IN11C-3619:
Alaska Satellite Facility: The Quest to Stay Ahead of the Big Data Wave

Monday, 15 December 2014
Annette L Labelle-Hamer, Jeremy Nicoll and Stephen Munk, University of Alaska Fairbanks, Fairbanks, AK, United States
Abstract:
Big Data is getting bigger. Fast enough is getting faster. The number and type of products produced is growing. The ideas on how to handle the day-to-day management of data and data systems need to scale with the data and the demand. We have seen the effects of rapid growth spurts at the Alaska Satellite Facility (ASF) and anticipate we are not done yet. Looking back, ASF was conceived in 1982 to be a single-purpose imaging radar receiving station supporting a science team focused on geophysical processes. The primary construction at the University of Alaska Fairbanks (UAF) was completed in 1988 and full operational status achieved in 1991. The expected supports were estimated at 10 minutes per day and quickly grew to 70 minutes per day. In 1994, a Memorandum of Agreement (MOA) between NASA and UAF formed the ASF Distributed Active Archive Center (DAAC) complementing, the existing agreement for ASF. The demand for the use of ASF as a receiving station and as a data center grew as fast as, and at times faster, than the capabilities.

Looking forward, as demand drives the system larger just adding on more of the same often complicates rather than simplifies the system. A growing percentage of efforts and resources spent on dealing with problems that originate from a legacy system can creep up on an organization. This in turn limits the ability to keep the overall sustaining costs under control and leads to a crisis. Such growth means more-of-the-same philosophy has to shift into change-or-die philosophy in order to boot strap up to the next level. In this talk, we review how ASF has faced this several times in the past as the volume and demand of data grew along with the technology to acquire and disseminate it. We will look at what is coming for ASF as a data center and what we think are the next steps to stay ahead of the Big Data wave.