Mihai Carabas

Monitoring and auditing mobile operating systems

The popularity and usage of mobile devices have increased and therefore performance demands changed. An important aspect is represented by the whole system stability which has a big impact on general user experience. This paper proposes a solution for improving the implementation process and stability of operating systems for mobile devices. It presents penetration testing techniques, in order to discover security vulnerabilities. The focus will be on the process of finding the bugs, the exploration part, and the exploitation will only be mentioned to highlight the importance of the vulnerabilities. The second part of the paper consists of a system service which collects the error messages and sends them to a server in order to perform a triage. Received data is saved, processed and then useful information is exposed to developers.

Smart malware detection on Android

Nowadays, because of its increased popularity, Android is target to a growing number of attacks and malicious applications, with the purpose of stealing private information and consuming credit by subscribing to premium services. Most of the current commercial antivirus solutions use static signatures for malware detection, which may fail to detect different variants of the same malware and zero-day attacks. In this paper, we present a behavior-based, dynamic analysis security solution, called Android Malware Detection System, for detecting both well-known and zero-day malware. The proposed solution uses a machine learning classifier in order to differentiate between the behaviors of legitimate and malicious applications. In addition, it uses the application statistics for determining its reputation. The final decision is based on a combination of the classifiers result and the application reputation. The solution includes a unique and extensive set of data collectors, which gather application-specific data that describe the behavior of the monitored application. We evaluated our solution on a set of legitimate and malicious applications and obtained a high accuracy of 0.985. Our system is able to detect zero-day malware samples that are not detected by current commercial solutions. Our solution outperforms other similar solutions running on mobile devices. Copyright

Energy-efficient virtualized clusters

Abstract In this paper we provide the state of the art for the virtualization techniques and means to reduce power consumption using it. Virtualization allows us to answer all the requirements with many-core servers and thus eliminate the one size does not fit all issue. The resulting pool of resources is beneficial from an economic as well as environmental point of view. It brings benefits of scale to all logistic elements of the problem: power supply, cooling, floor space. When talking about virtualization and power consumption, one important aspect to be taken into account is data center’s heterogeneity from the hardware architecture point of view (e.g., X86, PowerPC). Mapping virtualized operating systems on hardware nodes in order to minimize power consumption is still an open issue that will be addressed throughout this paper: given a number of physical machines, we try to map on them the available virtual machines (called virtual machine assignment) in order to have an efficient system when relating to power consumption. We expose new general bounds for the power consumption of a virtual machine assignment based on Jensen inequality. The lower bound has been previously obtained and used into literature, so here we only rediscover it in a simplified and more clear manner. The upper bound is new and general. Furthermore we practically evaluate some discrete cases and we proposed some graphics with the power consumption and its bounds for some particular real cases.

Lightweight Display Virtualization For Mobile Devices

Recent years have witnessed a constant evolution of mobile systems, from hardware improvements such as multi-core CPUs and virtualization extensions, to a rise in software complexity, including application stores that are the home for millions of mobile applications. This enables and encourages the rising trend of mobile virtualization. In this paper we present an approach for providing lightweight and fast display virtualization for mobile devices, allowing multiple operating systems to share the same physical display screen. We describe a design that provides the means to perform efficient display virtualization, making use of the concept of overlays implemented in recent underlying mobile display subsystems. We demonstrate our approach on a dual Android setup: two operating systems running simultaneously on the same mobile device, using our display virtualization framework.

Scalability of a Web Server: How Does Vertical Scalability Improve the Performance of a Server?

This paper presents how the throughput of a server is influenced by the applying a vertical scalability. The paper studies the results obtained in measuring the response time of the server and the processing time of the server when dealing with a large number of requests by modifying the configuration of the machine, increasing the number of cores the machine has and increasing the RAM capacity. This represents a test to see how many requests a server can process and complete when dealing with large amount of data in a short period of time.

Native Runtime Environment for Internet of Things

Over the past few years, it has become obvious that Internet of Things, in which all embedded devices are interconnected, will soon turn into reality. These smart devices have limited processing and energy resources and there is a need to provide security guarantees for applications running on them. In order to meet these demands, in this paper we present a POSIX-compliant native runtime environment which runs over a L4-based microkernel, a thin operating system, on top of which applications for embedded devices can run and thus be a part of the Internet of Things.

Horizontal scalability towards server performance improvement

This paper presents an evaluation of an experiment conducted on a server to analyse how horizontal scalability affects its performance. This paper studies the results obtained by measuring response times and processing times when dealing with many requests by adding more machines to the system. This paper presents the technologies used to build this system of machines as well as the results obtained comparing them to a basic configuration and offering a complete analysis of how horizontal scalability affects the performance of a server.

Evaluation of multi-source downloads for FTS

The data transfer in the Grid at CERN (the European Organization for Nuclear Research) has seen constant improvement, be it through optimizing the existing tools, GridFTP and GFAL2 or by adding new tools such as XRootd. Unfortunately, all these have reached the maximum limit in terms of throughput. They are limited not by the network infrastructure, but by the fact that they use a single source for transfers, despite the existence of multiple replicas. In this paper we take on the challenge of evaluating the effects of using multiple sources, over the throughput, by comparing the download speed of the tools mentioned above with Aria2 and an under development version of XRootd, both supporting the use of multiple sources.

File caching and purging strategies in the disk pool manager

Disk Pool Manager is a lightweight software which offers a simple way of creating a disk-based grid storage system. Its new, experimental features allow the deployment of very large file caches, which aim to provide a better experience for the users by storing previously accessed data in volatile disk pools and using them to serve subsequent requests much faster. After the disk pools are filled with data, a purging policy is responsible for evicting one or more items in order to make room for the new ones. In this paper, I will describe a file purging strategy that fits the requirements of the target software and underline why it is preferred over classical purging algorithms.

Generic Data Synchronization Algorithm in Distributed Systems

The increasing number of mobile users raises serious challenges upon middle-tier synchronization techniques and algorithms. The dynamics of backend systems, as well as those of frontend devices translates into an ever-growing demand for service standardization and flexibility. The current paper describes a flexible, scalable, platform-independent system relying on a distributed backend to provide data synchronization services in a multi-user system. The paper describes the desired architectural model used for this approach, continues with the description of the algorithm used by the backend subsystems and briefly describes an exemplary implementation of the aforementioned design.