Winter '20 SoS Quarterly Meeting

2020 Winter Science of Security and Privacy Quarterly Meeting

The 2020 Winter Science of Security and Privacy Quarterly Meeting will be hosted by the NCSU Lablet in the NCSU Department of Computer Science in Raleigh, North Carolina on January 15 and 16.

Registration:

Register to Attend (Register by January 10, 2020)

Program:

DAY 1 - WEDNESDAY, JANUARY 15

DAY 2 - THURSDAY, JANUARY 16

Venue

The meeting will be held at the NCSU Department of Computer Science in Raleigh, North Carolina. The department is located in the Engineering Building II on campus, right off of "the Oval" grounds and the Oval Drive street loop. For information on access to public transit from the NCSU campus, visit here.

Questions:

If you have questions about the meeting, please contact: Laurie Williams (lawilli3 [at] ncsu.edu), Munindar P. Singh (mpsingh [at] ncsu.edu), or Nirav Ajmeri (najmeri [at] ncsu.edu)

Invited Talks

'Deep' Thoughts on Information Sharing 
Neal Ziring (NSA) 
9:15 AM Wednesday, January 15, 2020

The Weakest Link 
Mike Bender (NSA Laboratory for Analytic Sciences) 
1:45 PM Wednesday, January 15, 2020

In this new digital universe, protecting privacy, shielding intellectual property, and securing IT infrastructure are becoming more difficult. The issue has grown more urgent amid new security and privacy laws—along with tougher enforcing of regulatory mandates and stiffer penalties for not complying.

Zero Trust 101: An Evolution in Enterprise Cybersecurity 
Alper A. Kerman, Scott Rose (NIST) 
03:30 PM Wednesday, January 15, 2020

Cybersecurity breaches expose millions of highly sensitive records that contain national security information, personally identifiable, health, and financial information. Traditionally designed, perimeter based enterprise network security models are no longer as effective in shielding against most recent elaborate cybersecurity threats, especially when networked information systems are spread across enterprise and cloud boundaries and often accessed remotely by enterprise mobile workforce in present-day network environments. This has led to a cybersecurity strategy called zero trust that recommends principles and approaches for empowering resilient environments orchestrated specifically for secure resource access based on dynamic policy creation and enforcement. Enterprises create base policies based on business workflows and the principle of least privilege. As the enterprise operates, collected data is used to refine and improve the security posture of the enterprise.  The identified challenges to zero trust include its impact on user behavior and how to use the various feedback systems used in zero trust to improve policies at scale. This talk will cover the concepts of zero trust, its advantages and challenges enterprises face when adopting a zero trust architecture.

Building Trust into the Future  
Anthony Grieco (Cisco) 
9:00 AM Thursday, January 16, 2020

The Internet of the future will require a new level of trust, in the most detailed of places and the most abstract. As technology emanates everywhere, global organizations will need the ability to verify the trust and cyber resilience of new innovations to achieve the speed, agility, and confidence to excel in the digital world. From embedding features into technology to authenticating how solutions are designed, manufactured, and operated. Engineers, manufacturers, operators and users, must continue to create new ways that verify trust. Anthony Grieco, Trust Strategy Officer, Cisco Systems, will share a perspective on the state of trust and resilience of solutions, operations, and processes today, and highlight future challenges in achieving the explicit and systematic trust we will need in the future.      

Privacy in a Decentralized World: Crypto tools for blockchains Applications and Governance 
Alessandra Scafuro (NCSU) 
11:15 AM Thursday, January 16, 2020

Decentralized systems (e.g. blockchain) promise transparency and "trustlessness" via public verifiability. This property,  however, naturally opens the door to potential privacy issues. These are arguably more severe than in a centralized setting, since even advanced crypto tools (such as secure computation or zero-knowledge)  don't necessarily guarantee  public verifiability and hence cannot be directly employed.

In this talk I will present cryptographic primitives that can enable privacy-preserving blockchain governance, and discuss challenges in the definitions and constructions that are unique to the decentralized setting. 

Lablet Talks

Usability Evaluation of the Obsidian Smart Contract Language 
Jonathan Aldrich (CMU) 
10:15 AM Wednesday, January 15, 2020

We designed the Obsidian language with a type system that enforces constraints on how smart contracts are used.  Many similar advanced type systems have struggled with user adoption in the past. We used formative user studies to try to reduce barriers to the usability of our type system.  In this talk, we report on recent results evaluating the usability of the final language design, evaluating how quickly developers can learn Obsidian and whether it helps them avoid making errors when writing smart contracts.

The Attestation Monad - A Principled Architecture for Remote Attestation 
Perry Alexander, Anna Fritz, and Adam Petz (ITTC - The University of Kansas) 
11:00 AM Wednesday, January 15, 2020 

Remote attestation is an emergent technique for establishing trust in remote systems. In its simplest form, an appraiser makes a request of a target, the target responds with evidence and meta-evidence, and the appraiser evaluates the evidence. If we are to scale remote attestation to complex enterprise systems involving thousands or millions of systems more flexible models and scalable infrastructure are required. We will present a policy driven model for remote attestation that supports protocol and service negotiation, distribution of attestation services, and sophisticated attestation architectures. The model centers on the Copland attestation protocol language combining it with a monad that provides sequencing, storage, and failure management. Policies govern both the selection of concrete services for gathering evidence and protocols for performing attestation. The model, called an Attestation Monad, has a formal semantics and emerging high-assurance implementations in CakeML and Haskell.

Discoverability & Attacker Behavior in a Penetration Testing Competition 
Andy Meneely (RIT) 
2:15 PM Wednesday, January 15, 2020

At RIT, we are collecting fine-grained data about how penetration testing competitors work. Using Splunk, we have collected over 9 TB of data across three 10-hour competitions with over 50 teams across those competitions. We have bash history, network metadata, process information, and a variety of other data sources. We are in the process of constructing timelines that provide a second-by-second breakdown of what each team was doing. Using these timelines, we are constructing stochastic models of attacker behavior so that we can estimate the discoverability of a vulnerability in relation to the attack surface. We believe this work can inform vulnerability equities processes as they estimate risk of a vulnerability being discovered.

Analytics for Cybersecurity of Cyber-Physical Systems 
Nazli Choucri (MIT) 
2:45 PM Wednesday, January 15, 2020

Cybersecurity policies and guidelines are presented   in stand-alone text-form. Text obscures feedback, delays, risk etc. and other features related to policy-technology complexities.  Our purpose is to create  a platform of methods and tools for cybersecurity analytics. Here we report on recent extensions with applications to policy-based risk analysis.

Efficient Estimation of the Cyber-attack Loss Distribution 
David Nicol (University of Illinois at Urbana-Champaign) 
10:00 AM Thursday, January 16, 2020

Cyber-security risk assessment includes estimation of the losses possible to a system (physical and/or data) due to cyber-attack. As there are random elements to this, and/or we model uncertainty with randomness, we seek to estimate a “loss distribution” The tail of the distribution represents low-probability high-impact events. However, quantifying the probability of rare events using standard risk measures such as the survival function and conditional expectation is inefficient using standard Monte Carlo techniques because of these significant low probability events.  This paper modifies a traditional network connectivity reliability model to cyber-attacks, with an emphasis on losses due to the attack.

When under rare event realizations where the loss is greater than a predefined threshold, we (i) derive the analytically optimal important sampling (IS) schemes for the two risk measures and (ii) proposed approximations to the optimal IS schemes, both of which satisfy the bounded relative error (BRE) property. Both approximation schemes rely on finding close approximation to the solution of an NP-hard problem, which we obtain with arbitrary precision using a branch-and-bound algorithm. A case study demonstrates our model and the numerical results confirm the accuracy and efficiency of the proposed IS schemes. 

Finding Solutions for Privacy Problems: Privacy Design Patterns 
Nick Doty (UC Berkeley, School of Information) 
10:45 AM Thursday, January 16, 2020 

Design patterns are abstracted solutions to common problems, applied in particular contexts. Historically applied to architecture, urban design and object-oriented programming, we consider how abstract solutions exist to address common privacy problems in technical systems. While some of those solutions may be well-known to some practitioners, we aim to seek out those solutions to privacy problems more systematically. In doing so, we can build a library of documented privacy design patterns based on experience and evidence in the field and then identify areas where solutions are missing. Our larger goal is to evaluate privacy design patterns as part of a systematic engineering approach to privacy. 

With collaborators, we have previously developed a library of 30-some privacy design patterns. Now, we are reviewing computer science publications analyzing mobile applications and their privacy pitfalls to identify where existing design patterns could apply or where patterns exist in industry practice to be documented. We also review the complaints brought by regulators (including the FTC) and the guidance they have published, to see what problems and what solutions can be extracted. How we categorize patterns identified, and what patterns are available to implement different privacy strategies, will show us where new design work is needed. 

For more information, see: https://privacypatterns.org/

PechaKucha Presentations

• Hui Guo (Identifying Spying Apps using App Reviews)
• Varun Madathil (Anonymous Lottery in Proof-of-Stake)
• Nasif Imtiaz (Prioritizing Security Updates of Software Dependencies)
• Shaown Sarker (Hiding in Plain Site: Detecting JavaScript Obfuscation through Concealed Browser API Usage)
• Yuhang Lin (Vulnerability Exploit Detection Over Aggregated Container Data)
• Fogo Tunde-Onadele (OPatch: Beyond Patch Tuesday for Containerized Applications)
• Md Rayhanur Rahman (Do Configuration Management Tools Make Systems More Secure? A Case Study)

Posters

• Abida Haque
• Fogo Tunde-Onadele
• Hui Guo
• Iffat Anjum
• Igibek Koishybayev
• Jordan Jueckstock
• Md Rayhanur Rahman
• Nasif Imtiaz
• Richeng Jin
• Sarah Elder
• Sathvik Prasad
• Shaown Sarker
• Vaibhav Garg
• Varun Madathil
• Yuhang Lin

Hotel Accomodations:

The StateView Hotel, Autograph Collection 
Address: 2451 Alumni Dr, Raleigh, NC 27606 
Front Desk: (919) 743-0055

• On campus, close to the meeting venue

• Booking (Discounted Rate): Call the front desk on (919) 743-0055 and ask for NCSU rate

  Rate at 12/10/2019; subject to change 
  Check in : 01/15 
  Check out : 01/16 
  Room type: 1 king bed or 2 queen/double 
  NCSU : 189 
  Federal Gov/Military: 120