In the past, network management mainly consisted of monitoring whether routers and switches were up or down — as long as the power light was on, it was available, and the network administrator could proclaim, “It’s not the network.”
With pervasive wireless connectivity and handheld devices capable of accessing the Internet from nearly anywhere, networks have become more advanced. In turn, users are focused on the level of service the network provides, rather than it simply being available. The tools to monitor and manage the network must also keep pace with growing end user expectations by being more usable than the old frameworks and offering more functionality than low-end red light/green light network utilities.
Great network management is achieved by carefully controlling the collision between user expectations (demands) and the technology available to the IT department.
With few exceptions, best practices result when the IT department focuses on the user perspective on the infrastructure, typically as a result of being business- or business-unit driven. Demanding users know what they want and do not want, and state their needs clearly, if not a little abruptly at times.
A Step-by-Step Look at User Demands
From the user’s perspective, expectations for network performance can be summarized in the following way:
- I shall always be able to work. I presume this means that the network shall always be up.
- That being said, it’s not enough for the network to simply be “up,” it must be running well.
- It’s not enough to tell me the network is running well, you must prove it to me, and tell me how well the network is running using metrics I can understand — and I’ll judge for myself.
- I should also be able to periodically check up on the quality of your service (and how you are taking care of me) on my own.
- I shouldn’t have to be an expert on all aspects of computer architecture in order to get support — all I care about is whether I am able to work. I want one support phone number to call — you sort out what’s wrong. I don’t want one phone number for the systems group, another for the application virtualization group, another for the network group, another for the database people …
- It may take a village to raise a child, but it shouldn’t take one to maintain a network.
These simply stated requirements are not always easy to fulfill, particularly when technologies are quickly evolving (SNMPv3, IPv6, 802.11n), applications are continuing to expand (in size, number, and memory and CPU consumption), application delivery mechanisms are continuing to morph (SOA, SaaS, virtualization), and new business initiatives are being required (VoIP, wireless, green IT).
Six Tenets of Green IT Departments
Green IT departments all seem to have these similar mantras:
- If it can be automated, automate it. If you think it can’t be automated, figure out why, and then automate it.
- The squeaky wheel gets the grease — the first time. The next time it squeaks, start planning to replace the wheel (and keep in mind that the “wheel” could be a network device or a member of staff).
- If a management tool is failing to live up to expectations, get it fixed or get rid of it. If it used to work, but isn’t keeping up with technology, replace it with one that is.
- If it is taking too much effort to run a management application, replace it with one that takes less effort. The personnel cost to run management software should not be more than the cost of that software, and it should not take months to learn and master.
- The investment in a good tool almost always pays for itself quickly. Monitor this and make sure that it is true.
- Monitor, measure, review, improve; monitor, measure, review, improve … nothing will ever be perfect, but everything should constantly be improving.
Meeting Demands With Best Practices
The specific capabilities of a network management solution emerging from the collision of two world views of user demands and infrastructure resources are fairly clear. What is called for is simply a modern network management solution that has serious functionality to address contemporary needs of business.
- It must support the latest technologies (devices, protocols) and initiatives (reducing power consumption, going wireless, optimizing resource utilization).
- It must be easy to learn, as well as easy and affordable to maintain and must provide a high degree of automation.
- It must significantly help with troubleshooting; i.e., it must have excellent event handling and root cause analysis.
- It must go beyond event-based troubleshooting and give clear insight into performance issues, with real-time and historical data, and even forward-looking projections.
- It must produce meaningful data and metrics — meaningful to the network staff as well as meaningful to the business users and management. That data must be open and accessible to a variety of users for a variety of purposes.
- It must look beyond the network and also concern itself with “non-network” components or elements that connect to the network (servers, workstations, applications) and/or it should integrate to management applications that do.
Given this list, it has become evident to many IT departments that it is time for an update in their management solution set, and that the solutions available for the last two decades are no longer sufficient to meet today’s technical needs and user expectations. The good news is that there is a variety of modern applications ready and able to do so.
Michael Jannery is president and CEO of Entuity, developer of the Eye of the Storm network management suite.