Logo des Repositoriums

Auflistung nach Autor:in "Schwarz, Christian"

1 - 4 von 4
  • Textdokument
    DMA Security in the Presence of IOMMUs
    (Tagungsband des FG-BS Frühjahrstreffens 2022, 2022) Schwarz, Christian; Reusch, Viktor; Planeta, Maksym
    Faulty, vulnerable or malicious PCIe devices can harm a system through DMA. IOMMUs can act as a security mechanism to protect against this problem by restricting the memory that is accessible via DMA. Unfortunately, there are methods to bypass the IOMMU restrictions. This paper is a survey over the currently existing bypasses and their feasibility. Current systems might be exploited from any untrusted source of DMA, which includes peripheral PCIe devices, virtual machines using SR-IOV, and even RDMA network cards, which enable remote attacks. Key strategies for the attacks presented here are Rowhammer, cache side-channels, and the exploitation of weaknesses in device drivers, e.g., for network cards, or protocols like PCIe or Ethernet OAM. An attacker can potentially achieve denial of service, the reading of confidential data, and even arbitrary code execution. Fortunately, there are some precautions to reduce the risks for affected systems.
  • Konferenzbeitrag
    Improving GPU Matrix Multiplication by Leveraging Bit Level Granularity and Compression
    (BTW 2023, 2023) Fett, Johannes; Schwarz, Christian; Kober, Urs; Habich, Dirk; Lehner, Wolfgang
    In this paper we introduce BEAM as a novel approach to perform GPU based matrix multiplication on compressed elements. BEAM allows flexible handling of bit sizes for both input and output elements. First evaluations show promising speedups compared to an uncompressed state-of-the-art matrix multiplication algorithm provided by nvidia.
  • Konferenzbeitrag
    Modelling a real-time control system using parameterized linear hybrid automata
    (INFORMATIK 2011 – Informatik schafft Communities, 2011) Schwarz, Christian
    Many real-time systems operate in safety-critical environments. Testing of these systems is hard in principle and can never guarantee full coverage. So a method for formally proving correctness is desirable. Another problem in real-time system design is the determination of deadlines and periods of time-critical processes as a prerequisite of the scheduler design. We want to use Hybrid Automata (HA) to tackle both problems. HA are a formal language that is equipped with semantics and thus accessible to formal analysis. We introduce the syntax and semantics of a new class of HA called Parameterized Linear Hybrid Automata and demonstrate its advantages by modelling a typical real-time control system using an extended version of the tool HieroMate.
  • Konferenzbeitrag
    Simulating multi-tenant OLAP database clusters
    (Datenbanksysteme für Business, Technologie und Web (BTW), 2011) Schaffner, Jan; Eckart, Benjamin; Schwarz, Christian; Brunnert, Jan; Jacobs, Dean; Zeier, Alexander; Plattner, Hasso
    Simulation of parallel database machines was used in many database research projects during the 1990ies. One of the main reasons why simulation approaches were popular in that time was the fact that clusters with hundreds of nodes were not as readily available for experimentation as it is the case today. At the same time, the simulation models underlying these systems were fairly complex since they needed to capture both queuing processes in hardware (e.g. CPU contention or disk I/O) and software (e.g. processing distributed joins). Todays trend towards more specialized database architectures removes large parts of this complexity from the modeling task. As the main contribution of this paper, we discuss how we developed a simple simulation model of such a specialized system: a multi-tenant OLAP cluster based on an in-memory column database. The original infrastructure and testbed was built using SAP TREX, an in-memory column database part of SAP's business warehouse accelerator, which we ported to run on the Amazon EC2 cloud. Although we employ a simple queuing model, we achieve good accuracy. Similar to some of the parallel systems of the 1990ies, we are interested in studying different replication and high-availability strategies with the help of simulation. In particular, we study the effects of mirrored vs. interleaved replication on throughput and load distribution in our cluster of multi-tenant databases. We show that the better load distribution inherent to the interleaved replication strategy is exhibited both on EC2 and in our simulation environment.