Network server aids with data storage

In my last column, we talked about transferring images from a camera to a computer.

In my last column, we talked about transferring images from a camera to a computer. This is done via some type of removable media to which the camera stores the images. This media can be removed from the camera, and the image files can be downloaded from it to a computer. The most common type of storage media are compact flash drives. Other types include SMART media and media sticks. These storage media come in different memory sizes and read/write speeds. The memory size refers to how many images, or number of files, that can be stored on a card at any one time. The more memory a card has, the more images it will store before the card is full and the images have to be downloaded. You then need to delete the images from the card to reuse it. Storage media act like reusable film. After images are stored on the compact flash card, they are transferred to your computer via a device, such as a “card reader.” This is a small device that is connected to your computer, usually via a USB port. The compact flash card is inserted into the reader and the image files are then transferred. The USB port that the card reader plugs into on a computer comes in different speeds, such as USB I and USB II. Most computers today have USB II ports, which transfer the images much faster than USB I ports.

Once these steps are taken, what happens next? This depends on a computer’s setup. You may have a single computer on which you store images, or a network in which several computers are connected to a main computer for data storage. The “server” is the main computer to which other computers found in your ops are connected. My seven-year-old office server needs to be updated. So I called on the services of Tim Thousand, a local network systems programmer, who shared with me the following information about different network setups and costs.

Basically, there are two different approaches to setting up a Windows-based network. The first is a peer-to-peer network. The second is a client-server, or domain-based network. Peer-to-peer networks can be implemented in offices with up to 10 computers. But, based on my experiences, they are best suited for offices with no more than two or three computers. Because of the administrative overhead often associated with peer-to-peer networks, a domain-structured network is often a better choice.

Generally speaking, dental offices that utilize digital imaging and radiography are better served with a client-server structure. With this, one computer in the office must be set up as a dedicated server, and must be running some type of network operating system, such as Windows 2003. Generally, a server’s hardware configuration is dictated by a combination of elements, including data capacity, availability, expandability, and budget constraints.

When working with customers to configure a viable server solution, the question of availability often arises. While there is no guarantee of 100 percent uptime, there are ways to minimize potential downtime. One of the best ways to ensure the availability of a server is to make sure it has certain redundant components, such as a RAID array, and dual - or even triple - power supplies. When compared to downtime costs, the nominal expense of implementing redundancy in a server is trivial.

RAID, which stands for Redundant Array of Independent Disks, is a late-1970s IBM technology that is essential in the quest for high availability. At a minimum, a server should be configured with a RAID 1 array. RAID 1 arrays can be either hardware- or software-based, and utilize two hard drives. The second hard drive is an exact duplicate of the primary. A RAID 5 array utilizes three or more hard drives, and spreads the data across all drives in the array. Both arrays provide redundancy; however, RAID 5 provides significantly more storage capacity, better throughput, and reduced or no downtime in case of a failure. While software-based RAID 1 arrays can be implemented in most Windows server operating systems, hardware-based RAID is a much better solution for performance reasons. You should keep in mind that with RAID 1, every byte of data that is written to the primary hard drive also has to be written to the secondary drive. This can result in an overall reduction in performance.

Obviously, there are many things to consider when contemplating a new server or network installation. Unless you are comfortable with network capacity planning and configuring a server, I would suggest you consider consulting with an IT person early in the planning stages - certainly before making any purchases. A well-devised plan of action will reduce frustration and expenses. When all is done, I think you will be much happier with the results of your investment of time and money.

Dr. Tony Soileau is a general dentist from Lafayette, La. He has taught digital photography at the Pacific Aesthetic Continuum in San Francisco, the Institute of Oral Art & Design in Tampa, Fla., and the Esthetic Epitome in Charlotte, N.C. Soileau is currently a co-director for the genR8TNext digital photography program. Contact Soileau at (337) 234-3551, or at tony@tonysoileau.com.

More in Money