Sanjay Castelino Issue: Europe II 2011
Article no.: 15
Topic: Network capacity growth pains
Author: Sanjay Castelino
Title: VP of Product Marketing
Organisation: SolarWinds
PDF size: 303KB

About author

Sanjay Castelino is the VP of product marketing at SolarWinds, an IT management software provider. Mr Castelino leads the company’s end-to-end IT solutions for network, systems, application, storage and virtualization management initiatives; he is also responsible for all its product, technical and community marketing functions. Previously, Mr Castelino worked at NetStreams as VP of marketing and business development. Mr Castelino was also the VP of product marketing and management at Motive computer and electrical engineering.

Article abstract

With the rapid uptake of cloud computing, software as a service (SaaS), virtualised network environments and mobile wireless devices, IT administrators need to manage existing network capacity to support business-critical applications. Network performance monitoring solutions help organisations optimise their bandwidth usage of existing and ensure that critical traffic such as voice over Internet Protocol (VoIP), among others, are allocated more bandwidth than less demanding traffic. Today, monitoring an enterprise’s network performance is one of a network administrator’s most significant activities.

Full Article

Data growth is a big challenge for organisations of all sizes. According to Gartner’s report (November 2010), nearly half of the enterprises ranked data growth as one of their three biggest daily challenges. Network utilisation for data services continues to rocket globally, and organisations are increasingly under pressure to manage existing capacity in the most effective way possible. With volumes of data set to rise further – IDC predicted a rise in digital stored data from about 1.8 Zettabytes in 2011 to more than 7 ZBytes in 2014, businesses are faced with the challenge of managing the complexity of data traffic and explosion of growth. With the rapid uptake of cloud computing, software as a service (SaaS), virtualised network environments and mobile wireless devices, IT administrators are turning towards ways to proactively manage existing network capacity in order to support business-critical applications. Maximising existing infrastructure Historically, an answer to ever-increasing volumes of network data has been to buy a bigger Internet pipe. However, due to current economic conditions, financing of network upgrades at this scale is no longer feasible. Instead, organisations are looking at ways to manage growing capacity needs by understanding how critical applications are running over existing bandwidth, where performance of these applications can be improved, where the bottlenecks are and how existing bandwidth can be optimised. Applications are the life-blood of businesses and understanding exactly how critical applications perform, is essential to ensuring both company and customer-centric services are running effectively over the network. Network performance monitoring Network performance monitoring lets businesses view their capacity availability and performance issues within and around their network. Without the proper network monitoring solution to monitor virtualisation, wireless and wired network gear, businesses can be vulnerable to overextending their bandwidth. The right performance monitoring solution provides important insight into how applications are utilised, whether resources on the network are appropriately sized and they also provide data for forecasts and future requirement trends. By monitoring network capacity and usage, companies can determine how applications are running on the network and the ability to control and manage how network resources are allocated in order to accelerate business-critical applications. Today, capacity planning is growing in importance as enterprises turn to virtualisation to help control costs. In a large data centre, utilisation and application workloads can fluctuate dramatically over the course of a few weeks, days, or even within a single day. For example, a slightly improper virtualisation setup in which the company assigns the VMware ESX server to the incorrect network interface controller (NIC) would create an immediate bottleneck. Taking user error into account when implementing a virtualisation strategy makes it all the more important to have a comprehensive view of network performance and the ability to efficiently and quickly plan capacity requirements. How technology can help Business traffic such as voice over Internet Protocol (VoIP) and other critical services should receive precedence over less demanding types of traffic. Prioritising bandwidth allocation means that more bandwidth can be given to specific users and certain services, which require higher quality of service (QoS) to work reliably, are supported more effectively. Technologies such as NetFlow, IP SLA (Cisco’s IP SLA verifies network performance and identifies network problems) and class-based quality of service (CBQoS) allow network engineers to prioritize traffic for business critical applications over other traffic and help manage bandwidth more efficiently. Network performance soars or suffers based on what users are doing. NetFlow is a network protocol designed to monitor network traffic and help businesses identify and prioritise applications and business-critical content. It provides a useful perspective on where and how bandwidth is being consumed (both critical and non-critical). So employees using the corporate network for bandwidth-hogging applications like streaming YouTube and music or playing World of Warcraft® can be swiftly stopped. IP SLA allows companies to test connections and protocols over the wide area network (WAN). If a company is hosting a VoIP call between an office in New York and another in London, you’d want to ensure that the service’s performance does not degrade on the network. For instance, let’s say Site A is having no voice quality issues calling Site B. However, a user at Site C is reporting choppy sound and dropped calls when calling Site B. IP SLA would have helped identify this issue before the user called to complain or submitted a help desk ticket. Site A doesn’t see the same issue because it is taking a different routing path to Site B than is Site C. Using IP SLA to test network performance and connections to other sites on the WAN, users can verify if VoIP performance is adequate to ensure satisfactory quality. In addition to NetFlow and IP SLA, CBQoS can provide additional insight into an organisation’s application utilisation and help enforce traffic prioritisation. CBQoS is a resource reservation system; it reserves and prioritises resources based upon a set of defined polices that ensure business-critical traffic such as VoIP and video conferencing are of the highest quality. On networks with a variety of shared resources and limited bandwidth, CBQoS is one of the most effective and inexpensive ways to reduce network congestion. These technologies help uncover hidden bottlenecks on the network and between company sites by verifying how sites communicate with each other; they also prioritise the more critical applications and services running on the network. While NetFlow, IP SLA and CBQoS are key technologies to manage bandwidth, analyzing the data can be time-consuming and difficult to understand. To save time troubleshooting, IT administrators should invest in network monitoring solutions that automate this process. However even with a network performance management solution, prioritisation rules and policies aren’t perfect and can have unintended consequences. Think of it this way, you can setup a rule as soon as you have a snapshot of your traffic, but your traffic can change and the rules based upon the snapshot might not behave as expected given the new conditions. That is why monitoring an enterprise’s network performance is one of a network administrator’s most critical activities.