Reuven Yagel

Toward self-stabilizing operating systems

This work presents several approaches for designing self-stabilizing operating systems. The first approach is based on periodical automatic reinstalling of the operating system and restart. The second, reinstalls the executable portion of the operating system and uses predicates on the operating system state (content of variables) to ensure that the operating system does not diverge from its specifications. The last approach presents an example of a tailored self-stabilizing very-tiny operating system. Prototypes using the Intel Pentium processor were composed.

Memory management for self-stabilizing operating systems

This work presents several approaches for designing the memory management component of self-stabilizing operating systems. We state the requirements which a memory manager should satisfy. One requirement is eventual memory hierarchy consistency among different copies of data residing in different (level of) memory devices e.g., ram and disk. Another requirement is stabilization preserving where the memory manager ensures that every process that is proven to stabilize independently, also stabilizes under the (self-stabilizing scheduler and the) memory manager operation. Three memory managers that satisfy the above requirements are presented. The first allocates the entire physical memory to a single process in every given point of time, the second uses fixed partition of the memory among processes, and the last uses memory leases for dynamic memory allocations.

Self-Stabilization as a Foundation for Autonomic Computing

This position paper advocates the use of the well defined and provable self-stabilization property of a system, to achieve the goals of the self-* paradigms and autonomic computing. Several recent results starting from hardware concerns, continuing with the operating system, and ending in the applications, are integrated: the self-stabilizing microprocessor, with the self-stabilizing operating system, the self-stabilization preserving compiler, and the self-stabilizing autonomic recoverer for applications

Towards Self-Stabilizing Operating Systems

This work presents several approaches for designing self-stabilizing operating systems. The first approach is based on periodical automatic reinstalling of the operating system and restart. The second reinstalls the executable portion of the operating system and uses predicates on the operating system state (content of variables) to ensure that the operating system does not diverge from its specifications. The last approach presents an example of a tailored self-stabilizing very tiny operating system. Prototypes using the Intel Pentium processor were composed.

Self-stabilizing device drivers

This work presents approaches for designing the input-output device management components of self-stabilizing operating systems. As an example, we demonstrate the non-stability of the ATA standard protocol for storage devices. We state the requirements that an operating system and I/O devices should satisfy in order to become self-stabilizing. Then we suggest two solutions to satisfy these requirements. The first uses leases in order to guarantee progress from the I/O device side. The second assumes stabilization of the I/O device, and uses snapshots to perform consistency checks. By supplying an infrastructure for practical self-stabilizing systems, robust and dependable systems can be achieved.

ROM: An Approach to Self-consistency Verification of a Runnable Ontology Model

In the quest for the highest possible abstraction of software systems, Runnable Knowledge has been proposed for MDA. But in order to be useful in practice as a system design, it must be amenable to verification. This work precisely describes the necessary steps by which ROM – a Runnable Ontology Model tool – concurrently generates a running model and its respective test script from the designed Runnable Knowledge, allowing verification that the model is self-consistent. The novel implementation idea is to use ready-made mock object libraries to efficiently obtain the code for a running model. Detailed examples are provided to illustrate each of the ROM generation steps.

Self-stabilizing operating systems

This work presents new directions for building a self stabilizing operating system kernel. A system is self-stabilizing [3, 4] if it can be started in any possible state and it converges to a desired behavior. A state of a system is an assignment of arbitrary values to the systems variables. The usefulness of such a system in critical and remote systems cannot be over estimated. Entire years of work maybe lost when the operating system of an expensive complicated device e.g., a spaceship, may reach an arbitrary state due to say, soft errors (e.g., [8]), and be lost forever. Last results of this research can be found in [5] and [6].

Purifying data by machine learning with certainty levels

A fundamental paradigm used for autonomic computing, self-managing systems, and decision-making under uncertainty and faults is machine learning. Machine learning uses a data-set, or a set of data-items. A data-item is a vector of feature values and a classification. Occasionally these data sets include misleading data items that were either introduced by input device malfunctions, or were maliciously inserted to lead the machine learning to wrong conclusions. A reliable learning algorithm must be able to handle a corrupted data-set. Otherwise, an adversary (or simply a malfunctioning input device that corrupts a portion of the data-set) may lead to inaccurate classifications. Therefore, the challenge is to find effective methods to evaluate and increase the certainty level of the learning process as much as possible. This paper introduces the use of a certainty level measure to obtain better classification capability in the presence of corrupted data items. Assuming a known data distribution (e.g., a normal distribution) and/or a known upper bound on the given number of corrupted data items, our techniques define a certainty level for classifications. Another approach suggests enhancing the random forest techniques to cope with corrupted data items by augmenting the certainty level for the classification obtained in each leaf in the forest. This method is of independent interest, that of significantly improving the classification of the random forest machine learning technique in less severe settings.

Self-stabilizing Byzantine-Tolerant Distributed Replicated State Machine

Replicated state machine is a fundamental concept used for obtaining fault tolerant distributed computation. Legacy distributed computational architectures (such as Hadoop or Zookeeper) are designed to tolerate crashes of individual machines. Later, Byzantine fault-tolerant Paxos as well as self-stabilizing Paxos were introduced. Here we present for the first time the self-stabilizing Byzantine fault-tolerant version of a distributed replicated machine. It can cope with any adversarial takeover on less than one third of the participating replicas. It also ensures automatic recovery following any transient violation of the system state, in particular after periods in which more than one third of the participants are Byzantine. A prototype of self-stabilizing Byzantine-tolerant replicated Hadoop master node has been implemented. Experiments show that fully distributed recovery of cloud infrastructures against Byzantine faults can be made practical when relying on self-stabilization in local nodes. Thus automated cloud protection against a wide variety of faults and attacks is possible.

Self-Stabilizing Virtual Machine Hypervisor Architecture for Resilient Cloud

This paper presents the architecture for a self-stabilizing hypervisor able to recover itself in the presence of Byzantine faults regardless of the state it is currently in. Our architecture is applicable to wide variety of underlying hardware and software and does not require augmenting computers with special hardware. The actions representing defense and recovery strategies can be specified by a user. We describe our architecture in OS-independent terms, thus making it applicable to various virtualization infrastructures. We also provide a prototype extending the Linux-based hypervisor KVM with the self-stabilizing functionality. These features allow augmenting KVM with robustness functionality in the coming stages and moving to cloud management system architectures such as OpenStack to support more industrial scenarios.