NASA's Use of the World Wide Web to Deliver Shoemaker-Levy 9 Collision Data in Near-Real Time

By midday both the author and Dave Williams, NSSDC's resident planetary scientist, had put together a dozen or so of the first observations. This was a difficult task to accomplish since we had to acquire the data from several observatory sites (usually via FTP) while networks had already begun to experience a global slowdown due to the JPL and STScI activity. We tested the system and made it available through our existing services on our primary WWW server Bolero. Much later that night the author made an announcement on several network news groups that the National Space Science Data Center was now providing SL-9 observation through its WWW service.

Figure 1

At the same time network administrators at Goddard were very concerned with the saturation of the backbone and consequent loss or severe slowdown of other services. The Space Telescope Sciences Institute was also experiencing similar difficulties as Fred Romelfanger, developer of their SL-9 WWW system noted,

"[unprecedented access]...cause our T1 link to saturate, and we went off the air that night while NSI moved our network link due to problems that the data transfers were causing other sites. The saturation lasted for about a week and a half. During this time access to the Internet was very slow, but not impossible to use."

• Minimize the number of HTML pages.
• Maximize file transmission rate.

We made the assumption that the user was essentially interested in the latest observations and not necessarily on the background or other textual information. The whole system was only two layers deep so as to minimize browse activity. We were interested in wanting the user to come in, go to the observatory or fragment they are interested in, view the browse image, click on the image and caption to get the originals, and leave. The home page listed all the observatories from which we had acquired data. The home page also include links to other sites that was providing data. We wanted to encourage the user to look elsewhere immediately if they did not see the observatory of their choice on our list. Next to the name of each observatory, we put a "time stamp" as to when the last update was made to that observatory page, therefore preventing the user from needlessly looking to see if any new data had been obtained since they last visited. We also put a list of the comet fragments for which we had data next to the observatory names. Again, this was done to discourage unnecessary browsing.

Once selecting the observatory of their choice, the user sees the second and last layer of the system. The "observatory page" simply had a collection of in-line "thumbnail" images sorted my observation date and fragment. Below each image, a graphic bullet linked to a text file containing the image caption. Full resolution images was retrieved by selecting the thumbnail.

We knew our SL-9 system was going to be very heavily used. However, we were essentially constructing an image server, so somehow we needed to minimize the impact of transmitting large image files. We decided that we would not use any full resolution images on any page, but instead used in-line browse images. We experimented with several sizes before deciding that 100x100 pixel was sufficiently large for the user to view the essential features of the image. With hindsight, halving the file size and going down to 70x70 pixel might have been better choice.

Textual information such as captions were not included within the HTML pages, thus discouraging browsing, but was made available using a link at the bottom of each image. We also consciously avoided including animated data and motion pictures realizing that their immense size would create a massive bottle neck for users who just wanted the latest observations. We made at least three updates daily to the SL-9 system to ensure that the most recent observations were available. Normally, this was done via FTP to the observatory sites very early each morning when network traffic was lowest.

Syed S. Towheed
Code 633
National Space Science Data Center
NASA Goddard Space Flight Center
Greenbelt, MD 20771, USA.
Phone: (301)286-4136
Fax: (301)286-1771
E-mail: towheed@nssdca.gsfc.nasa.gov