Riverbed on Real User Management

A conversation with Riverbed's Michael Cucchi highlighted the issues companies have with really knowing how their workloads perform and how to find problems when they arise.
Written by Dan Kusnetzky, Contributor

Michael Cucchi, Riverbed's Senior Director of Product Marketing, stopped by to discuss the integration of Riverbed Cascade and OPNET product families with the introduction of its single appliance with integrated application aware network performance management  (aaNPM) and application performance management (APM). It is clear that Riverbed is doing its best to offer ways for companies to get control of what has become a complex and messy distributed mess of technology.

What's involved

When someone begins to use a application, a long list of products and technology come into play. A problem with any one of them could lead to this person experiencing a problem. Let's get digital for a moment and follow the stream.

  • The person touches, clicks on or types something that engages a program that is executing on an "end point device." This device could be a smartphone, a tablet, a laptop, a desktop or some other form of computer. Programs running in this device reach out to a network to access data or an application. The device, the software running in the device or its access to the network could work perfectly and yet this person could have a bad experience. The application could be slow or not work at all.
  • A network link is engaged by the end point device so that the person can access remote applications and data. Even if this link worked perfectly, applications could be inaccessible or the performance could be unacceptable due to problems further upstream.
  • An server running an application or a Web server accepts the requests sent by the person's device, gathers up needed data, performs some application logic to produce the needed data and sends back a response to the person's device.
  • Another network link is engaged to carry messages to and from the storage subsystem.
  • The data management software software that controls access to the data could be executing on a physical system, a virtual system or somewhere else on the network in a cloud service provider's data center. The data that this application accesses could be contained in local files which could be on a local storage devices, a storage device attached to a local storage server or in a cloud service provider's data center. The data could also be in a database stored locally and accessed by a local database server, stored remotely and accessed by a remote database server or stored at a cloud service provider's facilities.

 This is just one fairly simple example. Real world applications are often far more complex relying on many tiers of services, each of which may also include a long string of components.

Riverbed, and its many competitors, would point out that any number of things could cause the person's experience to be bad. Overworked or badly configured servers  or storage devices could be creating slow downs. Problems with network links could create problems that might be difficult to find.

The issue is increasingly related to the complex environment we've created as an IT infrastructure. Many different things need to be monitored. The monitoring itself can't take up so much of the limited IT resources or it could create problems in and of itself. Understanding what the monitoring shows now requires extensive and deep expertise in everything from operating systems, virtualization, storage, databases, applications and networks. A problem with a failing storage device could present itself as a slow database or a failing application.

Riverbed and its competitors all see the problem and are attacking it from just about every conceivable direction. After speaking with Riverbed, it is clear to me that they've done their best to find a comprehensive, easy-to-use set of tools that could be an important part of an IT organization's performance monitoring toolkit.

Editorial standards