Local Area Network - multiple-user

Local Area Network ‚ multiple- user

Incorporating a network into the mix adds new choices to the deployment. Where do you locate the core data? Where do you locate the application? Where are the runtimes installed? How do you handle the network administrators drive mappings and standards for placing data on one drive and applications on another drive?

It seems obvious that a multi-user application would have the data on a central server if one is available. Even in a peer-to-peer network, you have to designate one of the machines to be the keeper of the database and data files. The difference between a peer-to-peer network and a true file server network is more an administration than technical issue. In a peer-to-peer network the biggest problem is the peer-to-peer ‚“server ‚½ is often someone ‚ s workstation. Again, this is not a technical problem, but it is administrative. The key is to inform the user not to shut off their workstation because other people in the office might be connected to it to use the data. The other issue is peer-to-peer ‚“servers ‚½ often are not strong machines, and are not running server operating systems, so performance as a ‚“server ‚½ is not as strong as a dedicated file server. Hardware is cheap so we recommend to our customers they buy a dedicated server for all LAN-based deployments.

A bigger question is where to load the executable. Do you want it on the server or on each workstation? The choice should be made during development so the application is flexible enough to find the data no matter where it is installed. Applications loaded on workstations start faster because they are not accessed over the network connection (wired or wireless). The issue with loading it on the workstation is upgrading and how to distribute the new version to each workstation. We address this issue later in the chapter under the section ‚“Executables. ‚½ If network performance is not an issue we recommend loading the executables on the server so it can be upgraded in one place and distribution to the workstations does not have to be maintained .

Most of the problems faced with a network install center on the network administrators. In some cases you might be the administrator so you have full control of the network. Some clients might have a separate vendor or in-house administrators for the hardware and network. They might have strict standards you need to follow as to where the data is installed so it is properly backed -up , properly mirrored on a separate drive or machine, and optimized for performance. The network admin might dictate the applications need to run on a certain network share because the users have access or restricted access based on corporate security standards. Security plays a much bigger role than with individual workstation applications. If the network admin does not grant the proper rights to a user your application might crash and you need to support the problem, a problem you didn ‚ t cause.

Drive mappings and shares play an important role and you need to account for this during deployments. Remain as flexible as possible and use DBF, INI, or Windows Registry entries to provide data-driven mappings. Hard coding settings like directories and drives can prove fatal to installations unless you have specific agreements with the client and the administrator in advance of the installation.