|
But there are legitimate uses for the Internet, so access can't be discounted entirely. Researchers use the Internet for their work and communicating with colleagues. Internet access is practically a requirement for them. Administrators also need access to Internet databases. Internet access can be restricted at either the remote-access server or at the WAN router or firewall. That's the easy part. Determining legitimate Internet needs are more difficult. Access to trading partners, commercial services and research sites should be examined and considered in your overall sizing considerations. The issue of Web access centers on the fact that the bandwidth requirements are relatively small, but users tend to read while connected, leaving an otherwise idle port connected.
To counter users' inclination to remain connected while the line is idle, you can set time limits for connections. For example, the research hospital is setting a limit of one hour per connection. If more time is needed, users will have the option of reconnecting. The time restriction makes users compete for ports, but also increases the number of users who can get connections. However, some users will complain that they need more time. To address those needs, client tools that let users transfer data and work offline and server tools, such as caching Web proxies, may be used. These speed the transfer of data to users and lower the time needed to stay online.
Of course, organizational and user remote-access needs may change during the course of the day, especially if you're supporting users across multiple time zones. Time-of-day restrictions can segment your user population and lower the overall user-to-port ratio across a time period. For example, if a portion of your user base is made up of telecommuter
s, these users can be granted sole access during the early morning and afternoon hours to gather e-mail and files for work at home. Field personnel, such as sales or service, may need to connect just once a day to transfer data to the home office. Determining time-of-day restrictions presents difficulties, however, because you have to allow enough time for a large portion of the designated users--if not all of them--to connect to get their data and disconnect.
Data Needs
Determining data needs is straightforward and the process is similar to determining bandwidth requirements on a LAN. The limiting factors in sizing remote access is the number of ports available and the time users spend online--not the speed of the connection. This seems counterintuitive, but users generally don't use remote access efficiently, and forcing efficiency on users is like trying to bathe a cat--it's often noisy and very frustrating.
Regardless of access method, ISDN or modem, users tend to stay online interactively,
rather than get data and drop off the connection. This is especially true with analog connections when reconnecting can take up to 45 seconds. Of course, if you limit users' network access to specific applications, you can tailor your remote-access and application support for use over slow links. If, on the other hand, you don't control access, determining utilization becomes difficult because of the variance in time online that is possible.
Not all users are alike, but they can generally be categorized along departmental lines and by user type. Power users tend to be sophisticated users who work efficiently over remote access, sending and receiving data and getting offline fast. They don't want to be hobbled by slow data transfers. Thus, power users tend to spend less time connected, but connect more often. New users tend to connect less often, but spend more time connected. They aren't familiar enough with remote access to work efficiently and tend to read and respond to e-mail online.
In fact, typic
al users with limited remote-access experience tend to follow a pattern (see "Typical User Profile" at left). They spend little time online in the beginning. Why spend the time learning how to get connected via modem when bringing a disk home from work is just as easy? As users become more comfortable and experienced with remote access, they spend more time exploring other resources. New possibilities like synchronizing files, remote application and intranet technologies open up for them. Eventually, users learn to work more efficiently and spend less time online. This is especially true for roaming users, where staying connected ties them to the location with a phone cord. Offline mail and database synchronization let them work while traveling, reducing remote-access applications to primarily data transfer. Of course, not all remote users follow this pattern, but the majority do.
Users in different departments have different bandwidth requirements. For example, the hospital administrators may need to tran
sfer e-mail, update GroupWise and download a 500-KB document. Researchers, on the other hand, may need to grab large image files in addition to the tasks just mentioned. Here, in the extreme cases, the speed of the connection becomes important. An ISDN connection with STAC compression can transfer data as much as 12 times faster than a V.34 modem connection at 31.2 Kbps.
Data throughput becomes a concern in the long picture--forecasting requirements into the next year or more. Applications, and their data requirements, are expanding at an alarming rate. Although advancements in compression, both hardware and software, over the past few years have helped to offset the increasing data needs, compression can only go so far.
|
REPORTS
ANALYTICS
Follow 
