5 Myths About Network Performance Management
by June 1, 2011

Filed under: Network Performance Management, Performance Management Tech

Network Performance Myths and FactsWith the meteoric rise in enterprise deployments of IP-based services, network performance management has joined security at the top of every CTO’s list of concerns. It has become imperative to monitor end-to-end network performance from the point of delivery to the point of consumption from the user’s perspective, in order to know whether service levels are being met. For many organizations, nothing short of their ability to conduct business is at stake. Yet misconceptions about network performance management still linger. Are one or more of these “myths” putting your business services at risk?

Myth #1: We haven’t yet moved to “the cloud” so we don’t need network performance   management. You need to manage network performance if your business relies on any IP-based applications – not just cloud-based applications. These include: Software-as-a-Service (SaaS) applications like Salesforce.com or SAP Business ByDesign; Online backup; Collaboration applications like Sharepoint; mobile communications and presence-enabled applications; IP storage; Virtual desktop.

Many of today’s network-dependent applications degrade rapidly and crash abruptly if network performance falls below a specific threshold. These applications need not only sufficient bandwidth, but also minimal latency, jitter and packet loss. Even minor performance issues often result in problems like degraded VoIP call quality, slow application responsiveness, or complete service failure.

Myth #2: Can’t we do that with ping checks? While ping checks can give you a rough idea of the latency users are experiencing with web-based applications, they cannot provide the detailed network performance data you need to monitor user experience and ensure adherence to service levels.

Myth #3: Our cloud/SaaS provider will take care of that for us. Cloud-based application delivery totally depends on high-bandwidth, low-latency network performance. But that’s not your cloud provider’s problem. Most providers are not responsible for anything beyond their data centers and networks. They are not contractually obligated to monitor the delivery of hosted services end-to-end to your users, across networks they don’t control. One reason for this is that most cloud services run on TCP, which is highly location dependent. The distance that packets traverse between the cloud and the service consumers is a significant factor in overall performance. To monitor service levels across your distributed user base, your business needs the capability to remotely and continuously monitor whatever network infrastructure (WAN, remote LAN, VPN and/or web-based) supports the IP-based applications your users are consuming.

Myth #4: Our data center performance management tools have us covered. SNMP-based network management tools aren’t capable of monitoring the experience of users accessing remote, IP-based services, which depends on real-time network performance. Traditional netflow analysis tools can help, but they are expensive to deploy and manage and often consume significant bandwidth themselves. Most organizations therefore don’t have netflow analysis capability against remote offices – which is where it’s needed most. Without network performance management, it’s very difficult to know whether service levels are acceptable, and how what individual applications are consuming what slice of available capacity.

Myth #5: The technology is too expensive and infrastructure intensive for our remote sites. While this may have been true in the past – it is no longer. AppNeta currently offers affordable, instant-value network performance management solutions that enable you to ensure service levels across all data center, remote office, mobile and cloud environments. AppNeta’s cloud-based Pathview Cloud service delivers unequaled insight into users’ experience of network performance, while providing immediate time-to-value and on-demand scalability.