Fetahi Wuhib

Gossip-based resource management for cloud environments

We address the problem of resource management for a large-scale cloud environment that hosts sites. Our contribution centers around outlining a distributed middleware architecture and presenting one of its key elements, a gossip protocol that meets our design goals: fairness of resource allocation with respect to hosted sites, efficient adaptation to load changes and scalability in terms of both the number of machines and sites. We formalize the resource allocation problem as that of dynamically maximizing the cloud utility under CPU and memory constraints. While we can show that an optimal solution without considering memory constraints is straightforward (but not useful), we provide an efficient heuristic solution for the complete problem instead. We evaluate the protocol through simulation and find its performance to be well-aligned with our design goals.

A Gossip Protocol for Dynamic Resource Management in Large Cloud Environments

We address the problem of dynamic resource management for a large-scale cloud environment. Our contribution includes outlining a distributed middleware architecture and presenting one of its key elements: a gossip protocol that (1) ensures fair resource allocation among sites/applications, (2) dynamically adapts the allocation to load changes and (3) scales both in the number of physical machines and sites/applications. We formalize the resource allocation problem as that of dynamically maximizing the cloud utility under CPU and memory constraints. We first present a protocol that computes an optimal solution without considering memory constraints and prove correctness and convergence properties. Then, we extend that protocol to provide an efficient heuristic solution for the complete problem, which includes minimizing the cost for adapting an allocation. The protocol continuously executes on dynamic, local input and does not require global synchronization, as other proposed gossip protocols do. We evaluate the heuristic protocol through simulation and find its performance to be well-aligned with our design goals.

Robust monitoring of network-wide aggregates through gossiping

We investigate the use of gossip protocols for continuous monitoring of network-wide aggregates under crash failures. Aggregates are computed from local management variables using functions such as SUM, MAX, or AVERAGE. For this type of aggregation, crash failures offer a particular challenge due to the problem of mass loss, namely, how to correctly account for contributions from nodes that have failed. In this paper we give a partial solution. We present G-GAP, a gossip protocol for continuous monitoring of aggregates, which is robust against failures that are discontiguous in the sense that neighboring nodes do not fail within a short period of each other. We give formal proofs of correctness and convergence, and we evaluate the protocol through simulation using real traces. The simulation results suggest that the design goals for this protocol have been met. For instance, the tradeoff between estimation accuracy and protocol overhead can be controlled, and a high estimation accuracy (below some 5% error in our measurements) is achieved by the protocol, even for large networks and frequent node failures. Further, we perform a comparative assessment of GGAP against a tree-based aggregation protocol using simulation. Surprisingly, we find that the tree-based aggregation protocol consistently outperforms the gossip protocol for comparative overhead, both in terms of accuracy and robustness.

Robust Monitoring of Network-wide Aggregates through Gossiping

We investigate the use of gossip protocols for continuous monitoring of network-wide aggregates under crash failures. Aggregates are computed from local management variables using functions such as SUM, MAX, or AVERAGE. For this type of aggregation, crash failures offer a particular challenge due to the problem of mass loss, namely, how to correctly account for contributions from nodes that have failed. In this paper we give a partial solution. We present G-GAP, a gossip protocol for continuous monitoring of aggregates, which is robust against failures that are discontiguous in the sense that neighboring nodes do not fail within a short period of each other. We give formal proofs of correctness and convergence, and we evaluate the protocol through simulation using real traces. The simulation results suggest that the design goals for this protocol have been met. For instance, the tradeoff between estimation accuracy and protocol overhead can be controlled, and a high estimation accuracy (below some 5% error in our measurements) is achieved by the protocol, even for large networks and frequent node failures. Further, we perform a comparative assessment of GGAP against a tree-based aggregation protocol using simulation. Surprisingly, we find that the tree-based aggregation protocol consistently outperforms the gossip protocol for comparative overhead, both in terms of accuracy and robustness.

Decentralized detection of global threshold crossings using aggregation trees

The timely detection that a monitored variable has crossed a given threshold is a fundamental requirement for many network management applications. A challenge is the detection of threshold crossing of network-wide variables, which are computed from device counters across the network, using aggregation functions such as SUM, MAX and AVERAGE. This paper contains a detailed description and a comprehensive evaluation of TCA-GAP, a protocol for detecting threshold crossings of network-wide aggregates in a distributed way. Elements of its design include tree-based incremental aggregation for estimating the value of aggregates, a local hysteresis mechanism to reduce overhead and dynamic recomputation of local thresholds to ensure correctness. The protocol is evaluated through extensive simulation using real traces in scenarios with network sizes up to 5232 nodes. From the measurements, we conclude that the protocol is efficient in the sense that the overhead is negligible when the aggregate is far from the threshold. It is scalable as the protocol overhead is independent of the system size for the network sizes and scenario configurations considered. We demonstrate that the local hysteresis parameter can be used to control the tradeoff between protocol overhead and detection delay. We further report on results on how node failures impact overhead and detection quality of the protocol.

Distributed monitoring and resource management for large cloud environments

Over the last decade, the number, size and complexity of large-scale networked systems has been growing fast, and this trend is expected to accelerate. The best known example of a large-scale networked system is probably the Internet, while large datacenters for cloud services are the most recent ones. In such environments, a key challenge is to develop scalable and adaptive technologies for management functions. This thesis addresses the challenge by engineering several protocols for distributed monitoring and resource management that are suitable for large-scale networked systems. The protocols are evaluated through theoretical analysis, simulation studies and testbed experimentation. The evaluation results show that the protocols achieve their respective design objectives with respect to quality, efficiency, scalability, controllability and adaptability.

A gossiping protocol for detecting global threshold crossings

We investigate the use of gossip protocols for the detection of network-wide threshold crossings. Our design goals are low protocol overhead, small detection delay, low probability of false positives and negatives, scalability, robustness to node failures and controllability of the trade-off between overhead and detection delay. Based on push-synopses, a gossip protocol introduced by Kempe et al., we present a protocol that indicates whether a global aggregate of static local values is above or below a given threshold. For this protocol, we prove correctness and show that it converges to a state with no overhead when the aggregate is sufficiently far from the threshold. Then, we introduce an extension we call TG-GAP, a protocol that (1) executes in a dynamic network environment where local values change and (2) implements hysteresis behavior with upper and lower thresholds. Key elements of its design are the construction of snapshots of the global aggregate for threshold detection and a mechanism for synchronizing local states, both of which are realized through the underlying gossip protocol. Simulation studies suggest that TG-GAP is efficient in that the protocol overhead is minimal when the aggregate is sufficiently far from the threshold, that its overhead and the detection delay are largely independent on the system size, and that the tradeoff between overhead and detection quality can be effectively controlled. Lastly, we perform a comparative evaluation of TG-GAP against a tree-based protocol. We conclude that, for detecting global threshold crossings in the type of scenarios investigated, the tree-based protocol incurs a significantly lower overhead and a smaller detection delay than a gossip protocol such as TG-GAP.

Decentralized computation of threshold crossing alerts

Threshold crossing alerts (TCAs) indicate to a management system that a management variable, associated with the state, performance or health of the network, has crossed a certain threshold. The timely detection of TCAs is essential to proactive management. This paper focuses on detecting TCAs for network-level variables, which are computed from device-level variables using aggregation functions, such as SUM, MAX, or AVERAGE. It introduces TCA-GAP, a novel protocol for producing network-wide TCAs in a scalable and robust manner. The protocol maintains a spanning tree and uses local thresholds, which adapt to changes in network state and topology, by allowing nodes to trade unused “threshold space”. Scalability is achieved through computing the thresholds locally and through distributing the aggregation process across all nodes. Fault-tolerance is achieved by a mechanism that reconstructs the spanning tree after node addition, removal or failure. Simulation results on an ISP topology show that the protocol successfully concentrates traffic overhead to periods where the aggregate is close to the given threshold.

Decentralized real-time monitoring of network-wide aggregates

The traditional monitoring paradigm of network and systems management, characterized by a central entity polling individual devices, is not adequate for todays large-scale networked systems whose states and configurations are highly dynamic. We outline principles for monitoring such new systems and stress the need for protocols that continuously monitor network-wide aggregates. To keep the overhead at acceptable levels, such protocols must be tunable, e.g., allow controlling the trade-off between accuracy and overhead. We describe and compare two of our efforts in developing protocols for decentralized monitoring of aggregates; one is based on spanning trees, the other on gossiping.

Implementation and Evaluation of a Protocol for Detecting Network-Wide Threshold Crossing Alerts

Threshold crossing alerts (TCAs) indicate to a management system that a management variable, associated with the state, performance or health of the network, has crossed a certain threshold. In this paper, we report on implementing and evaluating TCA-GAP, a distributed protocol for detecting network-wide TCAs, which reports threshold crossings on aggregates, such as SUM, AVERAGE, or MAX of device counters. We present a concept for assessing the quality of detecting network-wide TCAs, which we apply to evaluate TCA-GAP on a lab testbed. First, we evaluate the correctness of the protocol by determining the correctly detected threshold crossings, the false positives and the false negatives. Second, for the correctly detected threshold crossings, we measure the delays between the time a crossing was reported by the protocol and the time of its actual occurrence. Finally, we demonstrate that the fundamental tradeoff between the quality of TCA detection and the management overhead can be controlled in TCA-GAP by modifying the maximum message rate on the management overlay.