Internet Service Providers & Bandwith

Internet Service Providers

There have been many advances in internet service provider (ISP) services over the last several years in terms of bandwidth, availability, and reliability. It is true that they do not offer a full service level agreement if you are crossing multiple Internet providers between a remote site and your data center, but it has become reliable enough to use in many situations.
From the chart you developed to define the prioritization of infrastructure components, identify sites that do not perform critical business functions and investigate moving these locations to an ISP connection. The costs will decrease and usually bandwidth will increase to provide quicker response times to the business at this location. For locations with two wide area network (WAN) connections installed in order to provide a more reliable service, consider moving one of these locations to an ISP connection. This will lower the overall cost for supporting each location. Also, set up routing so the business traffic will route over the existing WAN circuit and utility traffic will route over the ISP connection. Then you can route the business traffic over the ISP connection in the event of a failure of the WAN connection. However, since utility traffic is nonessential, you do not need to route it over the WAN connection in the event of a failure with the ISP connection. You will save the expense of upgrading the WAN connection to support all the utility traffic in the event of an ISP connection failure. Examples of utility traffic are software distribution, antivirus updates, and FTPs. You are also able to route Internet bound web traffic over the ISP connection. Routing this traffic over the ISP connection relieves congestion on the primary WAN link and provides better service for business traffic. Since the ISP connection provides higher bandwidth for lower costs, using this connection to support video applications is a more cost-effective solution than using the WAN service.


Users never have enough bandwidth. Complaints of slow response time will come in on a regular basis unless you have oversupplied locations with bandwidth, which is not advisable if you are trying to control costs. Ensure that locations are actually using the bandwidth provided for business traffic before making an upgrade. Bring in tools to provide the ability to analyze traffic at a location. Determine if users at this location are performing nonbusiness bandwidth-consuming functions, such as downloading MP3 songs or listening to streaming audio. If this is the case, have this traffic stopped by talking to management at the location and identifying the staff performing these functions. Taking the time to do this will help foster an environment where employees use computer resources only for business functions.
You can implement many products and services to stop employees from using the Internet for certain nonbusiness traffic. This is a preferable solution as it stops the downloading of MP3 songs, listening to streaming audio, etc. It saves personnel time in tracking down violators and prevents slower response times at a business site. However, this is not a cheap option. If funds are limited, spend the money on analysis tools because you need to know traffic patterns either way.
As a rule, you want to provide enough bandwidth capacity to support business requirements at a location. There may be sites or vendor price points that make oversubscription of bandwidth a good selection. If there are business-critical sites for your organization, then managing bandwidth too closely may cause a significant impact to the business that is not worth the savings. Spending more at these locations to provide extra capacity is worth the cost. The other decision in bandwidth management is looking at vendors' price points. Some WAN vendors have started providing full T1 Multiprotocol Label Switching (MPLS) connections at a lower price point than higher-speed fractional T1 connections.


Obtain Inventory Information

A very important and often overlooked step (Step 2) was to obtain the information and inventory.
This step is critical when beginning to look for cost reductions in the technical infrastructure area. Without an accurate inventory, you will make cost savings decisions on hunches or perceptions of what is in place and how it contributes to costs. It is important to know what your services cost. You also must understand what infrastructure components are required to support typical applications and what components you share across services. As you will see later in this chapter, it will also be critical to have an accurate inventory to ensure vendors are billing correctly for services.
Some companies may have a fully populated Information Technology Infrastructure Library (ITIL)-based Configuration Management Data Base (CMDB) or other type of database in place with all relationships among the IT items established. If not, do not attempt to create such a CMDB as an initial step like this is a large and complex effort. Initially concentrate on establishing enough information in a database or spreadsheet to manage resources and audit vendors' invoices. Then, implement processes to ensure the information stays up to date as you make changes to the infrastructure. Later, you can develop a CMDB or similar type database.

Develop Prioritization Scheme

Step 3 of the cost reduction to analyze the costs. Develop a prioritization scheme for the IT infrastructure. A prioritization scheme helps identify where you should look for infrastructure savings and where you want to avoid cutting back. Tier the infrastructure components to indicate the priority to the corporation. Figure 1 provides an example of an infrastructure prioritization framework. The exact prioritization scheme would vary in the number of categories and tiers for each company. Tier 1 applications would be areas in which the business stops if they become unavailable. Examples of infrastructure areas that would merit a tier one rating are components that support applications critical to the business, such as the search engine would be for Google or infrastructure components for airport passenger check-in programs would be for an airline. The bottom tier would be equipment that supports systems that you consider noncritical to business operations. Examples of noncritical systems would be an accounts payable system, help desk software, or a sales projection system. What may be noncritical for one company may be critical for another. While these are important systems for the organization, the business will continue to function while these systems are down. The importance of each tier and column vary by company, as company systems and locations will vary.

Entire company
Major site
Minor site
Tier 1 applications
Data center, key applications
Customer facing location
Tier 2 applications
Tier 3 applications
Tier 4 applications
Supply ordering
Small remote site

Figure 1: Infrastructure prioritization framework
Making infrastructure changes to reduce costs is risky. One approach that lowers risk, if you have the time, is to start cost-cutting activities on lower priority infrastructure areas. You will then gain experience in these areas with minimum or no impact to critical business areas first. As you gain experience, and staff skills are developed, you and upper management will have the confidence to attack cost savings in more critical infrastructure components.