On the Convergence of Statistical Techniques for Inferring Network Traffic Demands

Accurate knowledge of traffic demands in a communication network enables or enhances a variety of traffic engineering and network management tasks of paramount importance for operational networks. Directly measuring a complete set of these demands is prohibitively expensive because of the huge amounts of data that must be collected and the performance impact that such measurements would impose on the regular behavior of the network. As a consequence, we must rely on statistical techniques to produce estimates of actual traffic demands from partial information. The performance of such techniques is however limited due to their reliance on limited information and the high amount of computations they incur, which limits their convergence behavior. In this paper we study strategies to improve the convergence of a powerful statistical technique based on an Expectation-Maximization iterative algorithm. First we analyze modeling approaches to generating starting points. We call these starting points informed priors since they are obtained using actual network information such as packet traces and SNMP link counts. Second we provide a very fast variant of the EM algorithm which extends its computation range, increasing its accuracy and decreasing its dependence on the quality of the starting point. Finally, we study the convergence characteristics of our EM algorithm and compare it against a recently proposed Weighted Least Squares approach.

On the Relationship Between File Sizes, Transport Protocols, and Self-Similar Network Traffic

Recent measurements of local-area and wide-area traffic have shown that network traffic exhibits variability at a wide range of scales self-similarity. In this paper, we examine a mechanism that gives rise to self-similar network traffic and present some of its performance implications. The mechanism we study is the transfer of files or messages whose size is drawn from a heavy-tailed distribution. We examine its effects through detailed transport-level simulations of multiple TCP streams in an internetwork. First, we show that in a "realistic" client/server network environment i.e., one with bounded resources and coupling among traffic sources competing for resources the degree to which file sizes are heavy-tailed can directly determine the degree of traffic self-similarity at the link level. We show that this causal relationship is not significantly affected by changes in network resources (bottleneck bandwidth and buffer capacity), network topology, the influence of cross-traffic, or the distribution of interarrival times. Second, we show that properties of the transport layer play an important role in preserving and modulating this relationship. In particular, the reliable transmission and flow control mechanisms of TCP (Reno, Tahoe, or Vegas) serve to maintain the long-range dependency structure induced by heavy-tailed file size distributions. In contrast, if a non-flow-controlled and unreliable (UDP-based) transport protocol is used, the resulting traffic shows little self-similar characteristics: although still bursty at short time scales, it has little long-range dependence. If flow-controlled, unreliable transport is employed, the degree of traffic self-similarity is positively correlated with the degree of throttling at the source. Third, in exploring the relationship between file sizes, transport protocols, and self-similarity, we are also able to show some of the performance implications of self-similarity. We present data on the relationship between traffic self-similarity and network performance as captured by performance measures including packet loss rate, retransmission rate, and queueing delay. Increased self-similarity, as expected, results in degradation of performance. Queueing delay, in particular, exhibits a drastic increase with increasing self-similarity. Throughput-related measures such as packet loss and retransmission rate, however, increase only gradually with increasing traffic self-similarity as long as reliable, flow-controlled transport protocol is used.

An Integrated Neural Network-Event-Related Potentials Model of Temporal and Probability Context Effects on Event Categorization

We present a neural network that adapts and integrates several preexisting or new modules to categorize events in short term memory (STM), encode temporal order in working memory, evaluate timing and probability context in medium and long term memory. The model shows how processed contextual information modulates event recognition and categorization, focal attention and incentive motivation. The model is based on a compendium of Event Related Potentials (ERPs) and behavioral results either collected by the authors or compiled from the classical ERP literature. Its hallmark is, at the functional level, the interplay of memory registers endowed with widely different dynamical ranges, and at the structural level, the attempt to relate the different modules to known anatomical structures.

Effect of solder volume on joint shape with variable chip-to-board contact pad ratio

The objective of this paper is to investigate the effect of the pad size ratio between the chip and board end of a solder joint on the shape of that solder joint in combination with the solder volume available. The shape of the solder joint is correlated to its reliability and thus of importance. For low density chip bond pad applications Flip Chip (FC) manufacturing costs can be kept down by using larger size board pads suitable for solder application. By using “Surface Evolver” software package the solder joint shapes associated with different size/shape solder preforms and chip/board pad ratios are predicted. In this case a so called Flip-Chip Over Hole (FCOH) assembly format has been used. Assembly trials involved the deposition of lead-free 99.3Sn0.7Cu solder on the board side, followed by reflow, an underfill process and back die encapsulation. During the assembly work pad off-sets occurred that have been taken into account for the Surface Evolver solder joint shape prediction and accurately matched the real assembly. Overall, good correlation was found between the simulated solder joint shape and the actual fabricated solder joint shapes. Solder preforms were found to exhibit better control over the solder volume. Reflow simulation of commercially available solder preform volumes suggests that for a fixed stand-off height and chip-board pad ratio, the solder volume value and the surface tension determines the shape of the joint.

Influence of network topology and data collection on network inference.

We recently developed an approach for testing the accuracy of network inference algorithms by applying them to biologically realistic simulations with known network topology. Here, we seek to determine the degree to which the network topology and data sampling regime influence the ability of our Bayesian network inference algorithm, NETWORKINFERENCE, to recover gene regulatory networks. NETWORKINFERENCE performed well at recovering feedback loops and multiple targets of a regulator with small amounts of data, but required more data to recover multiple regulators of a gene. When collecting the same number of data samples at different intervals from the system, the best recovery was produced by sampling intervals long enough such that sampling covered propagation of regulation through the network but not so long such that intervals missed internal dynamics. These results further elucidate the possibilities and limitations of network inference based on biological data.

Material properties, geometry and their effect on the fatigue life of two flip-chip models

This paper describes how modeling technology has been used in providing fatigue life time data of two flip-chip models. Full-scale three-dimensional modeling of flip-chips under cyclic thermal loading has been combined with solder joint stand-off height prediction to analyze the stress and strain conditions in the two models. The Coffin-Manson empirical relationship is employed to predict the fatigue life times of the solder interconnects. In order to help designers in selecting the underfill material and the printed circuit board, the Young's modulus and the coefficient of thermal expansion of the underfill, as well as the thickness of the printed circuit boards are treated as variable parameters. Fatigue life times are therefore calculated over a range of these material and geometry parameters. In this paper we will also describe how the use of micro-via technology may affect fatigue life

The effect of reflow process on the contact resistance and reliability of anisotropic conductive film interconnection for flip chip on flex applications

The work presented in this paper focuses on the effect of reflow process on the contact resistance and reliability of anisotropic conductive film (ACF) interconnection. The contact resistance of ACF interconnection increases after reflow process due to the decrease in contact area of the conducting particles between the mating I/O pads. However, the relationship between the contact resistance and bonding parameters of the ACF interconnection with reflow treatment follows the similar trend to that of the as-bonded (i.e. without reflow) ACF interconnection. The contact resistance increases as the peak temperature of reflow profile increases. Nearly 40% of the joints were found to be open after reflow with 260 °C peak temperature. During the reflow process, the entrapped (between the chip and substrate) adhesive matrix tries to expand much more than the tiny conductive particles because of the higher coefficient of thermal expansion, the induced thermal stress will try to lift the bump from the pad and decrease the contact area of the conductive path and eventually, leading to a complete loss of electrical contact. In addition, the environmental effect on contact resistance such as high temperature/humidity aging test was also investigated. Compared with the ACF interconnections with Ni/Au bump, higher thermal stress in the Z-direction is accumulated in the ACF interconnections with Au bump during the reflow process owing to the higher bump height, thus greater loss of contact area between the particles and I/O pads leads to an increase of contact resistance and poorer reliability after reflow.

Effects of reflow process on the reliability of flip chip on flex interconnections using anisotropic conductive adhesives

In this paper, the effects of the solder reflow process on the reliability of anisotropic conductive film (ACF) interconnections for flip chip on flex (FCOF) applications are investigated. Experiments as well as computer modeling methods have been used. In the experiments, it was found that the contact resistance of ACF joints increased after the subsequent reflow process, and the magnitude of this increase was strongly correlated to the peak temperature of the reflow profile. Nearly 40% of the joints were opened (i.e. lifted away from the pad) after the reflow process with 260 °C peak temperature while no opening was observed when the peak temperature was 210 °C. It is believed that the CTE mismatch between the polymer particle and the adhesive matrix is the main cause of this contact degradation. It was also found that the ACF joints after the reflow process with 210 °C peak temperature showed a high ability to resist water absorption under steady state 85 °C/85%RH conditions, probably because the curing degree of the ACF was improved during the reflow process. To give a good understanding, a 3D model of an ACF joint structure was built and finite element analysis was used to predict the stress distribution in the conductive particles, adhesive matrix and metal pads during the reflow process.

The effect of curing on the performance of ACF bonded chip-on-flex assemblies after thermal ageing

Purpose – Anisotropic conductive film (ACF) is now an attractive technology for direct mounting of chips onto the substrate as an alternative to lead-free solders. However, despite its various advantages over other technologies, it also has many unresolved reliability issues. For instance, the performance of ACF assembly in high temperature applications is questionable. The purpose of this paper is to study the effect of bonding temperatures on the curing of ACFs, and their mechanical and electrical performance after high temperature ageing. Design/methodology/approach – In the work presented in this paper, the curing degree of an ACF at different bonding temperatures was measured using a differential scanning calorimeter. The adhesion strength and the contact resistance of ACF bonded chip-on-flex assembly were measured before and after thermal ageing and the results were correlated with the curing degree of ACF. The ACF was an epoxy-based adhesive in which Au-Ni coated polymer particles were randomly dispersed. Findings – The results showed that higher bonding temperatures had resulted in better ACF curing and stronger adhesion. After ageing, the adhesion strength increased for the samples bonded at lower temperatures and decreased for the samples bonded at higher temperatures. ACF assemblies with higher degrees of curing showed smaller increases in contact resistance after ageing. Conduction gaps at the bump-particle and/or particle-pad interfaces were found with the help of scanning electron microscopy and are thought to be the root cause of the increase in contact resistance. Originality/value – The present study focuses on the effect of bonding temperatures on the curing of ACFs, and their adhesion strength and electrical performances after high temperature ageing. The results of this study may help the development of ACFs with higher heat resistance, so that ACFs can be considered as an alternative to lead-free solders.

Moisture Effects on the Reliability of Anisotropic Conductive Film Interconnection for Flip Chip on Flex Applications

Anisotropic conductive film (ACF) which consists of an adhesive epoxy matrix and randomly distributed conductive particles are widely used as the connection material for electronic devices with high I/O counts. However, for the semiconductor industry the reliability of the ACF is still a major concern due to a lack of experimental reliability data. This paper reports the investigations into the moisture-induced failures in Flip-Chip-on-Flex interconnections with Anisotropic Conductive Films (ACFs). Both experimental and modeling methods were applied. In the experiments, the contact resistance was used as a quality indicator and was measured continuously during the accelerated tests (autoclave tests). The temperature, relative humidity and the pressure were set at 121°C, 100%RH, and 2atm respectively. The contact resistance of the ACF joints increased during the tests and nearly 25% of the joints were found to be open after 168 hours’ testing time. Visible conduction gaps between the adhesive and substrate pads were observed. Cracks at the adhesive/flex interface were also found. For a better understanding of the experimental results, 3-D Finite Element (FE) models were built and a macro-micro modeling method was used to determine the moisture diffusion and moisture-induced stresses inside the ACF joints. Modeling results are consistent with the findings in the experimental work.

Experimental and Modelling Analysis on the Moisture Induced Failures in Flip Chip on Flex Interconnections with Anisotropic Conductive Film

This paper reports the investigations into the moisture induced failures in flip-chip-on-flex interconnections with anisotropic conductive films (ACF). Both experimental and modeling methods were applied. In the experiments, the contact resistance was used as a quality indicator and was measured continuously during the accelerated tests (autoclave tests). The temperature, relative humidity and the pressure were set at 121°C, 100%RH, 1atm respectively. The contact resistance of the ACF joints increased during the tests and nearly 25% of the joints were found to be open after 168 hours' testing time. Visible conduction gaps between the adhesive and substrate pads were observed. Cracks at the adhesive/flex interface were also found. It is believed that the swelling effect of the adhesive and the water penetration along the adhesive/flex interface are the main causes of this contact degradation. Another finding from the experimental work was that the ACF interconnections that had undergone the reflow treatment were more sensitive to the moisture and showed worse reliability during the tests. For a better understanding of the experimental results, 3D finite element (FE) models were built and a macro-micro modeling method was used to determine the moisture diffusion and moisture-induced stresses inside the ACF joints. Modeling results are consistent with the findings in the experimental work.

Effects of solder reflow on the reliability of flip-chip on flex interconnections using anisotropic conductive adhesives

This work describes the work of an investigation of the effects of solder reflow process on the reliability of anisotropic conductive film (ACF) interconnection for flip-chip on flex (FCOF) applications. Experiments as well as computer modeling methods have been used. The results show that the contact resistance of ACF interconnections increases after the reflow and the magnitude of the increase is strongly correlated to the peak reflow temperature. In fact, nearly 40 percent of the joints are open when the peak reflow temperature is 260°C, while there is no opening when the peak temperature is 210°C. It is believed that the coefficient of thermal expansion (CTE) mismatch between the polymer particle and the adhesive matrix is the main cause of this contact degradation. To understand this phenomenon better, a three-dimensional (3-D) finite element (FE) model of an ACF joint has been analyzed in order to predict the stress distribution in the conductive particles, adhesive matrix and metal pads during the reflow process. The stress level at the interface between the particle and its surrounding materials is significant and it is the highest at the interface between the particle and the adhesive matrix.

Effects of thermal cycling profiles on the performance of chip-on-flex assembly using anisotropic conductive films

Anisotropic conductive films (ACFs) are widely used in the electronic packaging industries because of their fine pitch potential and the assembly process is simpler compared to the soldering process. However, there are still unsolved issues in the volume productions using ACFs. The main reason is that the effects of many factors on the interconnects are not well understood. This work focuses on the performance of ACF-bonded chip-on-flex assemblies subjected to a range of thermal cycling test conditions. Both experimental and three-dimensional finite element computer modelling methods are used. It has been revealed that greater temperature ranges and longer dwell-times give rise to higher stresses in the ACF interconnects. Higher stresses are concentrated along the edges of the chip-ACF interfaces. In the experiments, the results show that higher temperature ranges and prolonged dwell times increase contact resistance values. Close examination of the microstructures along the bond-line through the scanning electron microscope (SEM) indicates that cyclic thermal loads disjoint the conductive particles from the bump of the chip and/or pad of the substrate and this is thought to be related to the increase of the contact resistance value and the failure of the ACF joints.

Role of bonding time and temperature on the physical properties of coupled anisotropic conductive–nonconductive adhesive film for flip chip on glass technology

Using thermosetting epoxy based conductive adhesive films for the flip chip interconnect possess a great deal of attractions to the electronics manufacturing industries due to the ever increasing demands for miniaturized electronic products. Adhesive manufacturers have taken many attempts over the last decade to produce a number of types of adhesives and the coupled anisotropic conductive-nonconductive adhesive film is one of them. The successful formation of the flip chip interconnection using this particular type of adhesive depends on, among factors, how the physical properties of the adhesive changes during the bonding process. Experimental measurements of the temperature in the adhesive have revealed that the temperature becomes very close to the required maximum bonding temperature within the first 1s of the bonding time. The higher the bonding temperature the faster the ramp up of temperature is. A dynamic mechanical analysis (DMA) has been carried out to investigate the nature of the changes of the physical properties of the coupled anisotropic conductive-nonconductive adhesive film for a range of bonding parameters. Adhesive samples that are pre-cured at 170, 190 and 210°C for 3, 5 and 10s have been analyzed using a DMA instrument. The results have revealed that the glass transition temperature of this type of adhesive increases with the increase in the bonding time for the bonding temperatures that have been used in this work. For the curing time of 3 and 5s, the maximum glass transition temperature increases with the increase in the bonding temperature, but for the curing time of 10s the maximum glass transition temperature has been observed in the sample which is cured at 190°C. Based on these results it has been concluded that the optimal bonding temperature and time for this kind of adhesive are 190°C and 10s, respectively.