Search

In the event of an Arctic oil spill, ice in the water is a complicating factor. The presence of ice complicates the forecasting of the movement and spreading of oil as well as the planning of the oil spill clean-up process. The underside of Arctic Sea ice is not flat, rather it presents nongeometric, unpredictable protrusions into the water column. The depth to which these protrusions grow relates to the longevity of the ice itself. The challenge faced by the oil spill forecaster is that information on the under-ice storage capacity is not readily available. Hence, we explored how to estimate under-ice storage capacity based on the ice stage. Historical reports on Arctic sea ice stage (e.g., first year, thin ice) were obtained from the Alaska Ocean Observing System. Ice stage data is available daily in the form of an ice "egg" code, similar to the CIS/WMO egg. Next, we acquired historic data on under-ice storage capacity. During the winters of 2010-2013, Shell deployed upward looking sonar at several sites in the Beaufort and Chukchi Seas. The sonar made direct measurements of ice draft. The under-ice storage capacity--defined as the volume of pore space above the average ice draft level--was estimated based on the ice draft data. Analyzing the ice draft alongside the ice stage data, we developed low, medium, and high estimates of oil storage capacity. When the primary ice stage is assigned an ice egg code of 1 to 3, storage capacity is deemed "low," identified by less than 15,000 m³/km². For primary ice stage with a 7 to 10 ice egg code, storage capacity is "medium," between 15,000 and 50,000 m³/km². Older ice with an ice egg code of 11 or greater has a "high" storage capacity of 50,000 to 80,000 m³/km². [ABSTRACT FROM AUTHOR]

This paper presents a novel network architecture to provide cloud computing using Passive Optical Networks (PONs). It promotes a new application model of PON, especially for the design and development of next generation PONs. The attractive features include data-plane provisioning in addition to traditional transmission-pipe functionality, a new way to subsidize access network deployment investment, as well as high-speed fiber optic resource virtualization. We investigate this new architecture's feasibility by studying PON processing, storage and interconnection capabilities. Based on the feasibility study, we further explore the infrastructure as a service (IaaS) provisioning over PONs. The PON cloud performance is evaluated from two aspects: compute capability and storage capacity. Detailed analysis shows that cloud capability is determined by the PON configuration. Relevant parameters are ONU memory capacity, ONU CPU processing capability, OLT memory capacity, OLT CPU processing capability, PON port split, PON port density, OLT card density. Our result shows that PON cloud is an attractive value-added implementation of high-speed optic-fiber access systems. [ABSTRACT FROM PUBLISHER]

The authors compare the storage capacity and other properties of various neural associative processors (APs) and find that the Ho-Kashyap (H-K) AP has the largest storage capacity and can handle linearly dependent keys. General memory (random keys) and distortion-invariant pattern-recognition APs are considered. A discussion is presented of a content-addressable structure that further improves recall accuracy and noise performance and decreases the size of the memory matrix. Results from the new H-K content-addressable AP are given. In all cases, the APs use only one pass (no iterations) in recall [ABSTRACT FROM PUBLISHER]