As seismic surveys acquire more and more data at an ever increasing rate, the conventional technologies for data recording have been showing the strain over the last few years and recently the debate has been hotting up for as to the probable replacement for the 3480/3490 cartridges of the '80s. A modern seismic survey can generate around 2 Terabytes of data. Written to the last generation industry standard 3480 cartridge with 200 MB capacity, this survey would take up 10,000 tapes. Double that for dual recording and you can see that tape handling is starting to become a serious problem.
Some history
The race for a replacement has been taking place over quite a time span, and solutions have been cropping up from a variety of sources. Firstly the 3480, just like the 9 track before it, underwent considerable development itself, with increased tape length and number of tracks. The top of the line 3490E pushed the capacity to 800 MB, and illustrated a general point; each time a new tape appears, it undergoes a series of developments and usually sees its life span extended as its own technology is pushed to the limit. With the 3490E, the last generation of cartridges reached the end of its effective extensibility. This was in 1991/92. About that time the seismic industry was beginning to make demand much higher capacities, and it was easy to see then that a change of media would be necessary to keep pace. Interestingly, IBM was testing the next generation, the 3590 even before this date, in 1990, but such is the product development cycle and the rigorous nature of the testing program that the 3590 was not issued until 1995.
Window of opportunity
This left quite a gap in the tape market and represented a window of opportunity to IBM's competitors. This competition came mainly from the data versions of the digital broadcast video marketplace, with a serious challenge from Emass using the D2 tape designed and manufactured by Ampex. Some geophysical contractors installed these machines in the early 90's, and today, with the mature, D-D2 version of the technology Ampex holds the record for capacity per tape and also for the lowest tape cost per megabyte. Notwithstanding this initial success, the D2 systems do not appear to have totally convinced the seismic industry. The rugged conditions of field seismic acquisition are a far remove from a bank's data warehouse, and during the last year or two, a new battle has been raging, between IBM's 3590, which was finally issued in early 1995, and a new offering from StorageTek, again a derivative of a video broadcast format, the DD-3. Other players in the game were Sony, with a two data formats derived from digital video SD-1 and DTF, Quantum's DLT and Cybernetics with DTF.
Bad timing
As ever, in such matters timing is everything and the DD-3 format came out just a bit late to change the geophysicists minds, who by mid 1996 were pretty well made up in favour of IBM. Despite a relatively low capacity of 10GB (it its first incarnation) and probably the highest tape cost per megabyte of all, the IBM 3590 has won the battle to be the next industry standard for acquisition. Does that then mean that it will prevail throughout the industry? Certainly, once you have cornered the acquisition marketplace, the processing house is yours, as indeed is the tape archive - at least the field tapes. This is starting to look like total domination! However, the chain of recording, processing, interpretation, trading and storage of seismic data involves a considerable amount of movement and reformatting of the data en route. The data will probably move from one media to another - generally from tape to disk several times during its lifetime. This means that it is possible that more than one media may be used in the process. So there may be a place for the competition yet. For instance on a company's data server where the higher capacity of the helicoidal technology may be a persuasive argument. Indeed PECC use STK's DD-3 technology in their Queensland National data Bank in Australia, but this is naturally not the case for the IBM dominated PetroBank and GeoBank sites.
Record every geophone?
Another force for change would be a quantum jump in the data to be recorded. Supposing that someone decided that we should record every geophone? That would put the cat among the pigeons, and might open up the race again. In this issue we look behind the commercial arguments involved in this battle of the tape formats into some of the technical issues that have driven the seismic industry to their decision. We have polled the major manufacturers of tape solutions and offer our readers an opportunity to compare for themselves the merits of the various offerings. Graph 1 shows the capacity for the different formats. Most all of the media are well below the 50 GB mark, with the 3590 a skimpy 10GB. Having said that, this format is well suited to a modern 3D seismic boat, which turns over a leisurely 5 cartridges per day for a 6 streamer boat. Indeed the massive capacity of the largest DD-2 drive is achieved with a fairly substantial physical cartridge size, perhaps 10 times the volume of the 3590 cartridge. Our second graph is of tape cost per Gigabyte. This should be treated with caution since media costs are akin to state secrets and are very dependent on time, place and market conditions. What is true is that the 3590 is one of the most expensive at over $5 per Gigabyte. This price is anticipated to fall as volumes rise, and if and when IBM allows a second manufacturer in on the act. This survey has been something of a new venture for PDM and we would appreciate your feedback. Some of the issues are still quite contentious and we hope if nothing else, to spark off some debate.
Click here to comment on this article
If your browser does not work with the MailTo button, send mail to pdm@the-data-room.com with PDM_V_2.0_199703_2 as the subject.
© Oil IT Journal - all rights reserved.