Personal tools
You are here: Home Events CETI Applied Research Colloquium Summer Quarter 2011

CETI Applied Research Colloquium Summer Quarter 2011

When Jul 08, 2011 01:20 AM to
Sep 21, 2011 01:20 AM
Where CL 400,
Contact Name
Contact Phone 614-292-9358
Add event to calendar vCal

July 8,2011

  • Presentation by iBrutus group(Tom Lynch,Hareendra Manuru,Rajagopal Vasudevan,Ashok Sasidharan)

 "Microsoft Kinect Drivers Features and Limitations"

July 15,2011

  • Presentation by Aishwarya Venkatraman

Cloud service providers make virtual desktop resource provisioning decisions based in users and their application profiles. The aim of such provisioning is to provide at least an acceptable level of user experience that meets the service level agreement (SLA) while maximizing the number of users supported.

Current generation resource allocation schemes based on utility, do this in two steps,(i) Placement—deciding the location of the datacenter where the user will be placed; (ii) Provisioning—deciding the amount of resources to be allocated to the user at that datacenter. This two-step approach results in sub-optimal placement and resource fragmentation due to the departure of the VDs.

In this paper, we consider the combined problem of making placement and provisioning (P&P) in an optimal way, such that the utility of the overall distributed data center system is maximized, particularly focusing on reducing resource fragmentation. We address two specific problems, one, what is the right way to do allocation, if migration is not desirable or possible?. Second, if migration is possible, which VDs to migrate and to where?. The first we call the Migration-free Utility-optimal Placement & Provisioning (MUPP) problem and the second we call the Defrag problem. We provide a Migration-free Utility-optimal Placement & Provisioning (MUPP) algorithm, makes P&P for the newly arriving VDs, without disturbing the placement of the current VDs in the system. While the MUPP decreases resource fragmentation to a large extent it does not completely eliminate it, which we address in the Defrag problem. Our Global Utility-optimal Placement & Provisioning (GUPP) algorithm to finds the global near-optimal P&P, using which we subsequently defragment the resources, i.e., migrate a set of VDs to new locations, which will improve their user experience, at the same time improving the resource utilization for the CSPs.

  • Presentation by Chaitanya Shivde

We have proposed a methodology that established traceability between service requests received at a call center and various business components that contribute towards its completion. The methodology uses Activity Based Costing and proposes the use of heat maps derived from the  Component Business Model from IBM and as a dashboard for continuous process monitoring. We illustrate our model as a possible solution towards service improvement for the 311 service at The City of Columbus.
  • Presentation by Zeng Ma

Designing a Pulse Doppler Radar. In tradition, most of the Radar detects the target by detecting the displacement moving.So if some "target" are moving back and forth just like the tree in the wind, it will realize it as an target by mistake.  But my Radar is quite different, it detects the phase of the Radar wave. In this way, we can avoid the problem. But there is some problem of the radar. Because the the detector(FPGA), can sample at the very phase accurately. That will result in some interference wave. we have to use a digital filter to filter it out. So my task is to verify the radar and design the filter and reduce the power consumption.

July 22,2011

  • Presentation by Da Zhang(Belinda)
Use Nagios and Selenium together to monitor the function of a certain website.  Nagios is a powerful monitoring system that enables organizations to identify and resolve IT infrastructure problems before they affect critical business processes. Nagios monitors your entire IT infrastructure to ensure systems, applications, services, and business processes are functioning properly. Selenium is a suite of tools to automate web browsers across many platforms. In our application, we need to monitor the function of the websites of OSC automatically. In addition, once there is some problems happened, the Nagios should send an email to the administrator.
  • Presentation by Anirban Joy Nandi

Acceptance Test Driven Development is a Software Development approach in which Developers, Testers, Analysts  and other technical and non-technical people collaborate to produce high quality error free software. Here it will be described in brief, how acceptance tests are used to ensure that at every stage of development, the code is written, tested, rewritten and refactored to ensure that it satisfies a set of requirements known as acceptance tests.  This is significantly different from the traditional water-fall approach and leads to less code rejection and hence better use of “man-months”.
  • Presentation by Ankush Srivastava and Satyajeet Raje

The vision is to create a single knowledge resource portal for the clinical and translational research community that would provide a “front door” for a variety of available resources.
These Resources include :
• web resources (Eg: shared services - microarray)
• public databases (Eg: pubmed,
• grid resources (Eg: caTissue data service)
July 29,2011
  • Presentation by Ritu Singh
  • Presentation by Abhijit Shirke
 We focus on automation to improve turn around time and reusable components to enable ease of code maintenance and improve reusability. We are trying to use agile process along with Acceptance Test Driven Development(ATDD) to increase end user experience, reduce development time, phase containment in SDLC phase, and reduce rework. Legacy modernization project at Nationwide Insurance focuses on leveraging latest technologies and methodologies in Software Development Life Cycle (SDLC) process to achieve efficiencies and improving profitability.
  • Thesis dry run by Nived
 Wireless Sensor Networks consist of thousands of nodes operating with severe memory and power constraints. These challenges drive the design on TinyOS. TinyOS has consequently become the platform of choice for sensor network research supporting a host of protocols and libraries. However, theses are systems where application logic and platform logic is interwoven making programmability hard. The .Net Microframework enables programmers to write embedded applications in c# and has an efficent runtime enviroment designed for resource constrained systems. However, the lack of widespread use in wireless sensing has meant lack of support for complex protocols and hence the inability to develop effective applications. Through this technique we aim at synergising the two systems by providing the best of both worlds without loss in performance.

August 5,2011

  • Presentation by Chenjie Wei

"The mobile phone is no longer all that is mobile. The consumer is mobile." Maintaining mobile version assets with limited resources is imperative. We are developing an auto-mobilizer aiming at converting common web pages to a mobile version, which will provide an efficient way to utilize users' mobile devices, more easily to help companies realize their products' mobile values and open their mobile market, and save companies' developing resources. This product will convert a common web page, which is designed for desktop or laptop web browsers, to a much simpler version, which will better utilize all the mobile phone's resources and give out a suitable page looking for mobile device screens. By employing web crawler techniques, it will analyze all links and assets from a certain web page, provide better interaction to let users choose which links they want to convert, what assets, including images, alignments, tables, and flashes, they want to eliminate or keep. 
  • Presentation by Da Cai
Internship Portal
The main feature of this website is that the companies can post opportunities of internship, the students can upload their resumes, search and apply for the internship.
Automatically, reliably and efficiently import data from file server to OSC home directories of users. Have a secure but user-friendly web-interface for users to download their data sets from OSC. Allow users to process the data in their home directories via remote desktops.
  • Presentation by Min Wang
Production Flow Analysis (PFA) is a comprehensive method for material flow analysis, part family formation, design of manufacturing cells, and facility layout design that was developed in the early 70.s.PFAST (Production Flow Analysis and Simplification Toolkit) is a software tool that has automated the manual methods of Production Flow Analysis (PFA). In a typical application, Production Flow Analysis (PFA) is implemented in four stages - Factory Flow Analysis (FFA), Group Analysis (GA), Line Analysis (LA) and Tooling Analysis (TA). Each stage in PFA seeks to eliminate delays in production flows and operational wastes in a progressively smaller area of the factory. In Factory Flow Analysis (FFA), the flows between shops (or buildings) on the factory site are evaluated to eliminate wastes due to transportation, communication delays, use of large containers to store WIP and use of bulk-handling material handling equipment to move the large containers over large distances. In Group Analysis (GA), the flows between machines in each shop within the factory are evaluated to implement manufacturing cells to produce families of parts with identical (or similar) routings. In Line Analysis (LA), the flows between machines in each cell are evaluated. A layout for the cell is designed for efficient inter-machine material handling, multi-machine tending by individual operators and minimum wasted motions by operators. In Tooling Analysis (TA), the flows at each machine in a cell are evaluated to optimize the workstation layout for ergonomics of machine operation and rapid execution of setup activities.
  • Thesis dry rum by Srinivas Hegde
Modern society consumes large amounts of fuel, and the energy industry is a crucial part of infrastructure and maintenance of society in almost all countries. High energy prices, climate change and energy security are converging as the new engine driving the development of clean energy. Experts from green energy industry are trying their best possible way to educate people about the importance of green energy resources, as well as the need to shift from the traditional energy resources to green energy resources. The present world has also been noticing a tremendous rise in green energy industry and a shift in the mindset of the manner of utilizing energy. So, in this research, we present an ideal application from a consumer and producer point of view that provides an interactive way of looking at the energy generation and its usage. We have also created a sample application for both consumer and producer using cyber-infrastructure to provide more information to the user to make important data-driven and intelligent decisions.

August 12,2011

  • Presentation by Subha Balaraman

Software architecture selection is a multi-criteria decision-making problem which involves knowing what the possible choices are and what criteria distinguish the choices.

(a)    With a standard and established SOAP web services, it is valuable to determine why large scale online applications deploy REST web services.

(b)   DataPower is a rack mountable network device and a hardware enterprise service bus from IBM. Built for simplified deployment, hardened security, providing XML acceleration and message transformations at wire speed, it’s interesting to analyze if DataPower can replace the enterprise middleware systems in the current architecture.

(c)    Software testing is a trade-off between budget, time and quality and it is crucial to decide on precise testing tools.

  • Presentation by Vinh Ngoc Khuc

In recent years, social networks have become very popular. Twitter, a microblogging service, is estimated to have about 200 million registered users and these users are creating approximately 65 million tweets a day.  Twitter users usually show their opinion about topics that they are interested. It is very useful for customers who want to have reviews on products that they are going to buy, and for producers who want to measure customer satisfaction of their products. The challenges are that each tweet is limited in 140 characters, and hence very short, and it may contain slang and misspelling words. Thus, it is difficult to apply traditional NLP techniques which are designed for working with formal languages for Twitter posts. Another challenge is that the total volume of tweets is extremely high, and it takes a long time to process. In this presentation, a large-scale distributed system for Twitter sentiment analysis is proposed. The system consists of the following components: lexicon builder, sentiment classifier and parallel processing module based on MapReduce Framework.

  • Presentation by Alhad Sapre
Abstract : 
Time Off Planning Service (TOPS) is a leave management system solution provided by Nationwide Better Health (NBH) to help companies better manage absence policies in their workplace and identify employees in need of health management services. The application has been developed and maintained for around 10 years.  Our job was to analyze and fix some defects in the system while adapting to the Agile Software Development Process.
Devcon 2012
Nationwide’s Application Development Centre holds a developers conference every year at Nationwide. Our objective is to develop a customized content management system which can be used for all such events at Nationwide.  Customization is along two dimensions: firstly to integrate the system with existing tools at nationwide and secondly to accommodate scalability and maintenance in a production environment.
August 19,2011
  • Presentation by Shriram Ramamurthy
This presentation is mainly centered around Network Monitoring. With the huge number of networks and communication systems, which come under the control of OARnet, monitoring them become an essential day to day activity. So plug-ins are required to be written to the Nagios System to make the understanding of operations of the networks more easy. Detecting the uptime of hosts, the throughput control of wireless connections and hardware management of routers are certain issues which were dealt, as a part of Monitoring the Networks.
  • Presentation by Aishwarya Venkatraman
There is a rising trend to take advantage of cloud-computing technologies given their appealing characteristics such as on-demand resource access, and rapid elasticity. Security in the cloud environment is challenging due to the dual nature of the problem. The cloud needs to be open and accessible to users from the outside the network but, if one user is compromised, rest of the cloud should not become vulnerable. There is a new focus in designing clouds, where there is a need include security as one of the primary objectives. In this work we look at designing Mission-oriented Resilient Clouds (MRC) where, the cloud should be able achieve the mission definition, even under security threats. Mission definition might include objectives across multiple dimensions, such as maintaining a minimum level of security, need to optimize Quality of Experience of users
(QoE), resource allocation and scalability. This creates a multi-objective optimization problem where the relationships between the different objectives need to be understood. In this project, we aim at designing a objective function that will maximize the resilience and security, under resource and Quality of user experience
constraints. We also use threat profiles generated by the Threat assessment module as an input to our resource optimization module. Our framework will  also include a cost-benefit analysis, that uses cost of failure as a feedback to the resource optimization module.
  • Presentation by Keith Chima
Motion-tracking is important for a number of modern applications. However, modern motion-tracking systems tend to be expensive.  Cheaper solutions tend to be less effective, and are only used for simple applications, or novelty. Our approach was to design and implement a low-cost motion-tracking system. The scope of this project includes designing, building and programming the system. We chose to implement the system in the form of an automated Nerf gun turret, to demonstrate speed, accuracy, and cost-effectiveness. We will discuss the challenges that we have encountered so far, with particular emphasis on difficulties encountered with the hardware involved. At this juncture, we are shifting our focus somewhat to emphasize the motion-tracking portion of the project. There are a number of challenges involved in implementing an effective, efficient, and relatively inexpensive motion-tracking system with limited resources. We will discuss these challenges,  potential solutions, and realistic goals for the end of the project.

  • Presentation by Zeng Ma

First Abstract:
There is a Doppler radar, which can detect the target moving, but there is some drift interfere the signal. I choose a method to filter the drift by improving the sample rate, filter the data and then down sampling. considering the resource limitation I have to design a multiplier for the filter to save some space, so it can be synthesized.
Second Abstract:Design a harness using FPGA. This harness is used to replace the logic analyzer to detect some specified patterns of the signals from GPIO lines.Different from logic analyzer, we don't care what the signal is, we only care about whether there are the patterns we needed in the signal .

August 26,2011
  • Presentation by Subha Balaraman

Software architecture selection is a multi-criteria decision-making problem which involves knowing what the possible choices are and what criteria distinguish the choices.

(a)    With a standard and established SOAP web services, it is valuable to determine why large scale online applications deploy REST web services.

(b)   DataPower is a rack mountable network device and a hardware enterprise service bus from IBM. Built for simplified deployment, hardened security, providing XML acceleration and message transformations at wire speed, it’s interesting to analyze if DataPower can replace the enterprise middleware systems in the current architecture.

(c)    Software testing is a trade-off between budget, time and quality and it is crucial to decide on precise testing tools.

  • Presentation by Xiaojie Su
"HCI for Large-Scale Collaboration"
Research on HCI has shifted from focusing on system and context to delivering value to end users. This presentation contains HCI requirements analysis, design and implementation of two collaboration systems:
1. Weave, under development and used by Open Indicators Consortium to build visualizations of public data in support of regional planning throughout the nation. This project is sponsored by MORPC (Mid-Ohio Regional Planning Commission) and CRP (Community Research Partner) to build a local version customized to present and manage data in central Ohio.
2. HEZO, innovation action network developed by CETI to provide state-of-the-art solutions for collaboration management, also part of CETI’s continuous research in Extreme-Scale Organizational (ExSO) collaboration.

More information about this event…

« September 2017 »