As your organization seeks to efficiently connect, integrate or even dissolve teammates, partners and infrastructures in other companies, time zones and cultures, there are certain risks posed to the performance of your networked applications. But there are ten important steps CIOs can follow to mitigate these risks and efficiently and seamlessly maintain the performance of business-critical applications as their organizational structure evolves. Let’s take a look at these ten steps.
Step No. 1: Take an inventory of applications
Do you know what’s running on the merged network? Many organizations don’t have any idea what applications are running on their network and how much of the traffic is business-related versus recreational. This has implications for performance and security as well as bandwidth provisioning.
Look for network performance monitoring tools that use a technology called NetFlow (created by Cisco but now an industry standard called IPFIX) to identify the who, what, when and where of applications traversing the network, as well as how much bandwidth is being consumed.
Step No. 2: Determine all application dependencies
Do you know what infrastructure components are used by each critical application? Understanding the infrastructure is critical to mitigating risk. For instance, a large commercial vehicle parts manufacturer that spun off from its giant parent corporation was immediately faced with delivering network services to thousands of employees in 30 countries.
The network group had to quickly understand its global network infrastructure and application environment. The group also had to quickly discover which resources belonged to them-including WAN links, servers, shared buildings and applications-in order to transition the appropriate resources away from the parent company and ensure optimal network and application performance. Network traffic analysis, application response time monitoring and device performance monitoring technologies are all critical in this situation.
Measure Infrastructure Performance
Step No. 3: Measure infrastructure performance
Do you know how well your mission-critical business services are being delivered? The job of the network infrastructure is to take an IP packet from one end of the network and move it consistently and reliably to the other.
Do you have metrics for that process for all your key applications and remote sites? For instance, how is SAP performing across the network for your remote users? The question is not just, “Is it available or not?”
Calculating application response times across the network to users is a good starting point for understanding the before and after impact of change. This should include how much time an application is taking between tiers of a multitier application so you can determine whether a slowdown is in the network, server or application. In addition, establishing baselines for all critical applications is essential to understanding what “normal” performance is for each (and thus, being able to know when deviations occur).
Step No. 4: Monitor VOIP and video
Will voice and video traffic between the new entities impact data applications? Is your network prepared for a convergence of unified communications (UC) and data services? The addition of streaming application services such as voice over IP (VOIP) and video can negatively impact the performance of data applications. Your IT organization must constantly monitor response times, bandwidth consumption and service quality.
Calculate Bandwidth Allocation
Step No. 5: Calculate bandwidth allocation
If consolidation is a goal, how much bandwidth is needed for the various remote locations to handle the merged application portfolio and employee base? Knowing which applications and hosts are consuming bandwidth across the network enables organizations to make informed decisions about bandwidth investments. For instance, a large chemical manufacturer helped ensure better performance of its critical applications and cut its global bandwidth budget in half by understanding its traffic composition.
The company was surprised to find that 50 percent of its global bandwidth was being consumed by e-mail, 30 percent by Internet browsing and 10 percent for print and other small jobs. This left less than 10 percent of the bandwidth for its business-critical applications. Because e-mail and Internet browsing are not time-sensitive, the company decided to use the less expensive public Internet to deliver this traffic. In doing so, they removed 80 percent of the traffic off its core Multiprotocol Label Switching (MPLS) network serving the business and lowered its service provider costs.
Step No. 6: Establish service-level agreements
What service-level goals, formal or informal, need to be put in place or modified? Establishing service-level agreements (SLAs) around application delivery-not just availability-provides a measuring stick for the level of service you are providing and when it degrades from normal. SLAs can also help you determine whether your external service providers are meeting their service guarantees.
CIOs should insist on performance-oriented SLA commitments from their service providers that are related to key business applications, not just availability guarantees.
Step No. 7: Implement and monitor QOS policies
If applications from the merged entities will share links, what quality of service (QOS) policies need to be created or updated? The ability to set QOS policies for which types of traffic get priority across your network is imperative, especially as your organizations uses more VOIP and video streaming. This requires the political will and ability to agree on priorities between the different lines of business using shared infrastructure. Ensuring that the policies are working is essential.
Consolidate Network Personnel
Step No. 8: Consolidate network personnel
Will there be a combined network strategy and management team? Ensuring optimal network and application performance often hinges on the answers to the following five questions:
1. Who will need visibility into what areas of the network?
2. How will management of the combination of networks be centralized?
3. Who will be responsible for network planning and architecture?
4. Will they have access to the historical and trend data needed to make educated infrastructure investment decisions?
5. How will visibility into performance metrics be restricted to authorized users?
An effective network strategy also requires good lines of communication between other IT groups, including the application and server teams-especially with the advent of technologies such as virtualization.
Step No. 9: Manage network change and configuration
How will changes be planned and implemented across the combined network? Facilitating collaboration across your IT organizations is critical to maintaining high performance levels. What processes need to be implemented? How will changes be communicated across the different IT groups? What methods will be used to measure the impact of changes on application performance?
The answers to these questions will help you to successfully plan and implement changes across the combined network. Performance management tools with views that can be easily shared among different groups are essential.
Step No. 10: Fortify network operations
Who will monitor the combined network for application performance? How will they know what the “new normal” performance is? Network operations or help desk groups are often the first to tackle a performance issue and are taking on more and more responsibility for application performance.
Do your operations personnel have the tools and skills needed to monitor application performance across the network? Can they solve basic problems themselves without having to escalate every ticket to the more expensive level two or three engineers? Ensuring the operations group has the right data and means to collaborate effectively with network engineering and other groups is the key to solving problems quickly, at the least cost.
Joel Trammell is co-founder and CEO of NetQoS, Inc. Since co-founding NetQoS in 1999, Joel has led the company to become one of the fastest growing in the U.S. In 2005, the Austin Business Journal recognized Joel as its Private Company Executive of the Year. In 2006, he received the Ernst & Young Entrepreneur of the Year award. Joel’s achievements in IT span more than two decades and include numerous positions of leadership and management, most recently with Advanced Micro Devices, UST Computers (where he was the President and CEO of the Austin-based provider of computer hardware and network services), and HomeSmart, a company that he co-founded in order to provide sales and marketing products for the residential construction industry.
Joel began his career as an Officer and Instructor/Division Director at the U.S. Navy’s prestigious Naval Nuclear Power School, instructing more than 400 naval officers in thermal sciences and nuclear reactor plant systems. He was awarded the Navy Achievement Medal and the designation of Master Training Specialist before moving on to build products and solutions that would empower clients to solve their long-standing IT problems. Joel graduated Magna Cum Laude from Louisiana Tech University with a Bachelor’s Degree in Electrical Engineering. He remains a recognized expert in network solutions and large-scale IT systems deployment. He can be reached at [email protected].