Assuring Robust Delivery of Critical Business Applications with Riverbed Cascade (formerly Mazu)

As IT organizations strive to reduce costs through server consolidation, virtualization and datacenter transformation, the task of delivering robust business services becomes more difficult. Visibility to how applications traverse the infrastructure is reduced by virtualization and WAN optimization technologies. The increased number of application delivery paths weighs against optimizing the entire delivery infrastructure through bandwidth upgrades, faster switching technologies and WAN optimization. Even the idea of monitoring application response times with threshold based tools absorbs more labor hours than the average IT staff can spare.

Cascade customers have unprecedented visibility to how applications perform across their infrastructure including:

• Layer 7 application traffic across the entire infrastructure
• The delivery paths of applications from back-end servers to the user's desktop including WAN interfaces traversed
• Normal patterns of response time, user connections, bandwidth utilization, QoS and application health statistics
• Visibility into http traffic to distinguish between web front-ended business applications and non-business use

Cascade’s holistic view of application delivery includes analyses on # of hosts serving and users connecting to an application

Cascade customers have unprecedented automation when monitoring and analyzing application performance including:

• Baselines of typical interactions of users, servers, network and applications by time of day, day of week, etc.
• Behavioral analysis to determine when current conditions are outside of predicted behavior patterns
• Actionable alerts with rich contextual information to streamline the troubleshooting process

Cascade alerts include an impact score which helps focus operations staff on the most likely root cause of performance issues

Riverbed Cascade combines the best approaches to monitoring, analysis, reporting, and alerting in one highly automated solution. Cascade incorporates advanced Network Behavior Analysis (NBA) technology. Cascade's NBA implementation is the most advanced in the industry allowing customers to implement automated network and application monitoring and analysis with a significantly reduced burden on expert operations staff. Cascade's unique User Defined Policies, allow users to easily construct advanced monitoring and control policies to avoid known contributors to degraded performance. These technologies combine to reduce the number of false positive alerts or 'alarm noise' common to threshold based monitoring systems. Additionally, Cascade provides rich contextual information for each alert that allows a) help desk staff to solve more problems and more correctly route escalated tickets; and b) expert staff to identify the root cause of performance issues more quickly shaving as much as 75% off of the time spent troubleshooting.

Cascade's unique capabilities enable our customers to significantly streamline operational planning and troubleshooting workflows across multiple operations teams. The most common use cases include:

• Managing the WAN to accelerate business applications
• Assuring robust delivery of critical business applications
• Minimizing the risk of datacenter moves
• Reducing the cost of regulatory compliance

Cascade in Action

A retail service provider operated call centers accessed by thousands of retail outlets. One Saturday morning, the VoIP traffic which intelligently routed calls to available experts, came to a standstill. For four hours of a day in which normally 60% of a week's transactions are done, business came to a halt. At the 4th hour, the escalation policy included security operations to assure the company was not under attack. Within 5 minutes of applying Cascade to the problem, security had determined there was no attack. Within 15 minutes they had identified the culprit as a new Exchange server which had swamped the main trunk line to the datacenter with replication traffic. Within 30 minutes business was restored to normal. The company estimates that a problem that would've taken business down for 8 hours was identified and resolved in 30 minutes using Cascade.