Categories
Papers Publications Uncategorized

new conference paper “Efficient Processing of Streaming Data using Multiple Abstractions

We have published a new paper “Efficient Processing of Streaming Data using Multiple Abstractions” at the IEEE Cloud 2021 conference. (to be available at https://conferences.computer.org/cloud/2021/

We show that one framework can efficiently support multiple abstractions. We provide three abstractions of Block, Windowed, and Stateful streaming and demonstrate that many application classes can be developed with ease, correctness, and low processing latency.

From the abstract of our paper:

Large websites and distributed systems employ sophisticated analytics to evaluate successes to celebrate and problems to be addressed. As analytics grow, different teams often require different frameworks, with dozens of packages supporting with streaming and batch processing, SQL and no-SQL. Bringing multiple frameworks to bear on a large, changing dataset often create challenges where data transitions—these impedance mismatches can create brittle glue logic and performance problems that consume developer time. We propose Plumb, a meta-framework that can bridge three different abstractions to meet the needs of a large class of applications in a common workflow. Large-block streaming (Block-Streaming) is suitable for single-pass applications that care about the temporal and spatial locality. Windowed-Streaming allows applications to process a group of data and many reductions. Stateful-Streaming enables applications to keep a long-term state and always-on behavior. We show that it is possible to bridge abstractions, with a common, high-level workflow specification, while the system transitions data batch processing and block- and record-level streaming as required. The challenge in bridging abstractions is to minimize latency while allowing applications to select between sequential and parallel operation, while handling out-of-order data delivery, component failures, and providing clear semantics in the face of missing data. We demonstrate these abstractions evaluating a 10-stage workflow of DNS analytics that has been in production use with Plumb for 2 years, comparing to a brittle hand-built system that has run for more than 3 years.

This conference paper is joint work of Abdul Qadeer and  John Heidemann from USC/ISI.

Plumb is open source software and will be available at: https://ant.isi.edu/software/plumb/index.html

Categories
Uncategorized

congratulations to Abdul Qadeer for his PhD

I would like to congratulate Dr. Abdul Qadeer for defending his PhD at the University of Southern California in March 2021 and completing his doctoral dissertation “Efficient Processing of Streaming Data in Multi-User and Multi-Abstraction Workflows”.

From the abstract:

Abdul Qadeer after his defense.

Ever-increasing data and evolving processing needs force enterprises to scale-out expensive computational resources to prioritize processing for timely results. Teams process their organization’s data either independently or using ad hoc sharing mechanisms. Often different users start with the same data and the same initial stages (decrypt, decompress, clean, anonymize). As their workflows evolve, later stages often diverge, and different stages may work best with different abstractions. The result is workflows with some overlap, some variations, and multiple transitions where data handling changes between continuous, windowed, and per-block. The system processing this diverse, multi-user, multi-abstraction workflow should be efficient and safe, but also must cope with fault recovery.

Analytics from multiple users can cause redundant processing and data, or encounter performance anomalies due to skew. Skew arises due to static or dynamic imbalance in the workflow stages. Both redundancy and skew waste compute resources and add latency to results. When users bridge between multiple abstractions, such as from per-block processing to windowed processing, they often employ custom code. These transitions can be error prone due to corner cases, can easily add latency as an inefficiency, and custom code is often a source of errors and maintenance difficulty. We need new solutions to manage the above challenges and to expose opportunities for data sharing explicitly. Our thesis is: new methods enable efficient processing of multi-user and multi-abstraction workflows of streaming data. We present two new methods for efficient stream processing—optimizations for multi-user workflows, and multiple abstractions for application coverage and efficient bridging.

These algorithms use a pipeline-graph to detect duplication of code and data across multiple users and cleanly delineate workflow stages for skew management. The pipeline-graph is our job description language that allows developers to specify their need easily and enables our system to automatically detect duplication and manage skew. The pipeline-graph acts as a shared canvas for collaboration amongst users to extend each other’s work. To efficiently implement our deduplication and skew management algorithms, we present streaming data to processing stages as fixed-sized but large blocks. Large-blocks have low meta-data overhead per user, provide good parallelism, and help with fault recovery.

Our second method enables applications to use a different abstraction on a different workflow stage. We provide three key abstractions and show that they cover many classes of analytics and our framework can bridge them efficiently. We provide Block-Streaming, Windowed-Streaming, and Stateful-Streaming abstractions. Block-Streaming is suitable for single-pass applications that care about temporal or spatial locality. Windowed-Streaming allows applications to process accumulated data (time-aligned blocks to sync with external information) and reductions like summation, averages, or other MapReduce-style analytics. We believe our three abstractions allow many classes of analytics and enable processing of one block, many blocks, or infinite stream. Plumb allows multiple abstractions in different parts of the workflow and provides efficient bridging between them so that users could make complex analytics from individual stages without worrying about data movement.

Our methods aim for good throughput, low latency, and clean and easy-to-use support for more applications to achieve better efficiency than our prior hand-tuned but often brittle system. The Plumb framework is the implementation of our solutions and a testbed to validate them. We use real-world workloads from the B-Root DNS domain to demonstrate effectiveness of our solutions. Our processing deduplication increases throughput up to $6\times$, reduces storage by 75%, as compared to their pre-Plumb counterparts. Plumb reduces CPU wastage due to structural skew up to half and reduces latency due to computational skew by 50%. Plumb has cut per-block latency by 74% and latency of daily statistics by 97%, while reducing code size by 58% and lowering manual intervention to handle problems by 73% as compared to pre-Plumb system.

The operational use of Plumb for the B-Root service provides a multi-year validation of our design choices under many traffic conditions. Over the last three years, Plumb has processed more than 12PB of DNS packet data and daily statistics. We show that our abstractions apply to many applications in the domain of networking big-data and beyond.

Categories
Data

fighting bit rot in log-term data archives with babarchive

As part of research at ANT we generate a lot of data, and our goal is to keep it safe even in the face of an imperfect world of data storage.

When we say a lot, we mean hundreds of terabytes: As of May 2020, we have releasable 860 datasets making up 134 TB of storage (510TB if we uncompressed it). We provide this data at no cost to researchers, and since 2008 we’ve provided 2049 datasets (338 TB, or 1.1PB if uncompressed!) to 406 researchers!

These datasets range from packet captures of “normal” traffic, to curated captures of DDoS attacks, as well as dozens of research paper-specific datasets, 16 years of Internet censuses and 7 years of Internet outages, plus target lists for IPv4 that are regularly used for traffic studies and tools like Verfploeter anycast mapping.

As part of keeping this data, our goal is to keep this data. We want to fight bit rot and data loss. That means the RAID-6 for primary storage, with monitoring and timely disk replacement. It means off site backup (with a big thanks to our collaborators at Colorado State University, Christos Papadopoulos, Craig Partridge, and Dimitrios Kounalakis for their help). And it means watching bits to make sure they don’t spontaneously change.

One might think that bits at rest stay at rest, but… not always. We’ve seen three times when disks have spontaneously changed a byte over the last 20 years. In 2011 and 2012 I had bit flips on my personal files, and in 2020 we had a byte flip on a packet capture.

How do we know? We have application-level checksums of every file, and every day we take 10 minutes to check at least one dataset against its checksums. (Over time, we cover all datasets and then start all over.)

Our checksumming software is babarchive–our own wrapper around collecting SHA-256 checksums over a directory tree. We encourage other researchers interested in long-term data curation to carry out active content monitoring (in addition to backups and RAID).

A huge thanks to our research sponsors: DHS (through the LANDER, LACREND, and LACANIC projects), NSF (through the MADCAT, MR-Net), and DARPA (through GAWSEED).

Categories
Students

congratulations to Calvin Ardi for his new PhD

I would like to congratulate Dr. Calvin Ardi for defending his PhD in April 2020 and completing his doctoral dissertation “Improving Network Security through Collaborative Sharing” in June 2020.

From the abstract:

Calvin Ardi and John Heidemann (inset), after Calvin filed his PhD dissertation.

As our world continues to become more interconnected through the
Internet, cybersecurity incidents are correspondingly increasing in
number, severity, and complexity. The consequences of these attacks
include data loss, financial damages, and are steadily moving from the
digital to the physical world, impacting everything from public
infrastructure to our own homes. The existing mechanisms in
responding to cybersecurity incidents have three problems: they
promote a security monoculture, are too centralized, and are too slow.


In this thesis, we show that improving one’s network security strongly
benefits from a combination of personalized, local detection, coupled
with the controlled exchange of previously-private network information
with collaborators. We address the problem of a security monoculture
with personalized detection, introducing diversity by tailoring to the
individual’s browsing behavior, for example. We approach the problem
of too much centralization by localizing detection, emphasizing
detection techniques that can be used on the client device or local
network without reliance on external services. We counter slow
mechanisms by coupling controlled sharing of information with
collaborators to reactive techniques, enabling a more efficient
response to security events.


We prove that we can improve network security by demonstrating our
thesis with four studies and their respective research contributions
in malicious activity detection and cybersecurity data sharing. In
our first study, we develop Content Reuse Detection, an approach to
locally discover and detect duplication in large corpora and apply our
approach to improve network security by detecting “bad
neighborhoods” of suspicious activity on the web. Our second study
is AuntieTuna, an anti-phishing browser tool that implements personalized,
local detection of phish with user-personalization and improves
network security by reducing successful web phishing attacks. In our
third study, we develop Retro-Future, a framework for controlled information
exchange that enables organizations to control the risk-benefit
trade-off when sharing their previously-private data. Organizations
use Retro-Future to share data within and across collaborating organizations,
and improve their network security by using the shared data to
increase detection’s effectiveness in finding malicious activity.
Finally, we present AuntieTuna2.0 in our fourth study, extending the proactive
detection of phishing sites in AuntieTuna with data sharing between friends.
Users exchange previously-private information with collaborators to
collectively build a defense, improving their network security and
group’s collective immunity against phishing attacks.

Calvin defended his PhD when USC was on work-from-home due to COVID-19; he is the second ANT student with a fully on-line PhD defense.

Categories
Papers Publications

new paper “Precise Detection of Content Reuse in the Web” to appear in ACM SIGCOMM Computer Communication Review

We have published a new paper “Precise Detection of Content Reuse in the Web” by Calvin Ardi and John Heidemann, in the ACM SIGCOMM Computer Communication Review (Volume 49 Issue 2, April 2019) newsletter.

From the abstract:

With vast amount of content online, it is not surprising that unscrupulous entities “borrow” from the web to provide content for advertisements, link farms, and spam. Our insight is that cryptographic hashing and fingerprinting can efficiently identify content reuse for web-size corpora. We develop two related algorithms, one to automatically discover previously unknown duplicate content in the web, and the second to precisely detect copies of discovered or manually identified content. We show that bad neighborhoods, clusters of pages where copied content is frequent, help identify copying in the web. We verify our algorithm and its choices with controlled experiments over three web datasets: Common Crawl (2009/10), GeoCities (1990s–2000s), and a phishing corpus (2014). We show that our use of cryptographic hashing is much more precise than alternatives such as locality-sensitive hashing, avoiding the thousands of false-positives that would otherwise occur. We apply our approach in three systems: discovering and detecting duplicated content in the web, searching explicitly for copies of Wikipedia in the web, and detecting phishing sites in a web browser. We show that general copying in the web is often benign (for example, templates), but 6–11% are commercial or possibly commercial. Most copies of Wikipedia (86%) are commercialized (link farming or advertisements). For phishing, we focus on PayPal, detecting 59% of PayPal-phish even without taking on intentional cloaking.

Categories
Announcements Projects

new project “Global Analysis of Weak Signals for Enterprise Event Detection” (GAWSEED)

We are happy to announce a new project, Global Analysis of Weak Signals for Enterprise Event Detection (GAWSEED).  GAWSEED project is studing weak signals across multiple large-enterprise datasets looking for signs of malicious activity so small they may be passed over by a single enterprise’s operational staff. More details are on the GAWSEED project web page.

GAWSEED is part of ANT Lab at USC/ISI (PIs: John Heidemann and Wes Hardaker in the networking division, and Aram Galystan from the AI division. It is joint work with researchers at PARSONS Corporation. It is supported by DARPA as part of the CHASE program.