Enterprise Monitoring

Published
  • January 25 2008, 4:39pm EST

Like the straw that broke the camel’s back, an abundance of IT utilities can have the same impact on network performance. With new types of enterprise IT technology coming out of the woodwork everyday – each with its requisite monitoring software – operations personnel certainly have no shortage of tools at their disposal to ensure performance. Additionally, operations personnel in growing enterprises typically have their collection of favorite tools that they discovered and accumulated from reading blogs, attending trade shows, visiting social networking sites, or just swapping stories around the water cooler. Deployed chronologically and with specific business benefit in mind, these applications – and the demand they place on network infrastructure – continually accumulate like the sands of time. But are these tools using the same techniques or overlapping queries in gathering forensics – negating their proposed benefit?

What many don’t take into account is that each new application is created for a single purpose and executes without regard to other applications running on the network – playing well in its own sandbox, but frequently using the same data gathering techniques as other utilities. Each application intentionally delivers on its intended benefit, but compounds the management traffic as it keeps its own pulse. The result is a bottleneck from the very utilities designed to improve network performance. Applications such as peer-to-peer technologies are flourishing and, in turn, are placing increased demands on the network by generating more data. In short, the utilities that were supposed to improve performance are actually sapping it from the business services and infrastructure users rely upon for their productivity.

While it may seem the best approach is to just put your head in the sand, ignore the problem or simply tell users to go away, this will not stop the unending tide of productivity-improving enterprise applications from flooding your network. The better solution is to turn to the often overlooked tool that can ease the pain: the network management solution. With so many interrelated and dependent network services keeping your company’s IT infrastructure running, getting your arms around the network performance issue may seem insurmountable. Answering the following questions will quickly help you lay out your action plan to get the network performance your business requires – and already paid for – while likely saving some money and time along the way. Now that is a bargain!

How Did We Get Into This Mess?

Several years ago, keeping the company’s IT infrastructure running was a simple matter of ensuring the mainframe and VT terminals were under power. The network was simpler with few, if any, business critical services running across it and a single management tool for the job. Now, more and more technologies, services and overlays are being applied (IDS, IPS, virtualization, QoS, MPLS, VoIP, CRM, ERP, middleware, IP-CCTV, video conferencing and intranet access), driving the network to become business critical and necessitating 24x7 availability without fail.

This mix of business services resulted in networks hosting a significant number of highly dynamic applications all depending on the same network infrastructure and all competing for resources. The typical approach to keeping things running was usually the original basic network management tool. This basic tool was either adequate or would be upgraded with a collection of niche solutions, which were usually poorly integrated. Now, with a bigger mix of more complex applications, what is needed to manage today’s complex network is a single network management tool capable of collecting data across the shared network to present real-time and historical availability, performance and configuration statistics on individual services and applications. Unfortunately, all too often, the need for this cutting-edge all-seeing network management tool is not recognized until a severe network outage occurs and the business realizes just how dependent it has become on its network and how expensive outages can be.

Adding insult to injury, Web 2.0 and proliferating peer-to-peer communication is resulting in new traffic flow patterns and volumes; management tools are now required to cope with more connections, as well as monitor and diagnose technologies that have come rapidly to the forefront of the network administrator/operator’s arsenal, such as QoS diagnosis and netflow traffic analysis.

How Does This New Application Help?

No company executive in his right mind – or with shareholders to answer to – would implement technology for technology’s sake. The goal of any IT initiative is to increase profitability, either by reducing expenses or increasing revenue. While it is true that IT initiatives such as VoIP, security, compliance and virtualization can offer potential productivity benefits and economies of scale, can the network handle the strain? Deploying any of the new “must have” IP services will place additional demand on the network – potentially negating any productivity improvements by creating bandwidth bottlenecks and eroding cost savings. Whatever the size of the business, a well-planned strategy, including due diligence on your infrastructure, can ensure a realistic level of investment. This will help ensure that the network is optimally deployed and performing efficiently, regardless of the IT initiative.

Where Do We Start?

To reach your final destination, you have to know where you currently stand – and that begins with a complete assessment of your IT infrastructure. However, a simple list or spreadsheet of your assets is not enough. Most network management technologies can automatically determine the devices that make up your network, but more advanced solutions will also indicate their connectivity. This is a critical step, particularly for network applications that are sensitive to latency or jitter. More advanced technologies will also indicate any excess capacity that you might have that may be more optimally used. For example, a network switch may have wires in every port, but 20 percent may not be in active use. This could give you the headroom you need without any additional hardware purchase.

It is equally, if not more, important to determine how these assets are performing to establish a baseline assessment. Performance is not just a measure of real-time status, but one of historical trending. Are there any current usage patterns that could impact the new service or, conversely, will the new service cause service interruption on currently deployed applications? It is doubtful the accounting department will be happy if they can't process payables or receivables, even if their new VoIP phone is working flawlessly. With several tools, one often has the same data in several places – duplicate polling, inconsistencies, time slew, etc.

Trying to grasp a meaningful understanding of performance across a collection of disparate tools is difficult and costly. Operating expenses are compounded for many reasons in multiple applications, including:

  • Managed inventory maintenance (adding new devices to several management applications).
  • Configuring business-relevant views in a number of applications.
  • Configuring user permissions in each application.
  • Software maintenance (upgrade costs, patch costs, cost of installing).
  • OS maintenance on many platforms.
  • High training costs due to the many disparate applications involved.
  • High contract management involving a range of vendors.
  • Complex and expensive management of many applications.
  • Expensive maintenance of custom integrations.

Who’s Talking to Whom?

On your network, not all conversations are created equal and a multitude of management applications are likely having redundant conversations with the same infrastructure. Knowing what application or service protocols are in use will not only help you provision for the “chattiest” offenders, but the most threatening as well. By their nature, viruses and Trojan horses keep their signatures low to avoid detection. A unified approach that combines this level of flow-based information with the depth of infrastructure, connectivity and event detail of a network management solution can deliver businesses better control and predictability of user activity, as well as capacity planning and traffic engineering, proactive infrastructure and security analysis, and service level reporting for accounting and billing.

Likewise, not all conversations require the same priority. ”Who or what deserves the highest priority?” and Who can live with a slower response time?” are questions that will help guide network performance by tuning quality of service QoS policies. With visibility into QoS configurations in the context of the overall network, managers can see the current and historical use and performance of network resources, monitor and report on congestion, correlate QoS configurations with network performance, and use the information to improve traffic and service delivery. This enables managers to match service delivery with the priority of the user or business unit for that service.

Where’s the Proof?

Although technologists will argue that you can never have too much information, too much data can make decision-making difficult. Today’s enterprise applications efficiently gather countless amounts of statistics, but typically keep that data trapped within the host application. These point solutions have few, if any, data integration or export options, preventing adequate reporting, archiving (for SOX, COBIT, ITIL), portal population, mashups, base lining, trending or capacity planning. Gathering up the supporting data, massaging it into compatible formats, and importing to some reporting or display technology is a process fraught with time, expense and compromise. It generally results in something that doesn’t quite hit the mark. Reporting across a number of disparate tools is very difficult, but markedly easier if all of the data is available in a single tool. To some extent, this can be mitigated by data export, but this is often not available in point solutions.

Dashboards and Web displays deliver visually compelling and graphically concise reports that enable organizations to make the right decisions faster and with more confidence. Executives and business unit managers demand concise and graphical summaries of key network statistics and capacity utilization trends to empower decisions and actions. Flexibility in data presentation is also necessary for Web-savvy users who individually demand tailored portals or displays. Separating the data layer from the presentation layer ensures a new level of openness and accessibility of your data to accelerate and simplify business reporting, corporate dashboards and mash-ups using the display or charting tools of your choice.

Where’s All This Going?

The only thing constant – particularly with technology – is change. Putting aside for a moment the nearly instinctive need for technology people to have the latest, greatest and shiniest gadget du jour, technology marketers are constantly working on the next big thing to change the way you work and separate you from your hard-earned cash. As users and their applications become more technologically advanced and revolutionary Internet technologies and Web 2.0 blur the line between online data stores, centralized enterprise applications and individual workgroup databases, the importance of round-the-clock accessibility to all data sources becomes paramount. Today’s knowledge workers are creating composite applications and mashups using Yahoo Pipes, Google APIs, Groove shares and XML to piece together the interesting bits of data to answer their most burning questions. Regardless if the connection is around the globe, down the street or to a workstation in the next cubicle, the foundation they share is the network. Only a comprehensive network management solution with visibility across infrastructures, services and business views provides proactive insight to key performance indicators and utilization statistics that keep all the connections alive and well.

Any Last Words?

Today's complex IT initiatives and the converged network they require demand comprehensive and detailed insight to ensure control and predictability of business-critical services with maximum return on investment. Answering a few simple questions can keep you out of hot water, your users in the pink and your budget in the black.

Comprehensive network management systems provide deeper analysis than a collection of broad, but shallow open source or point solutions for management. It’s this type of visibility that has gone from being a luxury to a necessity when managing today’s business-critical, service-oriented network. Networks have grown up – and the tools for their effective management have grown up too.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access