index.html

---
layout: default
---

<div class="header-container jumbotron">
    <div class="container">
        <center>
            <h1 style="color:black">Hyperloop</h1>
            <p style="color:black; font-size:24px">Solving performance issues for your web application</p>
        </center>
        <!-- <p><a class="btn btn-primary btn-lg" href="{{ "/docs/home/" | prepend: site.baseurl }}" role="button">Learn more</a></p> -->
    </div>
</div>


<div class="container">

    <div class="row">
        <div class="col-md-12">
            <h2 class="header-light regular-pad">What is Hyperloop?</h2>
            <!-- <blockquote> -->
              <p class="main-page-content">Hyperloop is an ongoing research project that aims at detecting and solving 
                    performance problems for web applications
                    built with ORM (Object-Relational Mapping) frameworks, 
                    mainly Ruby on Rails. Currently it contains:</p>
                    <ul class="main-page-content">
                        <li>A comprehensive study on existing open-source applications built with Ruby-on-Rails
                            (with a <a href='https://docs.google.com/forms/d/e/1FAIpQLScyfauBa8YgnlumPU-6YW5x5s9xvoU9cdBXysE_EUmslIcDFA/formResponse'>replication package</a>
                            of these applications).

                        </li>
                        <li><a href="docs/home/">Instructions and benchmarks</a> for reproducing the 
                            common performance inefficiencies found in our study.
                        </li>
                        <li><a href="powerstataion/">PowerStation</a>, a RubyMine plugin to identify performance bugs in
                            Ruby on Rails applications.
                        </li>
                        <li><a href="panorama/">Panorama</a>, a view-centric optimizer for Ruby on Rails applications.
                        </li>
                            
                    </ul>
                <p class="main-page-content">
                    We are working on providing better solutions to solve more inefficiencies
                    and more handy tools for application developers.
                </p>
            
             <h2 class="header-light regular-pad">Project Goals</h2>
            <!-- <blockquote> -->
             <p class="main-page-content">
                 From banking to online shopping, data-driven applications that comprise of a database management system (DBMS) and an application that interacts with the DBMS are pervasive in our daily lives and the scientific community. Needless to say, the performance of data-driven applications is crucial. Unfortunately, due to the quickly increasing data size and ever-changing operational environment of the data center that host such applications, many data-driven applications suffer from performance problems. Techniques and tools that can help identify existing performance problems, predict future performance problems that will arise with workload and operational environment changes, and fix these problems in data-driven applications are highly desired in the current age of big data.
             </p>
             <p class="main-page-content">
                 Unfortunately, although previous work in programming systems and database research communities have studied performance problems and query execution-time prediction, they are insufficient to address the performance problems that arise in data-driven applications. For instance, prior work in programming systems focus on optimizing applications, and treat the DBMS as a black box with unknown performance characteristics. Meanwhile, research in the database community focuses on analyzing query performance, and is oblivious to the fact that many queries are programmatically generated by applications with predictable structures.
             </p>
            
                        
             <h2 class="header-light regular-pad">Research Challenges</h2>
            <!-- <blockquote> -->
             <p class="main-page-content">
                 In this project, we argue that such “black box” approach forgoes many opportunities in identifying, and fixing performance problems in data-driven applications. Our key insight is to regard the DBMS and the application as white boxes and devise new opportunities to optimize both systems simultaneously. First, we plan to use various program analysis techniques to understand the semantics of data-driven applications. We will devise new program analysis techniques to study the structure and sources of parameter values for the queries that are issued by the application, the query plans that the DBMS use to process such queries, and how the query results are subsequently used by the application.
             </p>
             <p class="main-page-content">
                Next, we will use such information to identify performance problems in data-driven applications. Leveraging the end-to-end performance modeling of the DBMS and the application in data-driven applications, we will design new program analysis algorithms, performance testing techniques, and performance debugging tools to identify workload and code fragments in the application that might cause performance problems.
             </p>
             <p class="main-page-content">
                We will evaluate our techniques using real-world data-driven applications such as web applications that store persistent data in DBMSs.
             </p>
            
              <h2 class="header-light regular-pad">Broader Impacts</h2>
            <!-- <blockquote> -->
             <p class="main-page-content">
                 The project addresses one of the most important and fundamental challenges of the big data era, namely the performance of data-driven applications. It will broadly benefit other science disciplines and the society by improving the performance of data-driven applications. We will leverage their expertise in in database systems, program analysis, performance debugging, and software engineering to tackle the challenges. Furthermore, we have extensive connections with researchers and industry practitioners at Intel, Huawei, Google, and Microsoft, and will make use of such connections to validate their research findings.
             </p>
             <p class="main-page-content">
                  We plan to release all software and benchmarks that are developed to public domain. The techniques developed in this project will be incorporated into graduate-level and undergraduate-level courses in data management systems and programming systems offered at the UC Berkeley.
             </p>
            <!-- </blockquote> -->
        </div>
        <!-- <div class="col-md-6 text-center">
            <img src="img/jekyll-dark.png" alt="" class="img-responsive">
        </div> -->
    </div>

    <div class="row">
        <div class="col-md-12">
        <h2 class="header-light regular-pad">Publications</h2>
        <ul class="list-publication">
           <li>
                <span class="title">
                    <a href="http://cidrdb.org/cidr2020/papers/p21-yan-cidr20.pdf">View-Driven Optimization of Database-Backed Web
Applications.</a>
                </span>
                <span class="authors">
                    Cong Yan,
                    Junwen Yang,
                    Alvin Cheung and Shan Lu,
                </span>
                <span class="description">
                    Conference on Innovative Data Systems Research (CIDR'20)
                </span>                                                                           
            </li>
               <li>
                <span class="title">
                    <a href="278.pdf">Managing data constraints in database-backed web applications.</a>
                </span>
                <span class="authors">
                    Junwen Yang,
                    Utsav Sethi,
                    Cong Yan,
                    Shan Lu and Alvin Cheung,
                </span>
                <span class="description">
                    42th International Conference on Software Engineering (ICSE'20)
                </span>
                <span style="max-width:30px"> 
                            <img src="img/red.jpg" alt="" style="max-width:80px"> <img src="img/green.jpg" alt="" style="max-width:80px">
                </span>                          
                                                     
            </li>
            <li>
                <span class="title">
                    <a href="https://homes.cs.washington.edu/~congy/chestnut.pdf">Generating Application-Specific Data Layouts for In-memory Databases.</a>                    
                </span>
                <span class="authors">
                    Cong Yan and Alvin Cheung
                </span>
                <span class="description">
                    45th International Conference on Very Large Databases (VLDB'19)
                </span>
            </li>
            
            <li>
                <span class="title">
                    <a href="panorama.pdf">View-Centric Performance Optimization for Database-Backed Web Applications.</a>
                    <font color="maroon"><i>SIGSOFT Distinguished Paper Award</i></font> <i style='font-size:24px;color:red' class='fas'>&#xf5a2;</i>
                </span>
                <span class="authors">
                    Junwen Yang,
                    Cong Yan,
                    Chengcheng Wan,
                    Shan Lu and Alvin Cheung,
                </span>
                <span class="description">
                    41th International Conference on Software Engineering (ICSE'19)
                </span> 
                <span style="max-width:30px"> 
                            <img src="img/red.jpg" alt="" style="max-width:80px"> <img src="img/green.jpg" alt="" style="max-width:80px">             
                </span> 
            </li>
            <li>
                <span class="title">
                    <a href="./powerstation.pdf">PowerStation: Automatically detecting and fixing inefficiencies of database-backed web applications in IDE.</a>
                </span>
                <span class="authors">
                    Junwen Yang,
                    Cong Yan,
                    Pranav Subramaniam,
                    Shan Lu and Alvin Cheung,
                </span>
                <span class="description">
                    26th Foundations of Software Engineering (FSE'18 Demostration Track)
                </span>
            </li>
            <li>
                <span class="title">
                    <a href="./220-HowNotStructure.pdf">How not to structure your database-backed web applications: a study of performance bugs in the wild.</a>
                </span>
                <span class="authors">
                    Junwen Yang,
                    Cong Yan,
                    Pranav Subramaniam,
                    Shan Lu and Alvin Cheung,
                </span>
                <span class="description">
                    40th International Conference on Software Engineering (ICSE'18)
                </span>
            </li>
            <li>
                <span class="title">
                    <a href="./study_db.pdf">Understanding Database Performance Inefficiencies in Real-world Web Applications.</a>
                </span>
                <span class="authors">
                    Cong Yan,
                    Junwen Yang,
                    Alvin Cheung and Shan Lu,
                </span>
                <span class="description">
                    26th Conference on Information and Knowledge Management (CIKM'17)
                </span>
            </li>
            <li>
                <span class="title">
                    <a href="http://db.cs.
                             .edu/projects/quro/">Leveraging Lock Contention to Improve OLTP Application Performance.</a>
                </span>
                <span class="authors">
                    Cong Yan and  Alvin Cheung
                </span>
                <span class="description">
                    42nd Conference on Very Large Data Bases (VLDB'16)
                </span>
            </li>

        </ul>
        </div>
    </div>
      <div class="row">
        <div class="col-md-12">
        <h2 class="header-light regular-pad">News</h2>
        <ul class="list-publication">
            
             <li>
                <span class="title">
                    <a href="https://computerscience.uchicago.edu/news/article/panorama-slowdowns/">University of Chicago news release: Award-Winning ICSE Paper Offers Web Developers a “Panorama” View of Slowdowns and Fixes</a>
                </span>
                </span>
            </li>
            <li>
                 <span class="title">
                    Panorama is featured on Morning paper! Please check  
                    <a href="https://blog.acolyer.org/2019/07/12/view-centric-performance-optimization/">morning paper post</a> for details.
                </span>
            </li>
            <li>
                <span class="title">
                    Panorama has been released! Please check the  
                    <a href="https://github.com/hyperloop-rails/panorama-static-analyzer">repository</a> for details.
                </span>
            </li>
            <li>
                <span class="title">
                    PowerStation 1.0 has been released! Please check our 
                    <a href="https://medium.com/@uwdb/introducing-powerstation-26c8b3a53191">blog post</a> for details.
                </span>
            </li>
            <li>
                 <span class="title">
                    Our emprical study is featured on <a href="https://rubyweekly.com/issues/406">RubyWeekly</a>.
                </span>
            </li> 
            <li>
                <span class="title">
                    <a href="https://cs.uchicago.edu/news/new-study-finds-and-fixes-flaws-database-backed-web-applications">University of Chicago news release: New Study Finds and Fixes Flaws in Database-Backed Web Application</a>
                </span>
                </span>
            </li>
            <li>
                <span class="title">
                    <a href="https://blog.acolyer.org/2018/06/28/how-_not_-to-structure-your-database-backed-web-applications-a-study-of-performance-bugs-in-the-wild/">Morning Paper blog post</a>
                </span>
                </span>
            </li>
            <li>
                <span class="title">
                    <a href="https://news.ycombinator.com/item?id=17414383">Story on Hacker News: How not to structure database-backed web apps: performance bugs in the wild</a>
                </span>
                </span>
            </li>
        </ul>
        </div>
    </div>
    
  <div class="row">
        <div class="col-md-12">
        <h2 class="header-light regular-pad">Acknowledgement</h2>
            <p class="main-page-content">  
                This material is based upon work supported by the National Science Foundation under Grant No. 
                <a href="http://nsf.gov/awardsearch/showAward?AWD_ID=1546083">IIS-1546083</a> (2016-2019) and No.<a href="https://www.nsf.gov/awardsearch/showAward?AWD_ID=1546543">IIS-1546543</a> (2016-2019), "Holistic Optimization of Data-Driven Applications,"
                
                and is a collaborative effort between <a href="http://dsf.berkeley.edu">UC Berkeley</a> and the <a href="http://systems.cs.uchicago.edu">University of Chicago</a>.
            </p>
            
            <p class="main-page-content">
                Any opinions, findings, and conclusions or recommendations expressed in this
                material are those of the author(s) and do not necessarily reflect the views of
                the National Science Foundation.</span>
            </p>

            <p class="main-page-content">
                Last updated: Mar 2020
            </p>
    </div>
    
    <hr>
    
    <!-- <div class="row">
        <div class="col-md-12">
        <h2 class="header-light regular-pad">Replication Packages</h2>
        <p style="font-size: 20px">Click <a href='https://docs.google.com/forms/d/e/1FAIpQLScyfauBa8YgnlumPU-6YW5x5s9xvoU9cdBXysE_EUmslIcDFA/formResponse'>here</a>
        to get the replication packages for our studies.</p>
        </div>
    </div> -->


</div>

</body>
</html>