Storage Area Networks (SANs) are becoming as much a part of business as computers are to homes. Every business realizes the potential and need for it and is taking steps to secure the new technology. Businesses recognize the SANs benefits of convenience, ease of use, reliability, and cost reduction. But as with all new technologies, SANs also pose many challenges.

This ReferencePoint discusses SAN technology from the basics to the
steps involved in designing, building, and implementing a SAN.

The first step in the process is defining what makes up a SAN. Once you understand the basics, designing the system will be an easier procedure.

The best definition of a SAN by its function is a storage system designed to provide large amounts of storage capacity to a business or enterprise. It is fast, reliable, and upgradeable. Usually, it is identified by one or more servers connected to a number of storage devices as shown in figure 1.

Figure 1 shows the layout of a Storage Area Network (SAN)

Implementing a SAN will provide mass storage capabilities to any LAN system. In SAN systems, the servers are indirectly connected to the storage units. Connections are made by means of hubs, switches, and bridges.

When designing a SAN it is always good to know its proper structure, this way, during the development process, you are confronted with proposals for different methods to build a SAN, you will know what is valid and what is not.

For example, attaching a storage medium is directly to a LAN, without the proper hubs, bridges, or switches. This is not a SAN. Figure 2 shows this scenario.

Figure 2 shows two storage devices connected to a LAN, proving it is not a SAN system.

Sharing storage directly between servers is also not considered a SAN. Figure 3 shows an example of this arrangement.

Figure 3 shows a storage device connected to two servers, showing it is not a SAN.

Although, you cannot connect a storage device to a server, or LAN, and call it a SAN, you can connect this device to a LAN. The only problem will be a noticeable performance drop.

The SAN puts storage on the LAN so it is convenient and does not pose problems for the user.

A SAN takes the data and provides it to all servers. This will lesson the dependence on any one server. Further, it is easy to add more storage, which allows more data to be accessible to all servers.

Companies use SANs because SANs supply capacity for LANs. Many large companies have such huge amounts of data to work with; they need a massive amount of storage space to house it. They also require speed and reliability. SANs provide it all.

In any company, the need for data to be available 24/7 is vital. If a large business was down for even one hour, the cost would be astronomical.

Cost is another factor. Every business has to keep spending to a minimum when possible. One of the advantages of a SAN is it is easy to maintain and expand.

Although, these factors are important, one factor that cannot be overlooked is a SAN is its ability to help applications run better. SANs do this by storing data used often by applications in one place, this saves time and doesn't slow the network down, waiting for the application to search for the saved data. The data is in one central location that is easy and convenient to access.

When we create a SAN, we are provided with the opportunity to discover the appropriateness of having a SAN, but unfortunately, we also discover the problems that can come about as well. We will see this as each part or component is assembled.

We have to understand hardware reality. No system, no part, no component is infallible. There are bound to be sections that won't work correctly, while other sections will. This is just the way it is. It's just like building a PC. After the computer is complete, and you turn it on, there's likely to be a problem somewhere. Nine times out of ten, this happens. When it does, we deal with it.

That is what you will face as you build a SAN. You will face frustration at times, while attempting to create a system, something complete out of several intricate components that if stood alone, would be worthless.

Before a SAN is built it has to sit on the drawing board of a designer, planning every stage, every connection, every detail just right, to make sure to come out with a system that will work correctly, reliably, and give long service to the user.

Before that happens, some terms need to be covered. Without a knowledge and understanding of these terms, you will not be able to figure out what is being said and will be lost in the confusion.

Look at figure 4. Notice the components and how they are laid out.

Figure 4 shows a lay out of a typical SAN and the terminology used to describe each part of the system.

The top of figure 4 shows workstations. They are connected to the LAN. The next row is the servers, also connected to the LAN.

Row three is the SAN, made up of two hubs. In a typical SAN, there could also be switches and bridges connected here.

From this design, the SAN is the connection between servers and storage devices. The medium used for connecting all parts of the system is fiber optic cable.

In the last row are the storage devices. The typical storage device consist of disk arrays like RAID (Redundant Array of Inexpensive Disk) and JBOD (Just a Bunch of Disks). RAID and JBOD are commonly a collection of SCSI type disks.

The other type of storage device companies use is the tape library. Considering that SANs are used as a massive storage device, the tape library would require massive tape backup.

Before you put down on paper the idea or conception of what type of SAN you need for your business, you must consider various aspects of what is required to operate a SAN.

When designing a SAN there are several points that have to be considered. They are:

	Applications
	Support for protocols
	Cables
	Device distances
	Number of devices
	Using legacy or SCSI devices
	Growth potential
	Traffic
	Reassignment of sections or departments
	Redundancy
	Disaster recovery
	How is SNMP going to be managed
	Cost factor

Applications

Every application has its own requirements for distance, usability, and functionality on the network.

A SAN has to work with several of types of applications, and at the same time. Make sure you take into consideration bandwidth, power, speed, compatibility, and usability of all applications in relation to themselves and each other.

Support for protocols

Every network uses protocols to transmit signals across it from point-to-point and across the Internet. When designing a SAN, you have to make sure your system will support the standard protocols in use.

Cables

As far as cabling is concerned, both copper and fiber optic cable is used. Due to high cost of copper, fiber optic cable is the preferred brand for many LANs or SANs.

There are two types of copper wiring - passive and active. A passive copper cable offers no signal balancing and can only go 15 meters. But an active copper cable does have signal balancing and can go up to 30 meters.

However, fiber optic cable is flexible and can run up to 500 meters. Since it is immune to EMI, fiber optic cable is a favorite for LANs and SANs.

Device distances

Calculating the distance between devices is critical for the smooth transfer of signals between devices. If a device is too far apart, signal loss will occur. Depending on whether you use copper or fiber optic, you must be sure of the distance to insure proper signal strength. In some situations, 10 kilometers is sufficient, but that is not always the case. Most ports and connections will be close to 500 meters.

Number of devices

If you are using Fibre Channel, you can use millions of devices. Most switches have 8 to 16 ports available, but can support more if cascaded.

When calculating devices, make sure to include internal disk arrays. They have separate disks connected on their own connector to a local port.

The best way to be accurate with your device count is to calculate all devices you will need and add about five to 10 more to the list to be sure. This way you will have enough room on your network for that unanticipated growth.

Using legacy (SCSI) devices

It is important that you are able to support SCSI devices on your SAN. Disk arrays like RAID are SCSI made, so be sure to include them.

Growth potential

When designing your system you have to make room for growth. You will expand. Purchase the right amount of hubs, bridges, and/or switches that can support immediate growth. Always lean toward anticipated or unanticipated upgrades.

Traffic

You need to know how many users will be on the system. This information is vital. By having this number you can buy servers with the proper setup to handle the traffic and connections.

Departmental splits

When designing your system make sure to take into account departmental splits. You will have to support any hardware the department uses or buys should they split. You also have to be certain that any equipment from each department can interconnect.

Protection from failures

When designing your SAN, always allow for cable and hardware failures. Provide for extra cables and other removable parts, including hard drives and power supplies. This way if one fails you will have a replacement available.

Disaster recovery

Plan for a disaster, it will happen. Disaster recovery can be expensive. Make sure when designing your SAN, you account for disasters by providing backup systems including tape drives or some other storage medium. This way in case of network failure, you have everything stored safely, which would make recovery a smooth process.

How is SNMP going to be managed

SNMP is automatically provided for all LAN devices, since it is standard protocol that is used on all types of networks. There are many different levels of support. Make sure you choose the right level so you will have the full support you need.

Cost factor

Cost is an important design requirement. Since parts are not cheap, you need to provide a cost-effective alternative to your SAN design system. You have to be practical and use commonsense when sizing up your cost figures against what you actually need for your system. You have to keep your business requirements and your customers in mine when making your decisions.

Many IT vendors are creating components that are compatible with every hardware and software available.

Many companies that do not have SAN technology have found it to be a challenge to establish one, considering not all equipment and software has met proper standards. When they do decide where to purchase the equipment, the right vendor is critical.

If the vendor they work with cannot deliver the SAN hardware, another vendor will be required, which will necessitate establishing new relationships and building trust between company and vendor.

For example, Company A has a relationship with a certain vendor they have been doing business with for a long time. This vendor does carry SAN products. Now Company A is ready to face any challenge that comes their way while they produce top quality products.

When implementing a SAN, there are many areas of concern: 1) applications, 2) data, 3) resources, 4) the network, 5) each component, 6) educating the user, 7) storage space, and 8) backups.

Applications

Implementing a SAN requires installation of the right type of software or applications. Software is the most complex and unstable area to work with. Software has always been a critical factor for a SAN. The hardware exists, but having the software designed to take advantage of it is another matter. If you were to install an application that was not compatible with certain types of hardware, your system will not work. Time and money will be wasted. Also is the issue of large-scale use of the application.

Data

What about data. Remember that data is the lifeblood of your business. Without it, you would not have a business. Your SAN has to be designed so when it is implemented, you can work with and store your data knowing it is accessible and secure.

Resources

Then there are your resources. As with any computer system, when programs are running, printers are churning out paper, or any device is working, resources are being used. When any device is engaged, communication exist between the component or resource and the computer. This is where part of your system has to work with the other.

The Network

Your network is probably the most important part of the system. If your network is not working with your SAN system, then you will have major problems. Every part has to communicate with the other. Signals have to travel from section to section, to reach its intended target. If there is any break, incompatibility, insufficient power, or instability anywhere along the path, you will have major issues. You may spend more money than you planned.

Map and monitor every section of the SAN as it is installed. Check every component to make sure no problems exist before you proceed. This way you save time and money.

Each component

The one thing you have to be concerned about when buying components for your SAN is that each component has to interact with one another in a positive way. If you buy a hub for example, and that hub does not have the speed your bridge may have that you bought from another vendor, you will have speed issues. When you buy products from different vendors, you have to make sure every part has the same rating, speed, and can work well with your system.

Educating the user

Educating the user is the most important part of the SAN process. If a user is told what a SAN is, how it works, the importance of it, and how to use it that will help ensure a possible trouble-free environment. The user will be able to handle the equipment in the most appropriate way possible.

Also of importance is the training of support personnel. What better way to implement SAN technology than to train a technician, who understands the structure and layout of a SAN. If the technician has the required knowledge and understanding of the principles involved, putting a SAN together will not be a burden.

There are many technicians that know enough about computers to assemble and repair one with no problem. However, when it comes down to working with a network, they do not know what to do or how to work with it. That is why training is so vital to an organization. The technician has to be kept up-to-date on everything from the newest products on the market, to the topology of a LAN and SAN.

Storage Space

In a typical LAN system, when the server gets full, that's it. You have no more space to work with. That means you will need to spend money to increase storage in the server. With a SAN, you eliminate this problem because all storage is kept separately from the servers.

Backups

In typical LAN setups, a backup system usually involves a tape storage system of some sort. This tape system is directly connected to the server. This cost money and it slows the network. But, with a SAN, the SAN controls that function to make backing up data independent of the network. This way the SAN handles the movement of data from the server to the tape system.

A SAN helps save on data storage. On a typical LAN, data is only accessible to one or two servers. With a SAN, data can be accessible to many servers.

Now that you have designed your system it is time to build one.

When building anything, it needs to be evaluated piece by piece so you can see how it looks, how it is built, and how each part relates to the other.

When it comes to the structure of a SAN there are a number of points to take into consideration.

Building a SAN requires paying close attention to detail as to what equipment is necessary and what is not. When building a SAN, you need to keep focused on whether you want to expand it in the future. You need to be concerned about reliability of the parts or components, the speed at which those components can work, etc.

Consider mixing technologies. Just as you can mix technologies in a LAN, you can do the same in a SAN. This can include Fiber Distributed Data Interface (FDDI), Asynchronous Transfer Mode (ATM), and Fibre Channel. SANs also allow the use of different protocols like Transmission Control Protocol/Internet Protocol (TCP/IP) and different versions of SCSI.

You can use different kinds of storage like disks, tape, RAID, and different operating systems including NT, Unix, and XP.

When it comes to building a SAN, you do not have to separate it from the network, because the SAN operates like a subnet, thereby, working behind the scenes as it were.

A SAN has the ability to monitor and alert you for faults or failures of any type. This way if there should be a component failure you will know. In order to keep the signal moving, multiple paths are common.

Multiple paths to a device are important to a SAN. This way if the cable should develop kinks in it, the signal, or data can be re-routed to another cable around that problem area. The one aspect to a SAN is its concept of having no single point of failure. If any component should fail, there will be another component to take its place.

No matter how well planned your SAN may be, if it does not have the proper amount of power to operate it, your system will not work. There are sure to be power outages. The best power supply many companies use is the UPS (uninterruptible Power Supply).

Now you know the advantages of having, and building a SAN, and what requirements need to be met, you are going to be guided step-by-step through the building process of how a SAN is constructed.

You will start with a server and a storage unit. This is not a SAN, but you have to start somewhere. See figure 5 for a representation of this.

Figure 5 shows a simple server with a storage unit attached.

The speed between devices is fast and convenient but not reliable. The server, cable, and/or storage unit can fail.

To avoid any type of failures, you will need to add another connector to it. See figure 6.

Now there is an additional cable attached. If either cable fails, there is still a path between the server and storage unit to transfer data.

However, there is still a problem. The server can fail. The best recourse here is to add another server. See figure 7.

Figure 7 shows a storage device attached to two servers by means of fibre optic cable.

Now there are two servers connected to the storage unit. There is still a problem. Since there is one cable attached to each server instead of two, there is a possibility that either server can fail. If that happens, the whole network fails.

The system has to be designed so if one server goes down, the other one continues to operate. Another factor to consider is the disk storage unit. With two cables attached to each server means two ports on the unit are used. That is using unnecessary resources.

Adding another cable between servers will help. Look at figure 8 to see where the cable was attached.

Figure 8 shows a storage device attached to two servers, but now there is an additional cable attached between servers.

If one server fails, the other one will continue to work.

However, with this type of setup, you are limited to upgrading and expanding your system. You need to add more functionality to it so upgrading is possible. Right now, the system is configured such that when data leaves the disk storage unit, it will go to both servers at once, thereby slowing the network down.

Look at figure 9 and see the improvement of the current setup.

Figure 9 shows a storage unit attached to two servers by means of a hub.

As you can see, a hub was added. Now, when the data is sent out over the network, and if it is met for one server, that is the server that will get the data, not the other one. This speeds up the network. It also allows for upgrading later.

With the current setup, each server connects to the hub on two paths. Even with this configuration, there is still a problem. Everything is connected to one hub. If this hub were to fail, the network would stop working.

Adding a hub will remedy this situation. Look at figure 10 to see how this works.

Figure 10 shows a storage unit attached to two servers by means of two hubs.

You now have what is known as a fully-expandable SAN.

Each server connects to each hub on two paths. There is no single point or place where failure can occur. And if failure did occur somewhere, the network would continue to work.

From here you can add more servers. You can increase the distance between devices as long as you don't go beyond 500 meters. If you were to go beyond 500 meters, you would start losing data strength and power. You can add more hubs. Some companies have been known to cascade hubs to shorten the distance between them. This helps to keep the signal strength at maximum potential throughtout the network.

Taking the steps to build a SAN is not that difficult. No matter what type of equipment you have, it can be constructed to form a SAN, as long as the requirements for building a SAN is met.

For example, you have a SCSI system you want to add to the structure. Since most SAN equipment supports legacy or SCSI equipment, this would not be a problem.

Adding SCSI equipment, including RAID, or special devices like a tape library, is not difficult.

Building a SAN with SCSI equipment is identical to building one with standard equipment. The procedure is the same. The same considerations in regards cost, speed, reliability, and maintenance are taken into account.

The only difference is you are working with SCSI equipment and cables along with fibre optic cables instead of all standard equipment. Look at figure 11.

In this configuration, there is one server, one bridge, and a RAID system. Not well protected. If the server or bridge failed the network would stop functioning.

There is a better way of configuring this system to provide better stability and reliability.

Look at figure 12.

Figure 12 shows a RAID system attached to two servers by means of two bridges, using SCSI and fibre optic cable.

Now there are two servers connected to two bridges to a RAID system. This kind of configuration is better because if any one component should fail, the network would still work.

So far you have been working with SCSI devices. Thanks to the flexibility and adaptability of a SAN system, adding standard devices is not difficult.

Look at figure 13.

Figure 13 shows a SAN system with a standard tape library device being added to the SCSI system.

The only problem here is the tape system is not a SCSI device. It does allow a SCSI cable to connect to it, but in order for a signal to go from the tape library to the servers, the signal has to be translated into the SCSI format. This takes time and slows the network. To combat this situation, an adjustment to the configuration needs to be made.

Look at figure 14.

Figure 14 shows a SCSI tape library connected to a server through its own bridge.

This is a better system. A SCSI Library System is connected through a bridge to a server. But again, this there is a problem here. If the bridge that connects the library were to fail, there would be no tape backup system in place. In turn this would affect the whole network.

The best solution would be to change the configuration to something like that in figure 15.

Figure 15 shows the SCSI tape library connected to the SAN by way of cascaded bridges.

This resolves the problem. Now the network is stable, reliable, and error free. If any part of the network should fail, the network will still work.

But if you look at the setup, there are many types of equipment involved. You have four servers, three bridges, two hubs, along with a RAID system and tape library. Remember, everything cost money. Equipment is not cheap. Building a SAN involves more than creating stability, and reliability. It also involves building one at the cheapest cost possible, without sacrificing quality.

Since building a SAN does incur cost, you have to use the cheapest parts, and also the fewest parts, without sacrificing quality.

Look at figure 16, here is a way to keep cost down and create a very reliable network, that is cost effective, and will do the job.

Figure 16 shows a more simplified setup.

This setup is more economical but efficient in that it will provide you with the power, speed, reliability, and functionality you require, at less cost to you.

When you are building your SAN, always keep it simple. When there are less parts to consider, less maintenance to worry about, less cost to consider, and lower amount of failures, there is less worry and frustration.

Do your best to remember future upgrades and expansion, while taking advantage of the flexibility of your system. The final call is for you to make. Keep it simple or build it your way. You will have to decide that.

Before you even think about implementing your SAN, you had better test it to make sure it works correctly. If the system fails to work, all the time and money in building it will be for naught. Because, now you will have to spend more money to remedy the problem.

You need to test your SAN on a physical level. Every single part or component needs to be checked for reliability and functionality. If there are any problems associated with that part or component, you need to know now, not after it is in place.

As you test each unit, keep a log listing the part or component you are testing and the result. Do this with every single piece or unit until you have completed testing.

You have to make sure that every device works according to the manufacturer's guidelines. Test the hardware. Make sure all devices are responding to commands. Make sure signals are getting through. Take measurements between cable ends. Check for output.

Run your applications. Make sure each software program will load and run on the server, from each workstation, etc. Make sure your storage units can successfully read, save, and transfer data.

Why is all this testing necessary. Because, you do want a system that is reliable and dependable.

When testing your system, there will be issues around your applications. When this situation occurs, you will have to stop and evaluate the problems. See what caused the problems, and take the necessary steps to correct them. When you are ready to install the SAN you will have to communicate with your people and with your equipment. Expose any possible future threats to the system. Keep track of any inconsistencies you find. This way, if anything does happen in the future, they can be dealt with.

One way to start with a SAN is to do it in stages. Install a stage and let your IT personnel get a feel for its use and responsiveness. As the staff gains experience with each section, add more components gradually until all units are in place. This way you make a smooth transition to the equipment.

The best way to begin an installation is to start small. Master that technology first than progress from there.

Start by installing a server and hub to a storage unit. Make sure all connections are good and test the system. Use it like this for a short time. When your staff is adjusted to its use, increase the system by adding another server, hub, and perhaps more storage.

The best way to begin any network installation is to proportion out the hubs. Make sure there are connections to each one both ways. Every path or connection needs to have multiple lines to work with in case of failure.

One way to do this is shown in figure 17.

From here you keep adding devices until you decide upgrading is no longer an option.

In the near future, with advances in technology being the way it is, newer devices will result in better production and support.

Actually, there really is no need to wait for the future to deploy new devices in a SAN. There are such devices on the market now. There have been many changes to devices that have become part of the SAN standard. Vendors are providing switching systems instead of hubs to offer more reliability and stability.

Look at figure 19 for an example of this.

Figure 19 shows a switching system in place.

It doesn't stop here. With the capability of the storage switch, many builders and installers are adding storage hubs to the picture. This gives even more reliability to the existing SAN system. Check figure 20 to see an example of this.

Figure 20 shows storage hubs in place.

As technology changes, so does the equipment. New devices will mean better control and better performance. When this occurs, take advantage of the new technology and embrace it in your system. You and your users will benefit.

After your SAN is in place and it is ready to use, now will come more challenges.

Remember that parts are infallible. You may have tested the system and found everything working now, but that does not mean it will stay that way. That is why the major players invest time and money to insure continued access and uptime with their system. These businesses know how important having a successful running network is to their organization. That cannot be stressed enough.

For instance, you are right in the middle of a large project that you want to save. Just as you are about to save it, you get an "error reading from network" message. Since you depended on your SAN to save all your data, you do not have a way to save it locally.

That is why companies hire technicians and network troubleshooters to handle system problems before they occur. They need to ensure the safety of the data at all times.

There may be a time when they may have to install additional servers, or storage units. On the other hand, maybe, they may have to replace or install additional hubs or switches.

If certain protocols change in the industry, your system will need to be re-configured to accept and deal with that new protocol.

Sometimes, maintenance is required 24/7 for a major computer company, whose very output of products is vital to their organization.

By having maintenance in place future problems can be avoided. You would not need to bring your technician in from home to fix a problem that could have been prevented.

Of course, not every situation or emergency can be avoided. Accidents happen. Parts fail. By continued monitoring of the system, you can at least cut down on the amount of network failures that could occur.

SAN architecture is designed to allow all storage devices to have access to all servers on the network. To install a SAN so that it works any other way is defeating the purpose of what a SAN was design to do.

The best way to get full functional use from a SAN, is this:

Connect all storage devices or units to the network

as presupposed in your design documentation.

	Allow for expansion in your system. You have the capability to hold enormous amounts of data at terabyte speeds. Take advantage of it.
	Do your best to arrange your network so you can free up your LAN from tedious operations like data backup. Let your SAN do that.

Many vendors try to tell you how to handle a system, knowing you bought the parts from them. They will suggest this or that. The best way to work with a San is to run applications and use it as you normally would. Force the system to perform a task to see if it will handle the task appropriately, or it will respond with an error.

If you need to obtain information concerning a certain feature of a part you bought, go to the Internet. You have so much information at your fingertips. Let it be your knowledge base. After all, you can get just as much or maybe more information that way than you could from your vendor.

One way to assure your system works properly at all times, is by placing a copy of your configuration files in one place so you know where they are. Always keep backups of your system files. If you make changes to a configuration, record it and save the file as a backup. Protect yourself at all times. Backup. Backup. Backup, and backup often.

Another way to monitor the SAN is to know who has access and control that access. Don't let anyone have access to the system that is not authorized to do so. This will lessen the amount of traffic and security issues that may result from illegal or unauthorized use.

When it comes to sharing resources, there are always problems. Either someone does not have permission to use a file and is using it, or someone changed something, resulting in a loss of sharing capabilities.

One of the ways to resolve sharing problems is by making sure that each workstation is set up according to your quidelines.

Most businesses work with that framework in mind. They have a list of employees that work for the company and in what department. They know what applications or software each department needs to use. So you just go in and set up each workstation to ensure that the user at that workstation can only access the file, application, or data that he/she needs and nothing else.

Not only does this cut down on sharing problems but it also lessens the amount of people that is using a certain resource for any extended period of time.

When it comes to implementing a SAN, many requirements that you investigated earlier during the designing and building stages are now ready to put in place. You have a full functioning system that is ready to work for you.

You must realize of course, in having a SAN, you have a powerful force at your disposal. You have speed, accessibility, distance, availability, and sharing, right at your fingertips. You have exactly what you need for a smooth operating system and one for time to come.

If you have an accounting system and you want to take advantage of the speed of your system, use it. The same goes with your database program. Or your spreadsheet. Do not inhibit yourself from experimenting with your programs. Put the system to the test. Burn it in for a while.

The biggest drawback is concern for compatibility. Not every program will run on certain machines. Not every application is designed for that faster bus speed, or the switching hub. Not every part of your system is going to communicate with each other in a consistent manner.

That is why, when during your designing stage, you anticipated this type of situation, and you made adequate compensation for necessary man power to perform troubleshooting, when problems do occur.

Another consideration would be to enlist the services of certain skilled people. Hire programmers, data entry clerks, and computer operators, to put your system to good use.

Of course, that is not the end of it. Just because you have speed now, it does not mean you cannot upgrade to more speed later. When newer bus speeds become apparent, take hold of it and implement it in your system. Do your best to upgrade when parts and equipment become available.

Fibre Channel is the standard protocol for SANs today. Right now Fibre Channel runs at 1 Gbps. However, could there possibly be a protocol that can run even faster.

There is also the issue of availability. Networks have been known to run slow or lose connections. This creates problems when trying to access certain applications. With a SAN, you do not need to worry about not having constant access to your applications. A SAN is built to accommodate you this way. All you have to do is set up the application the way you want it to work and let the network do its job.

Right now, you may be concerned about sharing and sharing violations. This is a cause for alarm. But if you put in place a plan to equip your network administrator, or whoever is assigned to monitor and configure your network, with the information he/she needs to make sure the tape unit, or whatever medium you are using, is properly shared, so all servers and resources have access to that data, you will have no difficulties.

The same goes with backups. As long as you have the tape unit or device you are using as a backup system, in place and installed correctly, the backups will occur when they are suppose to, on schedule everyday.

There are two ways to implement your SAN network for best optimum performance.

One way is to consolidate or centralize your servers and storage units.

Figure 21 best describes this layout.

Figure 21 shows the storage unit is centralized.

If, on the other hand, you have critical information that needs access to 24/7, then you will need a system like that shown in figure 22.

Figure 22 shows a storage unit attached to a special type of application server by means of fibre channel switches.

Any one of these types of generic improvements can alter your SAN system to give it better performance, more power, and control. Plus, you will have the reassurance to know that your data is safe and reliable at all times.

A SAN is powerful, reliable, and stable. It is the best alternative to mass storage available. While a LAN has limited storage facilities, a SAN has large amounts to work from. If mass storage were attached to a LAN, network performance would suffer. It would suffer to such an extent that the network could stop working.

A SAN, on the other hand, is not directly attached to a LAN. Rather it is a part of the LAN, like a subnet, being connected by means of hubs, switches, and/or bridges. This set up will allow the LAN to function normally, without the excess travel that will slow it down.

A SAN enables a company to manage large amounts of data effectively than a LAN could. This makes it a sought after commodity, despite its intricacies and complexities.

Despite the widespread use of SANs, the technology is still in its early stages and continues to progress. The SAN makes up a complex array of hardware and software components. It works off few standards, many devices, and continually changing applications.

Management software that a SAN uses, works with all other types of software. It is hard to engage, considering all the resources it has to use and control. Once it is set and running, it works in a near flawless manner.

When designing and building a SAN, every step has to be taken into consideration. You cannot take short cuts. One missed step, your system will not functioning properly.

Prevent that now by being alert, careful, and assertive in your efforts to secure a reliable and dependable system.

Count the cost. Do research. Find out what is involved before you even start. Draw out a diagram listing everything you want in your system.

Designing a system or network takes careful planning, thought, and time. Take each step one-by-one. Write it down. Calculate. Do this with each section, device, and component, until you have exactly what you want and need.

When you are ready, build it. Build it and test it. The best way is to test it as you build it. If this is not feasible, build it first. After the network is finished, test it. When you are thoroughly convinced that everything is operational, implement it into your company plans and allow others to use it. Monitor it for a while. At least, by taking these steps, you will have succeeded in creating a network that will provide you years of service that will benefit you and secure for you trust in the safety of your data.