Business processes are application-dependent, and applications are network-dependent. With customer experience and business productivity on the line, traditional service level agreements (SLAs) focused on aggregate bandwidth delivered; uptime and mean-time-to-service-restoration are inadequate. Network SLAs must include how networks impact application performance (i.e. user experience and QoS).

So, for applications to perform as businesses and consumers demand, there’s no tolerance for “dumb pipes.” Network operations and engineering teams must deploy monitoring intelligence that understands how, why and when networks impact application service quality.

Intelligent instrumentation at work

Here’s an example of a financial institution with a worldwide network consisting of 15,000 network devices, including routers, switches, wireless access points, load-balancers and firewalls. They monitor basic SNMP MIBs for all the devices’ components, including 250,000 interfaces, and can see all the basic key performance indicators and trends for CPU, memory, disk, packet loss, etc. To make their pipes smarter, they also monitor approximately 70,000 QoS Class Maps, 90,000 QoS RED statistics, 285,000 IP SLA tests, 785,000 poll instances and a whopping 43,000,000 flows per minute.

When asked, “Why so much instrumentation?” the monitoring manager replied, “Because the network is the backbone of our business.”

Smart pipes are just as essential for communications service providers (CSPs), because for them, “the network is the business.” Just providing guaranteed bandwidth for Internet access, a remote office network or a mobile backhaul isn’t enough for their customers’ who, like all end-users, naturally judge network service in terms of application response (i.e. user experience).

Consider the example of a CSP who, besides monitoring essential SNMP metrics for network device health, collects NetFlow data from its enterprise customers’ networks to understand and manage bandwidth for critical applications and traffic patterns from office to office, office to data center and office to Internet and cloud resources.

They identify top bandwidth-consuming users, protocols and applications and highlight the IP addresses of the top talkers. This helps them understand if certain applications, locations and users are being starved – or are likely to be starved – bandwidth, so they can take remedial (or proactive) actions. Looking at network device performance indicators and bandwidth utilization together make pipes smarter and provide more intelligent insights.

This kind of intelligent instrumentation is necessary for the full range of CSP wholesale and retail services, including:

• Managed WiFi
• Optical and Microwave Transport
• Metro- and Carrier-Ethernet
• LTE (Long-Term Evolution for Mobile Wireless)
• Voice over LTE
• Ethernet Backhaul.

Are your pipes smart enough?

CSP and enterprise network operations and engineering teams must evaluate if their current monitoring tools can make their pipes smart enough to meet the demands of their business.

Here are 10 fundamental criteria for evaluating your network’s intelligence:

1. Does your network monitoring tool unify multi-technology, multi-vendor device monitoring, covering SNMP and Non-SNMP devices, in a single dashboard?
2. How comprehensive is the list of multi-vendor network devices that are certified for monitoring?
3. Does your tool enable self-certification of new devices and MIBs?
4. Does your tool provide device performance as well as network flow analysis and application delivery analysis in a single dashboard?
5. Does your tool have analytics and scorecards that show key performance indicators in relation to applications, business cycles, customers and locations?
6. Does your tool help you proactively manage capacity and combine trends with events to speed triage?
7. Does your tool enable instant visualization and scrolling of high-volume analytics, and eliminate waiting time for metric panels, dashboards and reports to populate?
8. Can your operators and engineers easily customize their own dashboards and reports without the help of administrator and with having to use scripts?
9. Can your tool integrate with CMDB, Service Catalogs and other OSS/BSS systems to automate monitoring and eliminate manual on-boarding?
10. Can your tool linearly scale with a minimum of additional physical or virtual hosts to keep capital and administrative costs from getting out of control?

How did you do? Leave me a comment below or Tweet @CAInc

Here are some helpful resources:

• a slide presentation about the top 10 Criteria for Communications Service Providers
• a slide presentation about the top 10 Criteria for Enterprises

The post Why there’s no room for dumb pipes in your network appeared first on Highlight.