|
|
Day 1, Sunday June 30, 201309:30 Registration 10:15 Welcome
10:30 Keynote #1:
Fighting Internet-based Sexual Exploitation Crimes Against Children
Abstract:
Working with law enforcement and criminologists, my research group has been developing strategies to fight online child sexual exploitation. Investigators in the US and other countries have been using our Internet-based tools daily, acquiring evidence that has resulted in many thousands of arrests for child pornography possession since 2009. Over 100 children have been rescued from abuse as a result of these cases in the last 12 months.
Using data and results from our project, I will provide an analysis of our view of the massive Internet-based child pornography trade. And I will illustrate our approach to problems in network-based criminal investigation. Our work demonstrates that even if "fully secured", the Internet would still play a role aiding and abetting crimes that harm persons. The project stands in contrast to security research focused on solving technical flaws that expose data or allow access, and to research where the victim can be protected with technical solutions. Addressing problems beyond technical flaws requires an interdisciplinary approach, as we have taken. 11:30 Break
Session 1: Virtualization 12:00
HSG-LM: Hybrid-Copy Speculative Guest OS Live Migration without Hypervisor,
12:25
Octopus: Efficient Data Intensive Computing on Virtualized Datacenters,
12:50
Efficient and Scalable Paravirtual I/O System
(Highlight paper, USENIX ATC'13),
13:15 Lunch break
Session 2: Metadata and Search
14:30
Examining Extended and Scientific Metadata for Scalable Index Designs,
14:45
Direct Lookup and Hash-Based Metadata Placement for Local File Systems,
15:10
Mercury: Bringing Efficiency to Key-value Stores,
15:35
A Self-Managed Self-Optimized Publish-Subscribe System,
15:50 Break
Session 3: Compute 16:20
GPUfs: Integrating a File System with GPUs (Highlight paper,
ASPLOS13),
16:45
Computational Caches, 17:00 Poster Session and Beer Garden 18:30 End of Day 1 Day 2, Monday July 1, 201308:30 Registration
09:00 Keynote #2:
Spanner: Google's Globally-Distributed Database
Abstract:
Spanner is Google's scalable, multi-version, globally-distributed, and synchronously-replicated database. It provides strong transactional semantics, consistent replication, and high performance reads and writes for a variety of Google's applications. I'll discuss the design and implementation of Spanner, as well as some of the lessons we have learned along the way. I'll also discuss some open challenges that we still see in building scalable distributed storage systems.
10:00 Break
Session 4: Flash Optimization
10:30
HEC: Improving Endurance of High Performance Flash-based Cache Devices,
10:55
Extending SSD Lifetime in Database Applications with Page Overwrites,
11:20
An Empirical Study of Hot/Cold Data Separation Policies in Solid State
Drives (SSDs),
11:35 Break
Session 5: De-duplication
12:00
Fuzzy Adaptive Control for Heterogeneous Tasks in High-Performance
Storage Systems,
12:25
Rangoli: Space management in Deduplication Environments,
12:40
Block Locality Caching for Data Deduplication,
13:05
A Scalable De-duplication and Garbage Collection Engine for
Incremental Backup,
13:30 Lunch Box 14:00 Tour & Dinner 21:00 End of Day 2 Day 3, Tuesday July 2, 201308:30 Registration
Session 6: Storage
09:00
Virtual Point in Time Access,
09:25
Performance Introspection of Graph Databases,
09:50
Leveraging Predefined Huffman Dictionaries for High Compression Rate
and Ratio,
10:05
Building Intelligence for Software Defined Data Centers: Modeling
Usage Patterns, 10:30 Break
Session 7: Storage with Flash
11:00
Beyond Block I/O: Implementing a Distributed Shared Log in Hardware,
11:25
Linux Block IO: Introducing Multi-queue SSD Access on Multi-core Systems, 11:50 Break 12:10 Keynote #3:
Big Data -- Little Software?
Abstract:
The steps of accessing, storing, and transmitting "Big Data" raise many interesting problems. But big data sets also amplify any system or software inefficiencies when large data sets require processing. So the efficiency of the generate code (and the runtime system) is crucial if we want to see widespread use of applications based on big data. Adaptive software exploits platform and data properties to custom-tailor program executions to the current environment. However, modern platforms have many features that make it difficult to support adaptive software. Multi-core systems with a non-uniform memory architecture expose various asymmetries and complicate the runtime system's task of data management, yet even modest multi-processors exhibit NUMA properties. Processor features like prefetchers are difficult to model by a compiler and may influence the execution in unexpected ways. Finally, performance monitoring units are supposed to allow a (just-in-time) compiler to obtain the information needed to adapt the generated code. But current performance monitoring units are incomplete and, worse, subject to change over time. An adaptive software system needs performance data that is readily available, reliable, and stable. In this talk I will discuss our experiences in modeling modern systems and argue for portable performance monitoring units that allow higher levels of the software tool chain to rely on live performance data. 13:20 Concluding remarks 13:30 Lunch 14:30 End of Day 3 Poster Session - Sunday June 30 at 5pm#1 Openstack Swift Multitenancy #2 An Analysis of Performance Prediction in Storage Systems #3 Deduplication with Block-Level Content-Aware Chunking for Solid State Drives (SSDs) #4 Towards Operating System Support for Heterogeneous-ISA Platforms #5 Data on-boarding in federated storage clouds #6 BigGraph: High-performance Distributed RDF Store #7 QUALITY OF SERVICE GUARANTEE FOR CLOUD-SCALE STORAGE SYSTEMS #8 Exploring Checkpointing and Closed Nesting in Distrubuted Transactional Memory #9 Evaluation of Inline Deduplication for Primary Block Storage #10 Close-to-Linear Performance Aggregation with Parallel NFS #11 Virtual Remote I/O (vRIO) #12 Combining ErasureCode and Replication Redundancy Schemes for Increased Storage and Repair Efficiency in P2P Storage Systems #13 Grid2Go: A Web Based Internet of Things GRID Computing Engine #14 Predicate Abstraction for Relaxed Memory Models #15 A method for hiding decryption keys in copy protected software #16 VSwap: Agile Page Swapping in Virtualization Environments #17 Workload Resampling for Performance Evaluation of Parallel Job Schedulers #18 Self-contained Information Retention Format (SIRF) in ForgetIT EU Project #19 Ring-IOMMU Efficient IOMMU Address Translation #20 Case Study: National Grid-Cloud Infrastructure as a key enabler of innovation in Kazakhstan #21 A Parallel Twig Join Algorithm for XML Processing using a GPGPU #22 Mitigating VM sprawl effects with adaptive over-commit #23 Efficient Use of Geographically Spread Cloud Resources |
 |