CSE Speaker Series – Dr. Drew Hamilton
On March 1 at 11:00-12:00 in Cramer 221, Dr. Drew Hamilton from Mississippi State University will discuss “What if a Simulation is Too Good?”
The need for simulation software vulnerability assessment is being driven by three major trends: increased use of modeling and simulation for training and operational planning; increased emphasis on coalition warfare and interoperability and finally increased awareness of the potential security risks inherent in sharing operationally useful software. This presentation will address calculating the parameterization of the simulation, as well as disassembly, decompilation and runtime execution analysis. Additionally, we will discuss training, tactics and procedures that can be gleaned from a high fidelity simulation. This presentation will describe in an unclassified manner the process developed by Dr. Hamilton and the Missile Defense Agency to evaluate potential vulnerabilities in shared simulation software.
CSE Speaker Series – Dongwan Shin
On February 10 in Cramer 221 from 11:00-12:00, Dr. Dongwan Shin will discuss Usability Considerations on Some Security Services: Authentication and Access Control
Abstract: Authentication and access control are critical security services for protecting computer systems; the former is to prove to systems who you are, while the latter is to entitle what you can do after successful authentication. A variety of cryptographic protocols, policy mechanisms, and/or techniques have been developed to support those services; and some of them have been successfully adopted and used in many systems and applications, whereas others were eventually discarded. In this talk, I will first discuss the assumption of open world, where a user is not known a priori to the system that she tries to access for using its resources, and then present some of the solutions to support authentication and access control services under the assumption. Lastly, I will discuss my prior and ongoing research problems related to the two important security services, especially in consideration of usability.
CSE Thesis Defense – Nick Davis
On Tuesday, December 20th at 1:00 pm, Nick Davis will defend his thesis “Approximating The Makespan Of Workflows Submitted To Heterogeneous Federated Cloud Computing Environments Using Neural Networks” in Cramer Hall 221.
Cloud Computing has become very popular in recent years. With open source cloud computing infrastructure available for businesses and consumers, anyone with the hardware can create their own private or public cloud. This, compounded with the fact that hardware is cheaper than ever, has led to an increase in the number of cloud datacenters. Large scale companies like Amazon and Microsoft have developed cloud computing datacenters that excel in executing lots of complex tasks, while smaller companies have developed their own internal cloud infrastructures. This proliferation of cloud computing technologies has resulted in a rush to develop larger, more powerful cloud datacenters.
Unfortunately, cooling costs and operation costs severely limit the total computational capacity of most large datacenters. Instead of increasing the size of individual datacenters, many cloud providers are linking datacenters and sharing resources through unified interfaces, creating what are known as cloud federations. Moreover, as clouds become more common, smaller providers and larger providers are looking at various federated solutions for improving resource utilization and task execution efficiency.
In this thesis, multiple algorithms for optimally distributing tasks among clouds in a federation are developed and tested using the CloudSim simulation framework. These algorithms attempt to minimize the global execution time of a series of tasks submitted to the cloud federation using task specific data and information visible to the federation. Various use case scenarios are developed to mimic potential real-world federation scenarios and data access is limited based on the type of scenario. Test results show that a multilayer perceptron algorithm could distribute diverse workflows within 73% of the calculated makespan execution time on average. These results show that is it possible to create algorithms that can successfully place tasks optimally in federated systems with varying background workloads and resource visibilities, and that resource visibility and task specific data both have a direct impact on task distribution optimization.
CSE Speaker Series – Patrick Bridges
On Friday, November 11 at 11:00 am in Cramer 221, Dr. Patrick Bridges from UNM will discuss Integrating Performance Modeling into Scalable System Software Design.
Abstract: Developing new scalable system software techniques is essential to the success of emerging large-scale scientific computing systems due to the increasing scale and complexity of hardware, programming systems, and applications. In particular, HPC operating systems and middleware must address challenges in areas such as fault tolerance, scheduling, synchronization, power management, and high-speed communication. Interactions between these areas also complicate software design; recent research has shown, for example, that both power capping and asynchronous checkpointing can have widely varying and hard-to predict impacts on system performance.
Because of these challenges, my research has increasing relied on performance modeling to expose research challenges, quantify performance tradeoffs, and evaluate the resulting system. This aspect of the research is challenging and rewarding because it requires understanding the underlying system, the strengths and limitations of different modeling approaches developed by the modeling community, and how to best integrate these techniques into system software design. In some cases, my students and I have been able to use simple analytical models; recently, however, we have recently been relying on more sophisticated stochastic modeling techniques. We have also begun exploring the viability of using large-scale computational models to inform the design of HPC system software.
In this talk, I discuss several systems research projects my students and I have conducted to meet HPC system software challenges in the areas of resilience, scheduling, and communication system design. In each of these areas, I describe both the research itself and how modeling techniques have informed the research. Finally, I will briefly discuss some new research directions we are currently exploring as well as provide some thoughts on the broader integration of modeling and evaluation in computer systems research and education.
CSE Speaker – Jorge Hernandez
On Friday, November 18 at 11:00 am in Cramer 221, Mr. Jorge Hernandez from Sandia will give a talk on U.S. Manufacturing and the Impact of Artificial Intelligence and the Digital Worker.
Abstract: U.S. Manufacturing is on a resurgence in terms of global relevance. What impact can artificial intelligence and advances in the digital workforce have in helping transform legacy manufacturing processes into innovative, agile, factories of the future? Considerations include the evolution of the role of factory workers and processes by technologies such as AI, machine learning, and augmented reality, to name a few. Such advanced are intended to enable a exponential effect on quality, output, speed, and personalization “value-add” to U.S. products. This brief presentation will provide an overview of the manufacturing scene and potential visions for the future of U.S. Manufacturing, leading to some insights on the opportunities for the discipline of computer science.