Analyzing Organizational Needs


Understanding the needs of a business or other organization is a fundamental step in creating a well-designed network. In this section, we’ll take a look at information flow—recognizing where data originates in your network and how it should be disseminated to the users and customers who require it. Next, we’ll discuss the importance of understanding an organization’s management structure and how you can use that information to design appropriate network services. We’ll also discuss some common priorities for an organization’s management group, as well as its more task- and project-oriented users. These range from factors such as performance and availability that affect an entire network, to more specific services and applications such as e-mail, file sharing, and audio/video services. All of these issues should be taken into account to ensure the overall success of your network design.

Information Flow Factors

If the “Information Age” moniker is to be believed, it only stands to reason that access to a company’s information needs to be a top priority of any network design. This means that all necessary personnel need on-demand access to their critical data in order to understand how their company’s profits and losses are occurring, to call up a customer’s account information at a moment’s notice, and to collate information from multiple sources to allow for effective decision making. The most successful organizations are those whose front-line employees have instant access to the information they need, rather than waiting for managers or central “gatekeepers” to disseminate scheduled or ad hoc reports.

Understanding information flow requires you to determine where your users are located, what data they need to access, when they need it, and how they need to access it to best perform their jobs—whether that job is running a quarterly sales report or a high-school fundraiser. Providing appropriate information flow can involve physical considerations such as sufficient bandwidth allocation, along with logical controls within the computer operating system. Remote and traveling users introduce their own unique challenges, because you will likely need to provide data access from varied and ever-changing locations around the globe. Whatever steps end up being necessary for your own network implementation, information flow can make or break a modern organization.

Management Model and Organizational Structure

Understanding a company’s organizational structure is imperative in designing a network to meet its needs. You should begin by becoming familiar with the high-level divisions within an enterprise and how they related to one another. Large divisions usually have their own organizational structure, and they might be broken into several smaller departments or workgroups. For example, the Division of Finance might encompass separate Payroll, Accounts Payable, and Collections departments. Most companies have developed an organizational chart to provide a graphical illustration of this overall structure.

Once you have an understanding of the organizational structure, you can take a closer look at the individual departments themselves. Does the management structure of your organization have many levels, with Assistant Directors reporting to Directors, who report to Senior Directors, and so forth? (You can see an example of this sort of structure in Figure 1.8.) Or is the management model more flat in design, with a single manager taking responsibility for an entire department? This information will greatly benefit you when designing network functions such as user groups and AD OUs, as well as when you are determining appropriate delegation of network management responsibilities.

click to expand
Figure 1.8: A Departmental Organizational Chart

Centralization versus Decentralization

Once you’ve determined the organizational structure of your client or company, you should also recognize whether that structure is a centralized or decentralized one. Some companies adhere to a strictly hierarchical reporting structure in which the organizational chart resembles a family tree, with each sublevel reporting to a subsequently higher level and a single individual or group at the top of the hierarchy. In an AD environment, this type of structure lends itself to a system of nested OUs like the ones shown in Figure 1.9.

click to expand
Figure 1.9: A Centralized Organizational Structure

Other organizational structures allow for greater autonomy within their business units, where various departments or project teams can function more independently. You might create an AD environment consisting of multiple domains, allowing each to maintain its own security requirements. And you can certainly mix-and-match these models to meet the unique requirements of your organization, as illustrated in Figure 1.10.

click to expand
Figure 1.10: A Combination of Centralization and Decentralization

Your network design should also consider the Information Technology (IT) management structure of the organization. A company with a decentralized management structure can still handle network management centrally and vice versa. The transitive trust relationships built into Windows Server 2003 can allow centralized management of a multidomain or multiforest environment, or for tasks to be split among departmental IT administrators. The IT management structure of your organization can help you to decide how tasks such as user and group management should be structured and delegated.

Management Priorities

The management perspective of network design can be more conceptual, or high level, than the end-user priorities that we’ll discuss in the next section. Rather than focusing on specific tasks and applications, a company’s management structure should focus on design attributes that are common to and can benefit the entire organization, not just specific departments or workers. These include network availability, security, scalability, performance, and cost. When designing a network for an enterprise organization, be sure to address as many of these concerns as possible.

Availability/Fault Tolerance

As people and companies have become more reliant on computer technology to function and perform personal and business tasks, network designers have needed to contend with increasing expectations for “always on” availability. A sales manager traveling in Europe or Asia will not be pleased to find that although she can access her e-mail client, the data on the server itself is available only during business hours in the Eastern Standard time zone, or that a hardware failure will prevent her from accessing sales figures for eight hours while the server is being repaired. To avoid such difficulties, business-critical applications such as database and e-mail servers should be placed on systems that are designed for high availability whenever possible. This rationale applies even more to retail Web sites (e-commerce sites) and other Web-based businesses. Planning for high availability and fault tolerance will help you to minimize the downtime experienced by your end users and customers. Windows Server 2003 offers two separate but related clustering technologies—server clustering and Network Load Balancing—that can provide the high availability required by most enterprises.

Fault tolerance specifically refers to the ability of a piece of hardware or software to withstand the failure of a key component. This can be implemented at the hardware level using redundant power supplies or a Redundant Array of Inexpensive Disks (RAID) hard drive array. Advanced fault-tolerance technologies will even allow an administrator to replace individual components within a server without powering down the server. Clustering provides the ultimate in fault tolerance: completely redundant systems.

Test Day Tip

The ability to replace hardware on the fly, without powering down or rebooting the server, is referred to as hot-swapping.

Security

To create an effective network design, you must perform a juggling act between providing easy access to data for those who require it and, at the same time, protecting the data against unauthorized or illicit access. Accessibility and security are always at opposite ends of a continuum—more of one results in less of the other. Establishing an information security strategy is critical in ensuring that your network design is prepared to address security concerns when, not if, they arise.

A well-developed network security policy is as much a business concern as a technological one; consequently, you should involve key decision-makers from all parts of an organization, including Risk Management, Legal, Human Resources, and so on. Your security policy will provide a common baseline of security procedures based on your company’s security requirements.

When addressing security concerns within your network design, your three primary concerns are the confidentiality, integrity, and availability of your data. These three security objectives answer the following key questions:

  • Who has access to your data?

  • Has the data been corrupted or altered in any way?

  • Will your users be able to access their data when they need to?

All technologies and practices within information security will ultimately address one or more of these key concepts.

Scalability

When planning a network design, scalability refers to how well a service or application can grow to meet client performance demands that will inevitably increase over time. It can refer to increasing system resources such as processors, memory, disk drives, and network adapters to an existing piece of hardware, or being able to seamlessly replace existing hardware with more powerful equipment. It can also refer to adding new servers to meet increased demands.

A scalable network is one that can expand over time to address network growth and improve (or at least maintain) client response time. Server clustering, mentioned earlier as a technology to ensure availability, can also be used to address scalability issues by allowing you to add nodes to a cluster when your network encounters a period of growth.

Performance

Network performance—good or bad—is one of the most noticeable outcomes of any network design plan. Performance has a direct impact on all aspects of end-user productivity and customer satisfaction. If your e-commerce Web servers are overloaded, you will probably lose customers who abandon their shopping carts out of impatience. This translates directly into lost customers and lost income for your company. Likewise, providing adequate performance on a corporate LAN will allow your corporate employees to focus less on waiting for their workstations to reboot and more on productivity, thus creating revenue for the company.

Cost

There is an old joke among software developers that goes something like this: “Cheap, fast, right… pick two.” Monetary considerations can make or break a network design. An improperly budgeted network installation can create any number of long-term difficulties and end up costing even more money to correct problems that cropped up during the initial installation. Almost everyone embraces the goal of cutting costs, but remember that it is almost always less expensive to do something right the first time than it is to correct or upgrade an insufficient installation.

User Priorities

No matter what network infrastructure your organization uses, you can be certain that it won’t be deployed in a vacuum. Whether your users are internal employees of your corporation or external customers paying for the services that your company provides, your network installation must provide for their needs if it is to be cost-effective and successful. You must create an environment that provides for the current needs of your users, as well as allowing room for future growth and changing requirements. We’ll describe some of the more common network services in use today: e-mail and other communications, scheduling and task management, project collaboration, data storage and retrieval, Internet research, application services, print services, and graphics/video/audio services. (Of course, a complete list of network services is limited only by the imaginations of your customers, clients, and users.)

Electronic Communications

Electronic communication, specifically e-mail, has become the de facto means of communication in the modern business world. Whether a company manages its own e-mail storage, using a technology like Lotus Notes or Microsoft Exchange, or outsources its e-mail to an external Internet Service Provider (ISP), modern computer users have come to expect a great deal from their e-mail service in terms of performance and availability. The outage of an e-mail server is now perceived to be just as disruptive as the loss of telephone service. In designing e-mail services for your network, you should make allowances for high performance and availability to meet the expectations of your network users.

You can provide high availability and performance for your e-mail services by making sure that you’ve allocated enough server resources to support all of your current clients, as well as planning for the growth of your user base. As with most other network services, fault tolerance can be achieved through the use of redundant hardware within an individual server, like redundant power supplies and NICs, as well as RAID arrays for your hard drives. Also, you can use server clustering to create two or more physical e-mail servers that your clients will see as one logical server; if one physical node of the cluster fails, the other will take over, usually without your clients noticing more than a few seconds’ outage.

Another common issue with e-mail servers relates more to how e-mail is used within your organization. Unsolicited commercial e-mail, commonly referred to as spam, can clog the inboxes of your client workstations, decreasing productivity as users sift through pages of junk mail looking for relevant messages. This can also lead to sexual harassment questions if the spam includes messages with adult content or graphics. As an e-mail administrator, you can implement spam-filtering centrally at the server level or install client-level tools for your users to configure according to their own tastes. Spam-filtering uses a number of different technologies, including blocking e-mail from lists of known spammers and filtering messages based on keywords such as “get rich.”

Along with deciding how to address unsolicited e-mail, you should create a policy describing how e-mail and other computing resources can and cannot be used within your environment. You’ll often hear this referred to as an Acceptable Use Policy (AUP). An AUP essentially provides a road map for your users to make decisions about what is and is not appropriate to do with their office computers.

Some organizations have a strict zero-tolerance policy, where there can be no personal use of any company resource, including e-mail. More often, though, you’ll see a phrase in an AUP that allows for “reasonable personal use” of computing resources. The purely financial argument might say, “If we can keep each of our 80,000 employees from spending one minute a day sending a personal e-mail, then we’ve saved the company X number of dollars.” But at the same time, you need to consider the potential for added productivity for the account manager who is happier being able to send a quick note to his daughter who lives halfway across the country. You need to carefully consider what type of policy will best suit your organization.

Scheduling/Task Management

Fully featured e-mail clients such as Microsoft Outlook and Lotus Notes can extend e-mail functionality to include a wide range of calendaring and task-management functions. Users can manage appointments for anything from small project teams to entire departments and offices. This can improve the efficiency of users’ time management by providing automatic meeting and resource scheduling, including notifications of appointments and time conflicts. Supervisors can manage schedules for an entire group of individuals, tracking meeting attendance, scheduled appointments, and vacations. Administrative assistants can even create, move, and delete appointments on their managers’ behalf.

Centralized task management can also assist managers or team leaders in directing the projects under their supervision. Managers can assign specific tasks and track their progress and completion date from a single location. As you can see in Figure 1.11, you can keep copies of tasks you’ve assigned on your personal task list, as well as receive status reports when an assigned task has been marked as complete. When integrated with e-mail and calendaring functionality, task-management functions can greatly streamline work processes for project teams and departments of any size.

click to expand
Figure 1.11: Assigning Tasks in Microsoft Outlook 2002

Along with using network resources to schedule and assign tasks for users and employees, you’ll also want to allow for scheduling of computer-based tasks. This can include scheduling recurring events such as nightly backups of user data, or the ability to run tasks on an as-needed basis to create user accounts, reset a forgotten password, and the like. Windows Server 2003 has a graphical Task Scheduler interface that allows you to schedule tasks on a daily, weekly, or custom basis. You can also integrate many Windows commands and utilities into scripted batch files or custom applications. For example, the administrator for a university department might want to automate the process of creating user accounts for incoming freshmen every year, rather than spending time creating each individual account manually. Well-developed scheduling and task-management functions will allow the administrator to accomplish this in an efficient and timesaving manner.

Project Collaboration

No matter the size of an organization, sharing information within an organization and with outside parties is vital to increasing productivity and creativity on projects of all kinds. In this case, a “project” can refer to any situation where people need to share information, from a formal business research project to a high school marching band. Project collaboration technologies must provide an intuitive and easy-to-use means of sharing documents, deadlines, and other key pieces of information among people working from multiple locations.

Packages such as Microsoft SharePoint offer users the ability to organize and access information through a Web browser or another familiar Microsoft Office environment. Figure 1.12 (from the SharePoint homepage on www.microsoft.com) illustrates the kind of information that a project collaboration technology can gather at a user’s fingertips.

click to expand
Figure 1.12: A Microsoft SharePoint Project Collaboration Web Page

Microsoft SharePoint comes in two varieties: SharePoint Team Services, and the more full-featured SharePoint Portal Server. SharePoint Team Services is actually integrated directly into the Windows Server 2003 operating system, and provides the ability for small or ad hoc project teams to share information. The full-blown SharePoint Portal Server is designed to work in an enterprise installation, allowing users to share and manage documents among multiple servers. The key differences between the two versions of SharePoint are listed in Table 1.1.

Table 1.1: Comparing SharePoint Portal Server and SharePoint Team Services

Feature

Team Services

Portal Server

Core function

Ad hoc team collaboration

Enterprise portal and search

Search capabilities

Documents within team Web site and subsites

Across multiple servers and data types

Discussion and notifications

Discussions, notifications, and surveys

Discussions and notifications

Customization

Browser-based, Microsoft FrontPage 2002, and SDK

Web Parts and SDK

Document management options

Publishing

Check-in and check-out, versioning, routing, and publishing

Client applications

Browser, Microsoft Office XP, and FrontPage 2002

Browser, Microsoft Windows Explorer, Office 2000, and Office XP

Security options

Customizable roles: Administrator, Advanced Author, Author, Contributor, and Browser

Administrator, Coordinator, Author, and Reader roles

Licensing requirements

One FrontPage 2002 server license, no separate client access license (CAL)

Server license and CALs

Data Storage and Retrieval

Providing a central location for users to store and access files is one of the oldest and most common uses for a network file server. This provides your users with the ability to access shared data within a department, an organization, or an enterprise. The Windows operating system has provided the means to share files and folders since the release of Windows 95. The Windows Server operating systems allow an administrator to add management, security, and scalability functions to their users’ ability to share information. When planning file services for your network, you should keep the following objectives in mind:

  • Simplify user access to files in a large organization, especially when those resources are located on multiple servers and shares. This can include the ability to retrieve data stored on multiple servers from a single access point.

  • Provide efficient data access for users accessing information from multiple locations. For example, if a sales manager in Chicago needs frequent access to reporting data from remote servers, he should be able to access that data without using an expensive leased line to do so.

  • You should be able to migrate data to various servers without affecting the way that users access that data. If you must visit each user’s workstation whenever you reconfigure a share or a server, it will greatly restrict the flexibility of your network infrastructure.

  • Minimize any delays that can occur when accessing a frequently used file or folder.

Windows Server 2003 has introduced new features (and improved on existing Windows Server functions) to improve file sharing services, including the following:

  • Volume Shadow Copy This allows network backups to take place while users are still accessing files and folders, increasing the availability of shared documents on the network.

  • Distributed File Service (DFS) Like its predecessor in Windows 2000, DFS allows you to take shared folders located on multiple physical servers and group them using a single namespace. With this feature, you can add or remove physical folders, drives, and even entire servers without affecting how your users access the resources they need. DFS can also be used to provide fault tolerance and load balancing for the file sharing services on your network.

  • NTFS permissions As in previous versions of Windows, file permissions prevent unauthorized access to the resources on your network. Windows Server 2003 also has continued support for file compression to save space used by infrequently accessed files on your hard drives.

  • Disk quotas As with Windows 2000, you can use the disk quota function of Windows Server 2003 to passively monitor or actively control disk usage on your file servers. Disk quotas can be enabled on a per-user basis on any of your server volumes. Properly implemented disk quotas will increase the availability of your file sharing services by preventing drive space from filling up without warning.

  • Removable storage Windows Server 2003 provides enhanced support for removable storage devices such as Zip drives, FireWire devices, and Universal Serial Bus (USB) storage devices.

  • Offline files Like Windows 2000, Windows Server 2003 will allow users to “check out” a network file and make changes to it on their local machine before the file is checked back into the network storage location. You can use this to improve performance, especially when accessing files over a WAN link or when you’re dealing with remote and traveling users who may need to work on network files while they are disconnected from the network.

  • Encrypted File System (EFS) This feature uses Public Key Infrastructure (PKI) certificates to digitally encrypt user files stored on a server or a local hard drive. This feature is largely unchanged from Windows 2000. It relies on users’ private keys to provide encryption for their stored files.

  • Indexing Service Another feature found in previous versions of Windows, the Indexing Service in Windows Server 2003 creates indexes of the contents of a server or workstation hard drive, as well as indexing the properties, or metadata, for various document files. This allows you to index files not only by name and location, but also by such properties as author, category, timestamp, and so on. You can create multiple indexes on a single machine to exert granular control over how the Indexing Service operates.

Internet Research

The Internet and the World Wide Web have created instant access to a wealth of information on countless topics for both personal and business use. This instant access to information has become crucial to the modern workplace, allowing access to a wide variety of resources (some of which we’ve already discussed), including e-mail, file transfers, business and personal collaboration, access to multimedia information, and more. The various resources available on the Internet allow users to research vast amounts of material and information.

Internet research differs greatly from traditional “paper” library research because information is not centrally catalogued in a single location, and it can move and change from day to day and week to week. Addresses of Internet sites can change and sometimes disappear altogether, creating a fluid and somewhat volatile environment. Information found on the Internet can also vary widely in terms of accuracy, credibility, and attention to detail, making it crucial to evaluate not only the information, but also the source of that information.

When designing a network for any setting, whether for corporate, educational, or personal use, it’s almost a given that you will be making some allowance for access to the Internet and the World Wide Web. Whether this access is universal to all users or restricted to only those who need to perform Internet research as part of their job functions, a good network design will provide secure access that will permit access to necessary resources while protecting the security of the internal network resources. You can accomplish this through the use of firewalls, proxy servers, and other hardware and software-based technologies.

Application Services

A well-designed network can allow you to host client applications from a central location, thus reducing deployment time and management costs as well as providing for centralized security. Centralized application management addresses some of the following user needs:

  • Central storage of application data so that users can access needed files from anywhere on the network

  • Centralized deployment, upgrading, and patching of applications without requiring user intervention (or sometimes even user knowledge)

  • Enabling offline access to network applications so that users can perform their tasks while disconnected from the network

Using a central application server such as Windows Terminal Services can enable you to deploy an application one time only to the server itself, rather than installing it on each user’s desktop. This can greatly improve both user and administrator efficiency in the case of custom applications that require frequent updates or applications that need to be deployed to users in geographically remote locations. Centralized application hosting can also increase security by maintaining sensitive data in a centralized location, rather than allowing it to traverse insecure network connections.

Note

When considering deployment costs of Terminal Services, remember to take into account Terminal Services licensing fees. Each client must have not only a CAL for the client operating system it is running, but also a Terminal Services license. See www.microsoft.com/windowsserver2003/techinfo/overview/termservlic.mspx for a white paper that discusses all the intricacies of the Terminal Services licensing structure in Windows Server 2003.

Print Services

Almost every environment relies on highly available printing services to produce all forms of paper output. A properly designed network will create shared printing resources across the network, allowing workstations to submit print jobs to printers that are attached to local servers or that are accessed across the Internet. Network operating systems such as Windows Server 2003 allow you to cluster network printers for high availability, and to automatically deploy printer drivers to clients of many different operating systems. Well-designed print services will also enable users to easily locate the printers they require. Administrators should be able to centrally manage and configure printers from any location.

The Windows printing architecture consists of two components:

  • Physical printer The printer is exactly what you think it is: the physical print device that is attached to a workstation or server’s parallel or USB port, or plugged directly into the network.

  • Logical print queue The print queue is the software piece that translates between the physical printer and the software application from which that the user is printing.

To improve printing efficiency, you can have a single print queue submit jobs to multiple printers, referred to as a printer pool. In the example shown in Figure 1.13, the Finance queue feeds to three separate printers. This is useful if a department produces a large amount of paper output, since you can manage the three physical printers as a single logical unit.

click to expand
Figure 1.13: Printer Pools and Prioritized Queues

Note

You should implement a printer pool only if the printers themselves are physically close to one another; otherwise, your users will be running from printer to printer looking for their output.

On the opposite end of the spectrum, you can have multiple logical print queues feed to a single physical printer in order to prioritize your users’ print jobs. You can assign a priority to a print queue between 1 and 99. Print jobs from higher-priority print queues will be processed before jobs submitted from lower-priority print queues.

You can also establish schedules in which printing to a certain queue may not be available at all. In Figure 1.13, there are three print queues set up for a single color laser printer. Let’s say that you want your graphics designers to have first priority when printing to this device, followed by any of your vice presidents. You can assign a priority of 99 to the Graphics Design print queue and a priority of 1 to the VP queue. Furthermore, you’ve recently discovered that some staff members have stayed after business hours to print personal material to the color laser printer. In order to keep from wasting the expensive color laser toner, you can establish a third Staff queue that can only be printed to between the hours of 9:00 a.m. to 5:00 p.m., Monday through Friday.

Graphics/Audio/Video Services

The increasing prevalence of high-speed Internet connectivity has created a market for high-quality streaming media services, ranging from streaming audio services offered by online radio stations to full-fledged audio/video streams used for training and conferences. Windows Server 2003 includes the latest version of Windows Media Services, allowing companies of any size to create and host powerful streaming media capabilities.

As a network manager, you need to be aware of the hardware, software, and network bandwidth considerations created by your organization’s current and potential future use of streaming media capabilities. Planning for the requirements of this technology is essential in creating an efficient network design.

Reviewing Legal and Regulatory Considerations

Depending on the business in which you are involved, your network design plan should address the legal issues associated with your industry, geographic location, and so on. Backup schedules and offsite data availability have become federally regulated matters, especially in the financial arena. Consult your Legal department during the design process, because like everything else in this venture, it’s certainly best to get it right the first time.

Don’t forget to include your client workstations when making allowances for legal and regulatory matters. For example, if your corporate data-retention policy calls for maintaining e-mail data for twelve months, but some users have copies of every item they’ve sent or received in the last five years, that fact could come back to haunt you in a legal proceeding.

Some fields of business are subject to very detailed governmental regulations regarding data security. For example, healthcare providers now fall under strict laws regarding electronic patient information since the Health Insurance Portability and Accountability Act (HIPAA) went into effect in 2003. Regardless of your field, if you work on government projects, your network might be required to meet specified security criteria.

Network communications can also subject your company to legal liability when employees misuse the network. For example, pornographic material on the company network can subject the company to charges of the “hostile workplace” definition of sexual harassment under Title VII of the federal Civil Rights Act of 1964 and various state laws. You should also consider intellectual property (copyright, trademark, and patent) laws in establishing your network policies.

Common factors that also need to be reviewed for legal compliance are any Service Level Agreements (SLAs) in place on your network. An SLA attempts to define the scope of a service provider’s responsibilities in maintaining applications or services on a network. This provider can be an external vendor to whom you’ve outsourced a critical service (your ISP, for example), or the SLA can be an internal document detailing the IT department’s duties in maintaining network availability. The following are the major components of an external SLA, using an ISP as a real-world example:

  • Scope of services This spells out exactly which service or application that an SLA is referring to and the level of responsibility that the internal IT department will have in maintaining this service versus the external vendor. This includes outlining the hardware, software, and resources that comprise the particular service, such as the modems, network connectivity equipment, ISP help desk, and engineering personnel in the case of an ISP.

  • Roles and responsibilities Your ISP should establish a coverage schedule so that at least one primary and one backup support avenue is available to report any service outages. You’ll also need to establish a system to escalate support calls if the scheduled support person is unavailable or cannot correct the problem. You can use this information to inform your users of the turnaround time they can anticipate in responding to and resolving any problems.

These are only a few of the legal considerations that are important in a corporate network environment. You should always include a legal advisor as a member of your network planning team.

Calculating TCO

“These upgrade proposals look interesting, but how will they impact our company’s TCO?” TCO is a calculation that was designed to assist consumers and corporate managers in assessing the direct and indirect costs and benefits associated with the implementation of new or upgraded computer technology. The purpose of TCO is to quantify the financial bottom line associated with a computer or technology purchase decision.

TCO calculations do not rely on a single formula. For example, a high-end computer will have a higher initial purchase price, but will probably incur fewer repair bills during its active life cycle. TCO is balanced against the benefits created by the technology purchase, such as improved user efficiency or perceived happiness with improved performance, in attempting to make a final purchase decision.

The first part of calculating TCO is relatively simple: What is the initial purchase price of the new technology? Include the cost of hardware, software licensing, networking equipment, installation charges, and so on. Don’t forget to factor in the necessary time to train your end users and IT staff in the use and administration of the new technology. Next, determine the ongoing costs for maintenance and support. These costs can include charges for vendor support, as well as in-house labor expended on interoperability issues with third-party and legacy software support. Try to estimate the total costs for the full anticipated life cycle of the proposed technology.

Determining the soft costs associated with a new technology is a bit more complicated. How much money will your company save by reducing the number of times your users are forced to reboot their computers each day? Conversely, how much money is lost when an account manager cannot access the order-entry application for 20 minutes, for an hour, and for a day? These costs are fairly difficult to quantify, but they can be critical when determining the total benefits afforded by a network upgrade. You can start investigating soft costs by talking to your users and reviewing TCO models from network analysts.

Your users can certainly tell you how much it aggravates them when their e-mail or order database is “running too slowly,” even if they can’t tell you what “too slowly” means in terms of actual response time. This can also point out performance bottlenecks that you may not have known about before. For example, a real estate lending office for a well-known bank shared a T1 line with the bank branch in the lobby of the office building. The real estate lenders encountered severe network performance degradation every day at around 4:30 p.m. Further investigation revealed that this time frame coincided with the bank tellers transmitting their daily totals to the bank’s main headquarters when the branch closed each day.

Preconfigured TCO models from organizations like the Gartner Group, IDC, or other independent network analysts can walk you step-by-step through plugging in various budget figures to arrive at the TCO of a specific technology, hardware, or software package. However, remember that these models are not set in stone, and they should be modified as needed to meet the specific needs of your organization. These models will rely more on actual calculations, such as dividing a help desk analyst’s salary by the number of support calls he or she is able to process in a day, or determining the “cost per e-mail message” of an e-mail server upgrade that increases the number of messages it can transmit in a day, week, or hour. You can then take these numbers and factor in the soft costs already mentioned. Using a combination of calculations and judgment calls will typically lead you to the most accurate assessment of TCO within your organization.

Planning for Growth

If there is one nearly universal truth to network design, it is that networks and their resource requirements always eventually grow. Your network design needs to account for not only what your users require today, but also what they are likely to require in the future. Even if your users or clients have not thought about future growth, you should provision your network design to accommodate for a reasonable increase in user population and bandwidth usage as time goes on.

One of the best ways to ensure that your design will support the future needs of your network is to implement well-known, standards-based technologies, rather than those that are proprietary or experimental. Expanding your network’s router core, for example, will be much simpler if the new hardware you purchase is compatible with the initial installation. (Otherwise, you might need to scrap the initial installation entirely and install all new hardware, greatly increasing your costs and overall headaches.) You should also deploy hardware and software in as consistent and well-documented a manner as possible, so that you can perform maintenance and upgrades as quickly as possible.

Examine the feasibility of allocating items like high-capacity network cabling and other infrastructure components at the initial installation of your network. For example, it may cost an extra 25 cents per foot to run 100MB Ethernet cable instead of 10MB Ethernet cable when you’re initially wiring your building, but it will cost significantly more if you find you need to rip out the cabling and redo it later.

In planning for network growth, you should again consult with your users, especially those in strategic planning and decision-making capacities. Although no one can accurately predict what will or will not happen to a company over months and years, these decision-makers will be able to give you some idea of the overall vision of the company. Are they hoping to expand dramatically through mergers and acquisitions? Or are they satisfied with their specific market niche and anticipate adding personnel and equipment in only smaller increments as production increases?

Finally, when considering desktop computers, laptops, and servers, keep in mind that most current hardware will come with a one- to three-year warranty, sometimes with an option to purchase an extended warranty at the time that you buy the equipment. It’s not necessarily true that your computer hardware will immediately break down the day after the warranty expires; however, the length of your warranty and/or service contract should factor into your projections regarding how often you plan to replace your equipment. For this reason, many organizations adopt a three- to four-year replacement cycle, budgeting sufficient funds to replace one-third or one-quarter of the installed computer base every year, or setting aside money to replace all of the equipment en masse when it reaches the end of its warranty cycle.




MCSE Planning and Maintaining a Windows Server 2003 Network Infrastructure. Exam 70-293 Study Guide and DVD Training System
MCSE Planning and Maintaining a Windows Server 2003 Network Infrastructure: Exam 70-293 Study Guide and DVD Training System
ISBN: 1931836930
EAN: 2147483647
Year: 2003
Pages: 173

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net