Graphic modeling in Distributed Autonomous and Asynchronous Automata (DA3)

Automated verification of distributed systems becomes very important in distributed computing. The graphical insight into the system in the early and late stages of the project is essential. In the design phase, the visual input helps to articulate the collaborative distributed components clearly. The formal verification gives evidence of correctness or malfunction, but in the latter case, graphical simulation of counterexample helps for better understanding design errors. For these purposes, we invented Distributed Autonomous and Asynchronous Automata (DA3), which have the same semantics as the formal verification base—Integrated Model of Distributed Systems (IMDS). The IMDS model reflects the natural characteristics of distributed systems: unicasting, locality, autonomy, and asynchrony. Distributed automata have all of these features because they share the same semantics as IMDS. In formalism, the unified system definition has two views: the server view of the cooperating distributed nodes and the agent view of the migrating agents performing distributed computations. The automata have two formally equivalent forms that reflect two views: Server DA3 for observing servers exchanging messages, and Agent DA3 for tracking agents, which visit individual servers in their progress of distributed calculations. We present the DA3 formulation based on the IMDS formalism and their application to design and verify distributed systems in the Dedan environment. DA3 formalism is compared with other concepts of distributed automata known from the literature.

SERVERLESS COMPUTATIONS RESOURCE SCHEDULING BASED ON DATA DEPENDENCY

The serverless computing model is becoming quite widespread. This model allows developers to create flexible and fault tolerant applications with an attractive billing model. The increasing complexity of serverless functions has led to the necessity to use serverless workflows – serverless functions invoking other serverless functions. However, such concept imposes certain requirements on the serverless functions that make distributed computations. The overhead of transferring data between serverless functions can significantly increase the execution time of a program using this approach. One way to reduce overhead is to improve serverless scheduling techniques. This paper discusses an approach to scheduling serverless computations based on data dependency analysis. We propose to divide the problem of planning of the computation of a composite serverless function into three stages. For each stage we provide a description by a mathematical model. We carried out a review of algorithms used to schedule resources by compilers and in parallel computing in multiprocessor systems to determine the best algorithm to implement in a prototype scheduler. For each algorithm, it was specified how it could be used for resource scheduling in serverless platforms. We provide a description of the developed prototype based on the Fission serverless platform. The prototype implements the critical path heuristic. It is shown that the improvements can significantly reduce the execution time up to two times for some types of serverless functions.

Cross-chain deals and adversarial commerce

Modern distributed data management systems face a new challenge: how can autonomous, mutually distrusting parties cooperate safely and effectively? Addressing this challenge brings up familiar questions from classical distributed systems: how to combine multiple steps into a single atomic action, how to recover from failures, and how to synchronize concurrent access to data. Nevertheless, each of these issues requires rethinking when participants are autonomous and potentially adversarial. We propose the notion of a cross-chain deal, a new way to structure complex distributed computations that manage assets in an adversarial setting. Deals are inspired by classical atomic transactions, but are necessarily different, in important ways, to accommodate the decentralized and untrusting nature of the exchange. We describe novel safety and liveness properties, along with two alternative protocols for implementing cross-chain deals in a system of independent blockchain ledgers. One protocol, based on synchronous communication, is fully decentralized, while the other, based on semi-synchronous communication, requires a globally shared ledger. We also prove that some degree of centralization is required in the semi-synchronous communication model.

Fortified Multi-Party Computation: Taking Advantage of Simple Secure Hardware Modules

In practice, there are numerous settings where mutually distrusting parties need to perform distributed computations on their private inputs. For instance, participants in a first-price sealed-bid online auction do not want their bids to be disclosed. This problem can be addressed using secure multi-party computation (MPC), where parties can evaluate a publicly known function on their private inputs by executing a specific protocol that only reveals the correct output, but nothing else about the private inputs. Such distributed computations performed over the Internet are susceptible to remote hacks that may take place during the computation. As a consequence, sensitive data such as private bids may leak. All existing MPC protocols do not provide any protection against the consequences of such remote hacks. We present the first MPC protocols that protect the remotely hacked parties’ inputs and outputs from leaking. More specifically, unless the remote hack takes place before the party received its input or all parties are corrupted, a hacker is unable to learn the parties’ inputs and outputs, and is also unable to modify them. We achieve these strong (privacy) guarantees by utilizing the fact that in practice parties may not be susceptible to remote attacks at every point in time, but only while they are online, i.e. able to receive messages. To this end, we model communication via explicit channels. In particular, we introduce channels with an airgap switch (disconnect-able by the party in control of the switch), and unidirectional data diodes. These channels and their isolation properties, together with very few, similarly simple and plausibly remotely unhackable hardware modules serve as the main ingredient for attaining such strong security guarantees. In order to formalize these strong guarantees, we propose the UC with Fortified Security (UC#) framework, a variant of the Universal Composability (UC) framework.

CALYPSO

Distributed ledgers provide high availability and integrity , making them a key enabler for practical and secure computation of distributed workloads among mutually distrustful parties. Many practical applications also require strong confidentiality , however. This work enhances permissioned and permissionless blockchains with the ability to manage confidential data without forfeiting availability or decentralization. The proposed Calypso architecture addresses two orthogonal challenges confronting modern distributed ledgers: (a) enabling the auditable management of secrets and (b) protecting distributed computations against arbitrage attacks when their results depend on the ordering and secrecy of inputs. Calypso introduces on-chain secrets, a novel abstraction that enforces atomic deposition of an auditable trace whenever users access confidential data. Calypso provides user-controlled consent management that ensures revocation atomicity and accountable anonymity. To enable permissionless deployment, we introduce an incentive scheme and provide users with the option to select their preferred trustees. We evaluated our Calypso prototype with a confidential document-sharing application and a decentralized lottery. Our benchmarks show that transaction-processing latency increases linearly in terms of security (number of trustees) and is in the range of 0.2 to 8 seconds for 16 to 128 trustees.