JVM-based Techniques for Improving Java Observability

Observability measures the support of computer systems to accurately capture, analyze, and present (collectively observe) the internal information about the systems. Observability frameworks play important roles for program understanding, troubleshooting, performance diagnosis, and optimizations. However, traditional solutions are either expensive or coarse-grained, consequently compromising their utility in accommodating today’s increasingly complex software systems. New solutions are emerging for VM-based languages due to the full control language VMs have over program executions. Existing such solutions, nonetheless, still lack flexibility, have high overhead, or provide limited context information for developing powerful dynamic analyses. In this thesis, we present a VM-based infrastructure, called marker tracing framework (MTF), to address the deficiencies in the existing solutions for providing better observability for VM-based languages. MTF serves as a solid foundation for implementing fine-grained low-overhead program instrumentation. Specifically, MTF allows analysis clients to: 1) define custom events with rich semantics ; 2) specify precisely the program locations where the events should trigger; and 3) adaptively enable/disable the instrumentation at runtime. In addition, MTF-based analysis clients are more powerful by having access to all information available to the VM. To demonstrate the utility and effectiveness of MTF, we present two analysis clients: 1) dynamic typestate analysis with adaptive online program analysis (AOPA); and 2) selective probabilistic calling context analysis (SPCC). In addition, we evaluate the runtime performance of MTF and the typestate client with the DaCapo benchmarks. The results show that: 1) MTF has acceptable runtime overhead when tracing moderate numbers of marker events; and 2) AOPA is highly effective in reducing the event frequency for the dynamic typestate analysis; and 3) language VMs can be exploited to offer greater observability.

Garbage collection optimization for non uniform memory access architectures

Cache-coherent non uniform memory access (ccNUMA) architecture is a standard design pattern for contemporary multicore processors, and future generations of architectures are likely to be NUMA. NUMA architectures create new challenges for managed runtime systems. Memory-intensive applications use the system’s distributed memory banks to allocate data, and the automatic memory manager collects garbage left in these memory banks. The garbage collector may need to access remote memory banks, which entails access latency overhead and potential bandwidth saturation for the interconnection between memory banks. This dissertation makes five significant contributions to garbage collection on NUMA systems, with a case study implementation using the Hotspot Java Virtual Machine. It empirically studies data locality for a Stop-The-World garbage collector when tracing connected objects in NUMA heaps. First, it identifies a locality richness which exists naturally in connected objects that contain a root object and its reachable set— ‘rooted sub-graphs’. Second, this dissertation leverages the locality characteristic of rooted sub-graphs to develop a new NUMA-aware garbage collection mechanism. A garbage collector thread processes a local root and its reachable set, which is likely to have a large number of objects in the same NUMA node. Third, a garbage collector thread steals references from sibling threads that run on the same NUMA node to improve data locality. This research evaluates the new NUMA-aware garbage collector using seven benchmarks of an established real-world DaCapo benchmark suite. In addition, evaluation involves a widely used SPECjbb benchmark and Neo4J graph database Java benchmark, as well as an artificial benchmark. The results of the NUMA-aware garbage collector on a multi-hop NUMA architecture show an average of 15% performance improvement. Furthermore, this performance gain is shown to be as a result of an improved NUMA memory access in a ccNUMA system. Fourth, the existing Hotspot JVM adaptive policy for configuring the number of garbage collection threads is shown to be suboptimal for current NUMA machines. The policy uses outdated assumptions and it generates a constant thread count. In fact, the Hotspot JVM still uses this policy in the production version. This research shows that the optimal number of garbage collection threads is application-specific and configuring the optimal number of garbage collection threads yields better collection throughput than the default policy. Fifth, this dissertation designs and implements a runtime technique, which involves heuristics from dynamic collection behavior to calculate an optimal number of garbage collector threads for each collection cycle. The results show an average of 21% improvements to the garbage collection performance for DaCapo benchmarks.

Self landcape - Hotspot

Self Landscape is an introductory image of the author's exhibition Lightsite, which toured Western Australian galleries from February 2006 to November 2007. It is a five-minute-long exposure photographic image captured inside a purpose-built, room-sized pinhole camera which is demountable and does not have a floor. Self Landscape depicts an inverted image of the outside environment - a revegetated farmland in the Great Southern Region of Western Australia. The light from the exterior landscape passes though the pinhole camera's aperture and illuminates the internal scene which includes that part of the landscape upon which the floorless room is erected, along with the author who is standing inside. The text "Hotspot" appears where the author has scribed that word with a torch during the long exposure. The image evokes the temporality of light. Here, light itself is portrayed as the primary medium through which we both perceive and describe landscape. It is through the agency of light that we construct our connectivity to landscape.

Experimental evaluation of hotspot identification methods

Identifying crash “hotspots”, “blackspots”, “sites with promise”, or “high risk” locations is standard practice in departments of transportation throughout the US. The literature is replete with the development and discussion of statistical methods for hotspot identification (HSID). Theoretical derivations and empirical studies have been used to weigh the benefits of various HSID methods; however, a small number of studies have used controlled experiments to systematically assess various methods. Using experimentally derived simulated data—which are argued to be superior to empirical data, three hot spot identification methods observed in practice are evaluated: simple ranking, confidence interval, and Empirical Bayes. Using simulated data, sites with promise are known a priori, in contrast to empirical data where high risk sites are not known for certain. To conduct the evaluation, properties of observed crash data are used to generate simulated crash frequency distributions at hypothetical sites. A variety of factors is manipulated to simulate a host of ‘real world’ conditions. Various levels of confidence are explored, and false positives (identifying a safe site as high risk) and false negatives (identifying a high risk site as safe) are compared across methods. Finally, the effects of crash history duration in the three HSID approaches are assessed. The results illustrate that the Empirical Bayes technique significantly outperforms ranking and confidence interval techniques (with certain caveats). As found by others, false positives and negatives are inversely related. Three years of crash history appears, in general, to provide an appropriate crash history duration.

The influence of underreported crashes on hotspot identification

Hot spot identification (HSID) plays a significant role in improving the safety of transportation networks. Numerous HSID methods have been proposed, developed, and evaluated in the literature. The vast majority of HSID methods reported and evaluated in the literature assume that crash data are complete, reliable, and accurate. Crash under-reporting, however, has long been recognized as a threat to the accuracy and completeness of historical traffic crash records. As a natural continuation of prior studies, the paper evaluates the influence that under-reported crashes exert on HSID methods. To conduct the evaluation, five groups of data gathered from Arizona Department of Transportation (ADOT) over the course of three years are adjusted to account for fifteen different assumed levels of under-reporting. Three identification methods are evaluated: simple ranking (SR), empirical Bayes (EB) and full Bayes (FB). Various threshold levels for establishing hotspots are explored. Finally, two evaluation criteria are compared across HSID methods. The results illustrate that the identification bias—the ability to correctly identify at risk sites--under-reporting is influenced by the degree of under-reporting. Comparatively speaking, crash under-reporting has the largest influence on the FB method and the least influence on the SR method. Additionally, the impact is positively related to the percentage of the under-reported PDO crashes and inversely related to the percentage of the under-reported injury crashes. This finding is significant because it reveals that despite PDO crashes being least severe and costly, they have the most significant influence on the accuracy of HSID.

Hotspot Identification: A Full Bayesian Hierarchical Modeling Approach

This study proposes a full Bayes (FB) hierarchical modeling approach in traffic crash hotspot identification. The FB approach is able to account for all uncertainties associated with crash risk and various risk factors by estimating a posterior distribution of the site safety on which various ranking criteria could be based. Moreover, by use of hierarchical model specification, FB approach is able to flexibly take into account various heterogeneities of crash occurrence due to spatiotemporal effects on traffic safety. Using Singapore intersection crash data(1997-2006), an empirical evaluate was conducted to compare the proposed FB approach to the state-of-the-art approaches. Results show that the Bayesian hierarchical models with accommodation for site specific effect and serial correlation have better goodness-of-fit than non hierarchical models. Furthermore, all model-based approaches perform significantly better in safety ranking than the naive approach using raw crash count. The FB hierarchical models were found to significantly outperform the standard EB approach in correctly identifying hotspots.

The Cougar Point tuff: Implications for thermochemical zonation and longevity of high-temperature, large-volume silicic magmas of the Miocene Yellowstone hotspot

The 12.7-10.5 Ma Cougar Point Tuff in southern Idaho, USA, consists of 10 large-volume (>10²-10³ km³ each), high-temperature (800-1000 °C), rhyolitic ash-flow tuffs erupted from the Bruneau-Jarbidge volcanic center of the Yellowstone hotspot. These tuffs provide evidence for compositional and thermal zonation in pre-eruptive rhyolite magma, and suggest the presence of a long-lived reservoir that was tapped by numerous large explosive eruptions. Pyroxene compositions exhibit discrete compositional modes with respect to Fe and Mg that define a linear spectrum punctuated by conspicuous gaps. Airfall glass compositions also cluster into modes, and the presence of multiple modes indicates tapping of different magma volumes during early phases of eruption. Equilibrium assemblages of pigeonite and augite are used to reconstruct compositional and thermal gradients in the pre-eruptive reservoir. The recurrence of identical compositional modes and of mineral pairs equilibrated at high temperatures in successive eruptive units is consistent with the persistence of their respective liquids in the magma reservoir. Recurrence intervals of identical modes range from 0.3 to 0.9 Myr and suggest possible magma residence times of similar duration. Eruption ages, magma temperatures, Nd isotopes, and pyroxene and glass compositions are consistent with a long-lived, dynamically evolving magma reservoir that was chemically and thermally zoned and composed of multiple discrete magma volumes.

High-temperature, low-H₂O silicic magmas of the Yellowstone Hotspot: an experimental study of Rhyolite from the Bruneau-Jarbidge eruptive center, Central Snake River Plain, USA

The phase relations have been investigated experimentally at 200 and 500 MPa as a function of water activity for one of the least evolved (Indian Batt Rhyolite) and of a more evolved rhyolite composition (Cougar Point Tuff XV) from the 12·8-8·1 Ma Bruneau-Jarbidge eruptive center of the Yellowstone hotspot. Particular priority was given to accurate determination of the water content of the quenched glasses using infrared spectroscopic techniques. Comparison of the composition of natural and experimentally synthesized phases confirms that high temperatures (>900°C) and extremely low melt water contents (<1·5 wt % H₂O) are required to reproduce the natural mineral assemblages. In melts containing 0·5-1·5 wt % H₂O, the liquidus phase is clinopyroxene (excluding Fe-Ti oxides, which are strongly dependent on fO₂), and the liquidus temperature of the more evolved Cougar Point Tuff sample (BJR; 940-1000°C) is at least 30°C lower than that of the Indian Batt Rhyolite lava sample (IBR2; 970-1030°C). For the composition BJR, the comparison of the compositions of the natural and experimental glasses indicates a pre-eruptive temperature of at least 900°C. The composition of clinopyroxene and pigeonite pairs can be reproduced only for water contents below 1·5 wt % H₂O at 900°C, or lower water contents if the temperature is higher. For the composition IBR2, a minimum temperature of 920°C is necessary to reproduce the main phases at 200 and 500 MPa. At 200 MPa, the pre-eruptive water content of the melt is constrained in the range 0·7-1·3 wt % at 950°C and 0·3-1·0 wt % at 1000°C. At 500 MPa, the pre-eruptive temperatures are slightly higher (by 30-50°C) for the same ranges of water concentration. The experimental results are used to explore possible proxies to constrain the depth of magma storage. The crystallization sequence of tectosilicates is strongly dependent on pressure between 200 and 500 MPa. In addition, the normative Qtz-Ab-Or contents of glasses quenched from melts coexisting with quartz, sanidine and plagioclase depend on pressure and melt water content, assuming that the normative Qtz and Ab/Or content of such melts is mainly dependent on pressure and water activity, respectively. The combination of results from the phase equilibria and from the composition of glasses indicates that the depth of magma storage for the IBR2 and BJR compositions may be in the range 300-400 MPa (13 km) and 200-300 MPa (10 km), respectively.

Low δ18O zircons from the Bruneau-­Jarbidge eruptive center: A key to crustal anatexis along the track of the Yellowstone hotspot

The Bruneau-Jarbidge eruptive center (BJEC) in the central Snake River Plain, Idaho, USA consists of the Cougar Point Tuff (CPT), a series of ten, high-temperature (900-1000°C) voluminous ignimbrites produced over the explosive phase of volcanism (12.8-10.5 Ma) and more than a dozen equally high-temperature rhyolite lava flows produced during the effusive phase (10.5-8 Ma). Spot analyses by ion microprobe of oxygen isotope ratios in 210 zircons demonstrate that all of the eruptive units of the BJEC are characterized by zircon δ¹⁸O values ≤ 2.5‰, thus documenting the largest low δ¹⁸O silicic volcanic province known on Earth (>10⁴ km³). There is no evidence for voluminous normal δ¹⁸O magmatism at the BJEC that precedes generation of low δ¹⁸O magmas as there is at other volcanic centers that generate low δ¹⁸O magmas such as Heise and Yellowstone. At these younger volcanic centers of the hotspot track, such low δ¹⁸O magmas represent ~45 % and ~20% respectively of total eruptive volumes. Zircons in all BJEC tuffs and lavas studied (23 units) document strong δ¹⁸O depletion (median CPT δ¹⁸OZrc = 1.0‰, post-CPT lavas = 1.5‰) with the third member of the CPT recording an excursion to minimum δ¹⁸O values (δ¹⁸OZrc= -1.8‰) in a supereruption > 2‰ lower than other voluminous low δ¹⁸O rhyolites known worldwide (δ¹⁸OWR ≤0.9 vs. 3.4‰). Subsequent units of the CPT and lavas record a progressive recovery in δ¹⁸OZrc to ~2.5‰ over a ~ 4 m.y. interval (12 to 8 Ma). We present detailed evidence of unit-to-unit systematic patterns in O isotopic zoning in zircons (i.e. direction and magnitude of Δcore-rim), spectrum of δ¹⁸O in individual units, and zircon inheritance patterns established by re-analysis of spots for U-Th-Pb isotopes by LA-ICPMS and SHRIMP. In conjunction with mineral thermometry and magma compositions, these patterns are difficult to reconcile with the well-established model for "cannibalistic" low δ¹⁸O magma genesis at Heise and Yellowstone. We present an alternative model for the central Snake River Plain using the modeling results of Leeman et al. (2008) for ¹⁸O depletion as a function of depth in a mid-upper crustal protolith that was hydrothermally altered by infiltrating meteoric waters prior to the onset of silicic magmatism. The model proposes that BJEC silicic magmas were generated in response to the propagation of a melting front, driven by the incremental growth of a vast underlying mafic sill complex, over a ~5 m.y. interval through a crustal volume in which a vertically asymmetric δ¹⁸OWR gradient had previously developed that was sharply inflected from ~ -1 to 10‰ at mid-upper crustal depths. Within the context of the model, data from BJEC zircons are consistent with incremental melting and mixing events in roof zones of magma reservoirs that accompany surfaceward advance of the coupled mafic-silicic magmatic system.

Numerical estimation of hotspot temperatures in electrodes subjected to pulsed electron beam heating in vacuum

A numerical solution for the transient temperature distribution in a cylindrical disc heated on its top surface by a circular source is presented. A finite difference form of the governing equations is solved by the Alternating Direction Implicit (ADI) time marching scheme. This solution has direct applications in analyzing transient electron beam heating of target materials as encountered in the prebreakdown current enhancement and consequent breakdown in high voltage vacuum gaps stressed by alternating and pulsed voltages. The solution provides an estimate of the temperature for pulsed electron beam heating and the size of thermally activated microparticles originating from anode hot spots. The calculated results for a typical 45kV (a.c.) electron beam of radius 2.5 micron indicate that the temperature of such spots can reach melting point and could give rise to microparticles which could initiate breakdown.

India's biodiversity hotspot under anthropogenic pressure: A case study of Nilgiri Biosphere Reserve

This paper presents data on the impact of biotic pressure in terms of grazing by livestock and wood cutting by humans on the plant community in the Nilgiri Biosphere Reserve of India. Grass, and herbaceous plant biomass, number of cattle dung piles, number of woody stems available and damaged by human activities and weed biomass were assessed at different proximity along transects radiating from village-forest boundary to forest interior to measure the ecological impact of livestock grazing and fire wood collection. The grass biomass was positively correlated to overgrazing indicating the adverse effect on natural vegetation by cattle. Woodcutting was intense along the forest boundary and significantly declined as distance increased. Similarly, weed biomass and number of thorny species declined positively with proximity from village-forest boundary and the weed biomass was significantly higher in the pastoral sites compared to residential sites. The results suggest that human impact adversely affects natural vegetation and promotes weed proliferation in forest areas adjoining human settlements in the ecologically important Nilgiri Biosphere Reserve. Continued anthropogenic pressure could cause reduction in fodder availability to large herbivores like elephants, which in turn leads to an increase in human-elephant conflict. (C) 2011 Elsevier GmbH. All rights reserved.

Phaeocystis colony distribution in the North Atlantic Ocean since 1948, and interpretation of long-term changes in the Phaeocystis hotspot in the North Sea

Monitoring of Phaeocystis since 1948 during the Continuous Plankton Recorder survey indicates that over the last 5.5 decades the distribution of its colonies in the North Atlantic Ocean was not restricted to neritic waters: occurrence was also recorded in the open Atlantic regions sampled, most frequently in the spring. Apparently, environmental conditions in open ocean waters, also those far oVshore, are suitable for complete lifecycle development of colonies (the only stage recorded in the survey). In the North Sea the frequency of occurrence was also highest in spring. Its southeastern part was the Phaeocystis abundance hotspot of the whole area covered by the survey. Frequency was especially high before the 1960s and after the 1980s, i.e., in the periods when anthropogenic nutrient enrichment was relatively low. Changes in eutrophication have obviously not been a major cause of long-term Phaeocystis variation in the southeastern North Sea, where total phytoplankton biomass was related signiWcantly to river discharge. Evidence is presented for the suggestion that Phaeocystis abundance in the southern North Sea is to a large extent determined by the amount of Atlantic Ocean water Xushed in through the Dover Strait. Since Phaeocystis plays a key role in element Xuxes relevant to climate the results presented here have implications for biogeochemical models of cycling of carbon and sulphur. Sea-to-air exchange of CO2 and dimethyl sulphide (DMS) has been calculated on the basis of measurements during single-year cruises. The considerable annual variation in phytoplankton and in its Phaeocystis component reported here does not warrant extrapolation of such figures.