|only for RuBoard - do not distribute or recompile|
There are SAN implementations that appear specialized, because of the severe processing demands of some applications. Those applications need extreme capacity and speed.
The paradox is that most general applications would appreciate a little more in the way of capacity and speed, so today s special SAN implementation has a way of becoming tomorrow s norm. At some point there will be no such thing as a specialized SAN implementation.
For example, a data warehouse is a storage- intense application, traditionally limited to larger enterprises with the resources to assemble outrageously large collections of data and go exploring. However, with SAN capabilities, medium and smaller enterprises have the potential for making a data warehouse part of their operation.
The same is true with motion picture and video production. Today, it s a specialized application, and a natural place for a SAN. Tomorrow, stored video may be employed widely in nonentertainment applications.
A data warehouse is a large collection of enterprise data, managed by a relational database running on a high-performance server. The underlying assumption is that enterprises collect a lot of information, which, if subjected to complex analysis, could help in the decision-making process.
Although the warehouse may be small in size at the start, it s assumed that the company will add a substantial amount of data to it daily. Because the warehouse always grows in size , storage for a data warehouse should be highly scalable. This is a natural application for a SAN, as it s easy to add storage devices. You can increase the number of devices, the capacity of devices, or both.
Data mining tools are used to research data in the data warehouse to find patterns, classifications, and associations. Data mining can help retail businesses see purchase patterns, payment patterns, and the relationships among various purchased items.
Data mining has been quite useful in the retail industry to analyze consumer buying patterns and form marketing programs to take advantage of the analysis results. For instance, data mining can find patterns in your data to answer questions like:
What item purchased in a given transaction triggers the purchase of additional related items?
How do purchasing patterns change with store location?
What items tend to be purchased using credit cards, cash, or check?
How would the typical customer likely to purchase these items be described?
Did the same customer purchase related items at another time?
Once the buying patterns have been discovered , the retailer can use this information to tailor a marketing strategy that appeals to each type of buyer, thereby maximizing profits or minimizing costs through optimizing inventory management.
Supermarkets are experimenting with clubs. When you join the club, you are issued a barcoded card to be presented when you make purchases. The number is tied to your name , address, e-mail, and other information, and presumably every purchase you make can be linked to you. Store discount coupons are gone, so the only way you can save on purchases is by presenting your club card.
Whether this data collection is useful is up to the supermarket chains. However, it suggests that a lot of data is being collected daily. From that we reason that supermarket chains have big data warehouses, and would suspect that a SAN of ever-increasing capacity is central to the picture.
The retail industry is not the only industry to take advantage of data mining. Other uses for data mining include: risk assessment and portfolio management in the finance industry; fraud detection and policy assessments in health insurance; and optimization, scheduling, and process control in manufacturing.
The most efficient way to edit video and motion pictures is in a digital editing suite. This requires large amounts of storage, and a SAN is the ideal way to accomplish the task.
When footage is digitized and stored on the SAN, multiple editors can work on the same project without breaking it into parts and downloading the video to different workstations.
Transoft Networks of Santa Barbara, CA (recently acquired by Hewlett-Packard), says:
FibreNet FC can support multiple streams of compressed or uncompressed video from a network with shared storage. This makes it an ideal media server for a number of non-linear video, audio, animation, and graphics editing applications. FibreNet FC allows multiple artists to work on the same source footage, without being interrupted by sneakernet, searching for shuttles, or duplicating footage. Valuable editing time can be spent on visual creativity instead of media management; footage can be viewed and mastered at ideal resolutions . The combination of speed and ability to collaborate on projects results in clear cost savings and increase in productivity.
What you want in video editing is simultaneous access to the same media files from multiple editing stations . This is superior to the time and effort it takes to place copies of material on multiple local disks. It makes sense that a SAN would be used for video applications.
|only for RuBoard - do not distribute or recompile|