Network Admin Nirvana: Fixing Performance Issues Before Users Notice

For many organisations, the status quo involves IT teams troubleshooting and firefighting network and application performance issues after users have called the help desk to complain about them.

This reactive modality can create a material impact to user productivity and has a knock on effect that impacts the organisation’s customers and partners if the issue is pervasive. Unfortunately, this is not uncommon. In recent times, we have seen multiple banking outages prevent customers from accessing their funds, credit cards, payment systems, and online and mobile banking platforms due to technical glitches. Even worse, what if technical glitches prevented access to patient medical records or heavy machinery equipment?

The exponential growth of digital transformation will continue, but so will the demands on IT further necessitating the requirement for organisations to get more proactive on health of their network and application performance.

What would it mean to your business if you were able to spot trends and have the insights to forecast degradation of network and application performance?

Organisations forecast their performance to measure their business health and investments as standard practice. The same cadence should be implemented for the organisation’s IT ecosystem to be able to predict and identify potential performance problems before users even notice.

In his recent article on ensuring productivity of staff working from home, Pinpointing Application Performance Issues with Unified NPM, Leigh Finch discussed how the combination of technology tools, processes and people gives you an effective solution for overcoming challenges on an ongoing basis.

Getting all of the data all the time

Data is key. When continuously monitoring the right data and accurate data, all the time, we can generate insights into IT ecosystem health and performance.

As Leigh said in his article, degraded application performance is not always a network problem. With Riverbed’s Network Performance  solution, you are able to continuously collect and monitor all packet data, flow data and device metrics and stitch the data together into meaningful insights.

Only an integrated (or unified) view from flow data, packet data and device metrics can provide multiple perspectives to see the data at different angles. Telemetry of the data consists of:

• Network flow data, such as NetFlow, jFlow, IPFIX, exported from routers, firewalls, etc.
• Device metrics, such as SNMP WMI, etc., polled from virtually any network device
• Packet data collected from on-premise, virtualised and cloud environments

That way nothing is missed—performance monitoring should be ‘always on,’ continuously monitoring everything everywhere.

Tools that sample data or do not scale across the entire IT ecosystem can likely be inefficient or ineffective. Organisations should consider the potential impacts to the business if the right data at the right time is not there during times of network and application degradation.

It’s better to have data that you don’t need than not to have data you do.

Getting the right data

To become more mature about dealing with performance health, the key is broader data visibility and multiple perspectives—with the right sort of insights from it to make actionable decisions. You want to anticipate and resolve degradation quickly. In other words, see it before your users, customers and partners do.

Imagine a non-typical scenario. A user calls the help desk to flag an issue. Instead of the help desk taking them through a routine of items to test, services to reboot, etc., they are able to tell the user that “we are aware of an issue with this application and we have identified the source. We expect to have it fixed in X minutes or Y hours.” The help desk can even let them know by email that a particular application is ‘acting up’ and under investigation to save them the trouble.

This shifts the paradigm and the perception of IT and the criticality of IT to the business, often overlooked, would positively change.

I’ve worked with a number of our customers who have made this paradigm shift and transitioned from being ‘reactive’ to ‘proactive’ using Riverbed’s Unified NPM solution.

Different teams and key stakeholders have visualisations of the right data: intelligent and insightful information that each needs—all rolled up into relevant fit-for-purpose dashboards, such as:

• High-level overview for the C-execs
• Business-centric and application views for lines of business owners
• Application-centric and performance detail views for application owners
• Network and device health views for network and infrastructure teams
• Key performance metrics and trending analysis views for operational teams

This enables them to get very accurate about what’s actually going on so potential issues can be identified and resolved fast.

Stronger security is a bonus

Getting all of the data adds also enables you to parse anomalous or bizarre behaviour within your network—such as a user accessing a server they’ve never accessed before. This gives you a stronger security posture because you can detect when it’s actually happening, rather than waiting for IDS/IDP reports to come in—which are only looking for signatures that something from outside your perimeter is trying to get in. But that’s a whole other topic for another article…