Today, our friends from Obkio will explore the top reasons why anyone should monitor network performance. You may be thinking that performance monitoring is complex and expensive, but it really is not. It can help you pinpoint the location of a problem—or sometimes prove that the network is not at fault. It can also help you identify, locate, and solve issues before they start affecting users. Furthermore, performance monitoring lets you easily troubleshoot network slowdowns and not just hard failures. It can be a great way to monitor remote sites without requiring local IT resources. It provides data that can be used in numerous ways to improve your networking environment and its operation. And last but not least, it can provide a smooth transition path to the cloud.
Reason Number One:
It’s Always a Network Problem! Or is it Really?
Any network administrator knows that as soon as things start to slow down, everyone is quick to point the finger at the network. There’s a simple reason for that: the network is often the culprit. However, often doesn’t mean always. In most jurisdictions, you are innocent until proven guilty. This is not so much the case with network administrators. It is usually their responsibility to prove that the issue is not with the network. Other teams will often not even look at the issue before the network has been categorically ruled out. So, how does one go about doing that?
The network is only one component of most IT environments, albeit an important one. Many subsystems are part of a typical service delivery environment. Some components can even be hosted in a public cloud, making it even harder to pinpoint the exact location of a fault and, more importantly, determine who is responsible for fixing it. One such typical case would be an OTT VoIP service hosted on Amazon AWS or Microsoft Azure. In this situation, an administrator’s first instinct will lead them to contact their VoIP provider after making sure that the problem is not within the local network. Chances are that the VoIP provider will say that the issue is not on their side and will blame the Internet.
We’ve all been on one of these never-ending conference calls with multiple providers where everyone seems to be pointing fingers and blaming someone else. No one wants the problem to be on their side mostly because no one has the proper tools to diagnose the issue.
By setting up a performance monitoring solution and by deploying monitoring agents in strategic locations, it becomes possible to lay out very clearly the responsibility domain of the various groups involved in delivering a given service. Furthermore, this performance monitoring solution lets every party involved pinpoint the source of the problem quickly and irrefutably, compelling the responsible group to take any necessary corrective actions.
Reason Number Two:
Detecting Problems Before Users – Or Your Boss – Do.
Identifying an issue from the user’s point of view can often be a challenge. Even more so when that user is the big boss. But detecting issues before they get noticed by users is the most difficult challenge. Although we have often means to measure performance between our data centers and branch office firewalls, it won’t be of much help to troubleshoot a performance issue that someone in a remote office encounters when communicating with another remote office. In addition, performance monitoring systems can often detect minute performance changes that could be advanced signs of some impending problem. They will literally let you detect issues before they start affecting users.
Traditional monitoring tools—like those that use SNMP polling, for instance—are not designed to provide performance statistics from a user’s point of view. They mostly are reactive tools that, while helpful for troubleshooting existing issues, are not the right tools for the early detection of performance issues, especially not from a user’s perspective. Deploying a performance monitoring solution gets all the blind spots left by most traditional monitoring tools covered. A performance monitoring tool will simulate real user traffic and will therefore measure the real performance experienced by your users. Additionally, most performance monitoring tools will let you establish a baseline measurement of each segment of your network. This is a very important feature as it will allow you to detect even the smallest performance degradation.
Deploying monitoring agents to constantly monitor your network’s performance from everywhere and from every angle is one of the best ways to quickly identify—and solve—issues as soon as they arise. You’ll often be able to fix issues before they are noticed by users—or your boss during his annual board meeting while live streaming a real-time video presentation—and, although that won’t make you look like a superhero, it will keep everyone happy, and that should be the primary goal.
Reason Number Three:
Troubleshooting Network Slowness Issues, not Just Failures.
Most network monitoring solutions on the market is actually fault monitoring systems. They will typically poll devices on a regular basis to ensure that they are up and running, and will alert you if and when they are not. The best ones will also read some operational parameter from the polled devices using SNMP. And likewise, they will alert you whenever some metric is abnormal. But as useful as some—or all—of this information may be, it does not say much about how the network is actually performing at any given time and between any two points from a user’s point of view.
On the other hand, true performance monitoring tools can help detect and troubleshoot any performance degradation. Haven’t you ever faced a situation where everyone complains about the performance of the network yet your monitoring tools are telling you that everything is alright? This is a typical case of using the wrong tool.
True network performance monitoring systems can simulate real user traffic between multiple points on your network and measure its actual performance—from a user’s standpoint. It’s like having network administrators everywhere constantly running performance tests, but more about this in a moment. Whenever a slowness issue is reported, the monitoring tool’s strategically located agents will help you quickly pinpoint the exact location of the issue and help resolve it quickly.
Reason Number Four:
Providing Historical Data and Establishing a Baseline.
When diagnosing network issues, a plethora of tools are available, allowing one to take steps to identify whether there really is a problem or not, its nature, and its location. Ping, which will send echo packets to a host and measure the host’s response time is the first such tool that comes to mind. Traceroute is another useful tool which will use successive echo packets to display the path to the destination as well as the response time of each hop. Web-based tools such as speedtest are also useful in evaluating the performance of your Internet access.
However, many of these tools’ usefulness is limited if you don’t have anything to compare the measured data to. For instance, perhaps that 60 ms response time between the head office and a remote location is “normal.” Without knowing what the usual value is, it’s hard to establish for sure that there is an issue. What you need, then, is a means to establish a baseline of your network’s performance, and this is precisely one of the things a performance monitoring tool can do for you.
With the historical data which is collected by a performance monitoring tool, you get the benefit of having access to a comparison point. Another huge advantage of these systems is that they are able to automatically and continuously compare the current data to the historical one and raise an alert as soon as performance starts to degrade.
Historical data also brings another important benefit: It can allow an administrator to troubleshoot a past event. Suppose you come in on a Tuesday morning after a long weekend and find an email from a user complaining about poor performance last Saturday afternoon. Having access to historical data could help you troubleshoot that issue even though things might have since returned to normal. It will let you identify exactly when and where the problem occurred and help you ensure it never happens again.
Reason Number Five:
Distributed Monitoring Provides a Smooth Transition to the Cloud.
For the past fifteen years or so, it’s been quite common for organizations with multiple sites to adopt a “centralized” telecom architecture and interconnect all locations to one single, unified network. This allowed businesses to keep their data and its processing within the head office and/or data centers. It was also typical to have a centralized Internet gateway shared by users from all locations.
The only performance monitoring that was needed was typically between the head office or data center and the remote locations. This provided sufficient coverage to catch most business-impacting issues and for that, traditional monitoring system was usually adequate.
Organizations—which were rarely using Internet-based VPN connections—used MPLS networks which, although expensive, offered guaranteed quality of service (QoS). Issues associated with packet loss, latency, or jitter were uncommon and so were performance monitoring tools. They simply were not needed and their return on investment was difficult to demonstrate. Bandwidth monitoring was typically all those enterprises needed.
Flash-forward to today and things are quite different. Organizations are increasingly using cloud-hosted applications services. And this is not only the case with office tools such as Office 365 from Microsoft or G Suite from Google. They also use that model for business-critical applications such as ERP, CRM, and other project management tools. Accessing these services through a centralized Internet gateway—often hundreds if not thousands of miles away—is no longer acceptable. The few milliseconds of round trip time added to each transaction quickly adds up to a considerable delay. Businesses are, therefore, reverting to using local Internet gateways in each location.
The transition to cloud-based services has led organizations to leave the centralized model and switch to a more distributed architecture. Unfortunately, this can quickly complexify the data flows between users, thereby rendering traditional monitoring tools less interesting. They’re simply not up to the task. This is where you need a distributed performance monitoring solution which can truly monitor the network’s performance from a user’s perspective and from every possible angle.
Reason Number Six:
Removing the Need for Remote IT Workers.
Centrally managed IT is an ever-growing trend, and since organizations no longer have IT resources in remote offices, most IT assets are being hosted at the head office—or in a dedicated data center—alleviating the need for remote IT resources. But while IT assets might be limited in remote locations, some equipment such as networking equipment is still there. A consequence of this centralization is that IT resources are getting scarce, teams are getting smaller, and the need for great tools is more pressing than ever.
When you need to manage multiple networks—some in very remote locations and with limited WAN connectivity—with no on-site staff, what you need is a way to “virtually” have people in each location. Network performance monitoring tools can offer that functionality, and they can operate over any type of WAN connection. It is irrelevant whether you’re using the VPN, SD-WAN, or MPLS as WAN connectivity. A good performance monitoring solution will have you covered.
As we hinted earlier, deploying a performance monitoring solution throughout your network with strategically placed agents will give you visibility in even the most remote sites. You can be confident that the agents are constantly measuring and evaluating the performance of the network from a true user’s perspective. Even more importantly, the agents do it on a continuous basis, day and night. Try to get that from IT staff!
Reason Number Seven:
Gaining Maturity by Monitoring Otherwise Undetectable Parameters.
Network performance monitoring will often let you identify real issues that more traditional monitoring tools may be overlooking. One of the reasons is that they simulate real usage scenarios. It’s like having IT staff everywhere constantly running tests! All the performance data that is gathered by performance monitoring tools can be used to your advantage.
One of the best ways this data can be used is capacity planning in a very broad sense. We’re not just talking about throwing in more bandwidth. Any bandwidth monitoring solution can help you with that. With true performance monitoring tools, you can make the proper diagnostic and ask yourself the right questions about various aspects of your network such as:
- The inherent design of the network,
- The necessary equipment acquisitions or upgrades,
- The sizing of the networking equipment,
- The performance of your service providers (including their SLAs),
- The operation and maintenance of the network.
This data can also be analyzed and used to help you make the right decisions about custom-building or expanding your network. It can also provide invaluable insight when managing costs and prioritizing network improvement investments. And last but certainly not least, it can also help identify the need for additional internal or external resources.
There are many reasons why one would want to deploy a network performance monitoring solution. Some have to do with assuring the performance of the network, while others deal with detecting and fixing issues. They can also address the needs of reduced IT staff teams with limited remote resources, and they can provide decision-making data that can be used in many different ways.
The key element here is that performance monitoring tools do provide much more useful operational data about your network than any other type of monitoring tools can. This does not mean that your good old SNMP monitor is not good anymore, but it could certainly be complemented by deploying a performance monitoring solution which, in a nutshell, starts where the other tools stop.