|only for RuBoard - do not distribute or recompile|
If you have a long enough lever, and a fulcrum upon which to place it, you can move the world. The Storage Area Network is the lever and the Internet is the fulcrum, and there s every reason to believe that you, as a SAN professional, will move the world in the not-too- distant future.
What would happen if people had faster access to more (and better) information? There have been some important moments when information in quantity was distributed widely:
The destruction of the library at Alexandria in 638 prompted scholars to save what books they could and flee. They spread the information throughout the Mediterranean and beyond.
The conquest of Granada in 1492 made the science and art of the Moorish empire available to European scholars.
In a move that did not require conquering a city or an empire, Gutenberg s development of moveable type and the printing press replicated ideas and made them available to many people.
When information in almost any field of human endeavor is distributed effectively, the world moves. In our time, the SAN is the basis for this process.
What applications can grow out of SAN technology? Let s look at a few.
High Definition Television (HDTV) exists, but it is not yet broadcast to people s homes . The formula in the video industry is that when the price of receivers drops to $1000, people will buy them.
Even after HDTV broadcasting becomes widespread, the interest will be in video on demand. That s a world where any video, on any subject, in any language, will be available to you when you want to see it, and the storage requirements are so intense that it s clearly a SAN application.
So how big is HDTV? We ve been advised by two video engineers that HDTV is a slippery subject to talk about, because a standard format and a standard for compression are both under debate.
But for the sake of argument, let s look at uncompressed HDTV, the highest quality available. Let s use the 1080i standard: 1080 lines, 1920 pixels, wide screen 16:9 aspect ratio, 30 frames per second, interlace scanned.
1920 pixels per line X 1080 lines X 30 frames per second X 2 bytes per pixel = 124,416,000 (124.416 MB) bytes for one second of HD video. That math suggests that a minute of HDTV requires about 7.46 GB and a 105-minute feature film requires about 783 GB. We also have a more conservative (and conflicting) calculation that yields about 1.169 TB for a feature film.
Either way, that s not an unmanageable amount of data for a SAN.
Transmitting HDTV into homes will require the video industry to agree on the compression sweet spot, the best acceptable quality produced by the highest rate of compression. Compression is essential, because there is currently no delivery medium for homes with a bandwidth to carry 124 MBps of uncompressed video. We have heard of compression formulas to send video at rates as low as 19.6 MBps but can t confirm that the quality is acceptable. Even so, 19.6 MBps cannot realistically be sent to homes yet.
However, a solution will be found, and as HDTV video on demand becomes a reality, it will surely alter all current thinking about broadcast TV and cable television.
What about the medium carrying uncompressed HDTV? It appears that 2 GB Fibre Channel could do it. So, imagine going to the movie theatre of the future at the mall. This cineplex has 100 small viewing rooms holding up to eight people each. There s a 300 TB SAN with 2 GB Fibre Channel connections to the rooms. What s playing? Just about whatever you want, because you have 300 movies to choose from. You pay one price for your whole party for the room and go in to select whatever movie you want. You watch a high-quality image on a large screen, and you can start it, pause it, and stop it whenever you want.
For you, you get a new kind of social, customized, entertainment experience. And the movie operator saves money in many ways: no scheduling problems; savings in advertising; no full houses requiring viewers to be turned away; no half-filled houses; no calculation of the number of screens to put a new feature on; additional revenues from old favorites; and no worries about booking a film for too short or too long a time period. Of course the films aren t films anymore, and they re downloaded from the distributor directly into the movie operator s SAN.
Bringing the Internet into every classroom in America sounded like a noble idea when it was first floated, but has evolved into a phrase used by politicians to show that they re interested in education and have some clue about technology. Having the Internet in classrooms is still a great idea, but it s only a start. Students also need content, and that content resides in well-crafted, formal, up-to-date textbooks .
When school districts share large SANs, there s a potential for putting every available textbook online. Those texts will be served to schools , and will be printable or accessible online.
That means the day of the expensive textbook is over, and school districts with tight budgets will welcome that. There s no budget-driven limitation on choice either. A teacher won t be limited to just one text or old texts; he or she will be able to choose from a variety of textbooks for the same class.
The reason secondary and college texts cost so much is simple: too complex a document bought by too few school districts or college students. With electronic books stored on a SAN and worldwide distribution, excellent textbooks will be within everyone s reach.
Even today, teachers can build custom video lessons by selecting stock still pictures and full motion video from video disks and assembling a sequence on VHS tape. With large SANs, teachers will have a much wider choice of such materials and will be able to easily create videotapes, CDs, and DVDs. Instead of a hodgepodge of specialized equipment, only a PC will be required.
The SAN has scalability, distance, and speed; so schools in different school districts can share learning data. Multiple school districts can share the costs of data on tap from central knowlege supply organizations.
The United States government has been called the world s largest publisher. In addition, it may have the world s largest collection of unintegrated data. That data falls into two grand categories: general information you can use, and specific information about you.
Even in the 21st century, access to government records and other information is inefficient and tedious . States and counties don t do a particularly better job. Government computers and storage are likely to be older and inadequate.
Imagine government with large SANs and well-managed data. We might see government that works better to serve the people.
It was too long in coming, but at last we can download tax forms and instructions electronically , and the Internal Revenue Service encourages electronic filing. That s fine, but wouldn t it be handy to look at your personal income tax data from past years , along with your Social Security data, Veterans Administration data, and Immigration and Naturalization Service data?
On the state and county level, it would be convenient to see your voter registration, automobile registration, assessor s information for your house, business license record, etc.
The data is there, but it s not integrated and it s not accessible. With a SAN and the Internet, that will all change. And when you can see and update your government data, you will exercise a greater level of control over it.
Unfortunately, information is power, and where there is power, there is sometimes abuse. If too much information about people is concentrated in one place, there s a serious risk of loss of privacy. What are the dangers of government collecting too much data about individuals in a well-integrated database?
If we assume a population in the United States of about 270 million people and dedicate 1 MB of storage for each person s data, it would only take 270 TB to create a dossier for every individual in the country. Would that be a good thing?
The biggest threat to getting quality health care in the United States is by no means the quality of the doctors , hospitals , or pharmaceuticals . It s the paperwork. In a recent National Public Radio interview, a doctor explained that in a 13- hour workday , she spent three to four hours filling out pieces of paper.
Every time you change doctors, you are asked to complete a new health history, providing the same information on new pieces of paper. Your chart is updated manually by the doctor during a visit. What s even more startling is that your electrocardiogram and X-rays are separate documents that can get separated from your chart and lost. Oh, and did we mention your lab results? Of course, records from your hospital stay are separate from records in your doctor s office. And getting a doctor s OK for a prescription refill requires a fax from the pharmacy to the doctor and another fax from the doctor to the pharmacy.
In the past, I (Barry) was an EDP auditor for a life and health insurance company, and I can assure you that insurance claims paperwork was and is inefficient.
To add to your concerns, none of the above information is of any use if, God forbid , you are in a serious traffic accident away from your home town. The Emergency Medical Technician (EMT) who treats you at the scene will have no clue as to your blood type, current medications, or allergic reaction to penicillin. This is not a very encouraging prospect.
By contrast, imagine your health information is stored in a SAN. This includes complete text and imaging data. In this scenario, almost every field of information on the documents described above is available for viewing and updating, over the Internet, by practitioners you authorize. All it takes is a robust SAN and the appropriate application software. Duplication of information and multiple pieces of paper are eliminated.
To return to the scene of the accident, with a SAN you have a much better chance of surviving. In the SAN-based medical scenario, the EMT takes your vital signs, identifies you, and sends a request for information through a wireless link from his or her helmet to the ambulance, and on to the local hospital. The request is transmitted to a national SAN-based medical data repository.
Your medical data comes to the emergency room doctors at the hospital and to the EMT who s trying to save you. The doctor directs the EMT about initial actions to take, views you through the video camera on the EMT s helmet, and reads your signs through transmitted signals. When you arrive at the emergency room, you get X-rays and a CAT scan. Minutes later, that data is appended to your electronic file. If you re in really bad shape, it will be shared electronically with remotely located trauma specialists.
Will all this cost a great deal more? Perhaps not. Mass storage is a very cost-effective way to consolidate information and eliminate unnecessary paper pushing. There may be enough savings to extend health care to many who do not currently have it available to them.
Clearly, it takes a lot of computing power and a lot of mass storage to make gains in some branches of science. The SAN will permit this to happen.
As mentioned at the beginning of this chapter, sequencing the human genome has advanced considerably, using supercomputers and a lot of stored data ”over 80 TB and growing. We would expect that any gene research would require storage of immense amounts of data.
Project Phoenix is the successor to the NASA SETI (Search for Extraterrestrial Intelligence) program that was cancelled by Congress in 1993. It tries to find extraterrestrial civilizations by listening for radio signals that are either being deliberately or inadvertently transmitted from another planet. According to the project s Web page, millions of radio channels are simultaneously monitored . Most of the listening is done by computers, and two billion channels are examined for each target star. It follows that a SAN would overcome storage limitations, and supercomputers would allow candidate signals to be found and evaluated faster. Unfortunately, the project is privately funded , and the necessary computing and SAN storage resources might be out of the project s reach.
It s apparent that better tools enable more efficient scientific achievement. From examining the origins of the universe to investigating sub-atomic particles, scientific research collects a lot of data. Research will advance faster with SAN technology to store that data.
I (Barry) was told years ago by the advertising manager of a Hollywood radio station that some day people would never need to own physical music recordings (they were LPs in those days, but the reasoning extends to cassette tapes, CDs, and DVDs). Say what? Why should they, the executive told me, when they could get all the music they wanted, whenever they wanted it, just by dialing a telephone? At the time, I was startled by this crazy concept, but as it turns out, she was absolutely right. She had inadvertently predicted MP3.
As this book goes to press, there is a raging controversy over fair payment of royalties for downloaded intellectual properties, rock music in particular. If we can work this out, we stand at the threshold of a new era for audio art: delivering more music to more people than any artist has ever dreamed of.
It makes sense. At this moment, online bookstores sell music CDs ”plastic disks in plastic boxes. They are missing the point; they should be selling music.
Given that the nominal capacity of a CD is 650 MB of data, it would take a SAN with only 650 TB of storage to deliver the content of one million CDs. The Internet would be the delivery mechanism, and various consumer music storage devices would be the target storage devices. Yes, we re doing some of this now, but the key difference is the one million CDs. That takes a SAN.
What if you could put all the information about something in one library? How about all the information about everything? The library wouldn t need to be near you if you could access it electronically. It need not even be one library, if a number of remotely located libraries worked in concert. What a concept!
To put it another way, isn t it about time all the books were in a SAN-based library? The operative word here is all.
There have been some very successful experiments with delivering books in HTML or PDF format over the Internet. For example, to read Huckleberry Finn or de Tocqueville s Democracy in America , head for the University of Virginia Web site. No fuss, no muss, no tax, no tips. It s fast and free.
Unfortunately, putting books on the World Wide Web is all too frequently a labor of love, or part of the vision of a university library. What is not a reality yet is a complete and codified body of learning. But with SANs, it can happen.
We could look to the public libraries and colleges to do this task. The trouble is that at the moment public and university libraries are not among the best-funded institutions in our country. It s a big task, but SAN technology could handle it ”if there is the will to pay for it.
A good start in organizing information is the Dewey Decimal Classification System and the ubiquitous card catalog. The system was conceived by Melvil Dewey in 1873 and was first published in 1876.
According to the Online Computer Library Center (OCLC), the Dewey Decimal Classification System is the most widely used library classification system in the world. It is used in more than 135 countries and has been translated into over 30 languages. In the United States, 95% of all public and school libraries, 25% of all college and university libraries, and 20% of special libraries use the DDCS.
Companies like Data Research Associates (DRA) of St. Louis, MO, and Monterey, CA, provide library systems software, including the card catalog component. The next step is for card catalog lookups to display links to all the found books. Those books will be stored on a number of SANs.
The conventional bricks and mortar library is not obsolete. People will always want to enjoy the quiet, page through paper books, and use the free Internet connections available there. And we ll still need the wise librarian, because a satisfactory electronic metaphor for that person still has not been developed.
|only for RuBoard - do not distribute or recompile|