Storage Area Networks For Dummies
Book image
Explore Book Buy On Amazon

Everyone likes new toys, and a storage area networks (SANs) are pretty cool. This, however, shouldn't be the main reason to install a SAN. Find out if you should make the move or not.

Reasons to use a SAN

Rather than buying large servers with tons of internal disk space, see if a SAN would make sense for your company.

You want better disk utilization

The number one benefit from installing a SAN is better disk utilization. When all your storage is tied together through a centralized storage network, you gain the ability to manage everything as a single entity. This gives you the ability to slice up the central pool of storage resources at the network level and assign that storage more intelligently to the servers that need it. The approach to disk management without a SAN is to buy tons of disks and stick them in huge expensive servers so that you can grow into them. All the disk space that's not currently being used is wasted until you need that space.

You need a good DR solution for multiple applications

When you have many critical servers in your data center running applications that simply can't go down and need to be recovered quickly if disaster strikes, then you definitely need a SAN. Although the upfront costs for implementing a SAN-based disaster recovery solution are high, the benefits can be realized in just a single hour if a disaster happens. The cost of downtime is critical in many organizations. A SAN-based disaster recovery (DR) solution is like having a good insurance policy on your business.

You need better availability for your applications

Those big bad storage arrays in storage networks are built from the ground up to never go down. They use technology like phone home (as in the movie ET), where the storage array itself makes a call to the manufacturer to report that one of its parts might be going bad soon. This is very cool stuff.

Equipment that you find in a SAN is expensive, but it's built like a tank. Applications fail because of data becoming corrupted by things going wrong with the disks that those applications use. The storage arrays in a SAN use very good data protection algorithms to make sure that your data stays consistent. The best thing that can happen to you is that you forget that your SAN is even there. It just works, like the dial tone on your phone.

Backup is taking too long

Decreasing the time needed to back up huge amounts of data is also one of the major benefits of installing a SAN. Technology is available in today's storage devices that enable making hardware-based, exact duplicates of your data almost instantly. The duplicates can be used as either the backup of your data or as a source for backing up that data to a tape library connected to your SAN. Many of the backup software vendors have the capability to use this functionality of the storage arrays to create these image copies for you.

Reasons not to use a SAN

A SAN may make sense for your company only if you have the right staff, budget, and business requirements to support it. If your storage requirements fall into one of the following scenarios, then a SAN might not be right for you.

You have only a few inexpensive servers

You really see the benefits of a SAN when you have a bunch of servers to hook up to the storage network. A main benefit of using a SAN comes from the centralized management of storage assets. Although companies with dozens or hundreds of servers benefit from this centralized management, managing just a few individual servers is easy enough to do without installing an expensive SAN. You may gain speed and reliability benefits from a SAN for just a few servers, but if you cannot justify the cost of the SAN against the needs of your applications, you should not install one. You need to weigh the costs against the benefits to see whether a SAN investment makes sense for you. The equipment that you need to buy to create an effective SAN can be pricey.

You want to save your company money this year

The payback period of a SAN sometimes might be less than 12 months, but you need to look at a lot of factors to see whether this could be the case for you. If your SAN vendor has the capability to give you a return on investment (ROI) report based on the solution that you're considering, then have the vendor prepare such a report.

Many of the cost benefits of a SAN come from the reduction of operational expenses over time. Things such as less floor space, power consumption, reduced manpower requirements, and application uptime are some of the benefits. But if you're not growing your business that fast and you're not planning any layoffs soon because of the reduced requirements, then the payback time will be longer. Depending on your environment, you can calculate the payback period between 12 to 18 months for a large SAN. Just installing the switches and running the cables can cost a lot of money. The storage itself is usually the least expensive part of the SAN.

You need a disaster recovery solution for a single application

If your recovery requirements are for a single application, then using a software-based solution might be more cost effective that a hardware-based solution that uses a SAN. Plenty of software solutions are available that can be used to create an effective host-based disaster recovery solution. And this software can cost a heck of a lot less than putting in a SAN just to use the storage array to move your data between locations. Software-based recovery solutions run on the servers that contain the data to be replicated, using your corporate network as the transport medium to mirror the data from the server at your primary location to the server at your recovery site.

You need to replicate your data for disaster recovery but can't afford fast WAN connections

This is a common problem. The most expensive part of data replication solutions is the money that you have to pay to lease network connections between your buildings. For those of you without very deep pockets, this can be an issue. One way to accomplish remote copy of data is to use the CTAM (Chevy Truck Access Method). That's right, Chevy truck, although you could instead use a Pontiac or even a van to accomplish the same task. Here's the deal: You back up your servers to tape, throw the tapes in the back of a truck, and transport the tapes to your remote site. In the event of a disaster, you restore the tapes to the new servers that you just bought and get back to work. It might take a week or two for you to get back to business as usual when using this method, but hey, you get what you pay for.

About This Article

This article can be found in the category: