance issues, quickly find the root cause and troubleshoot them.
To see how the Explorer Suite monitors HTTP service performance using IPSLA HTTP operations correlated with routing paths, check out the short demo video below: