Augusto Ciuffoletti

Architecture of a Network Monitoring Element

The scalability of a network monitoring system, a vital component of a Grid, is challenging. We propose a solution based on demand driven monitoring sessions that use passive network monitoring techniques on a domain oriented overlay network. Related aspects of security and group membership maintenance are also considered.

Automated Deployment of a Microservice-based Monitoring Infrastructure

Abstract We explore the specification and the automated deployment of a monitoring infrastructure in a container-based distributed system. This result shows that highly customizable monitoring infrastructures can be effectively provided as a service, and that a key step in this process is the definition of an expandable abstract model for them. So we start defining a simple model of the monitoring infrastructure that provides an interface between the user and the cloud management system. The interface follows the guidelines of Open Cloud Computing Interface (OCCI), the cloud interface standard proposed by the Open Grid Forum. The definition is simple and generic and it is a first step towards the definition of a standard interface for Monitoring Services. It allows the definition of complex, hierarchical monitoring infrastructure by composing multiple instances of two basic components, one for measurement and another for data distribution,. We illustrate how the monitoring functionalities that are defined through the interface are implemented as microservices embedded in containers. The internals of each microservice reflects the distinction between core functionalities which are bound to the standard, and custom plugin modules. We describe the engine that automatically deploys a system of microservices that implements the monitoring infrastructure. Special attention is paid to preserve the distinction between core and custom functionalities, and the on demand nature of a cloud service. A proof of concept demo is available through the Docker hub and consists of two multi-threaded Java appli- cations that implement the two basic components.

The wandering token: Congestion avoidance of a shared resource

In a distributed system where scalability is an issue, the problem of enforcing mutual exclusion often arises in a soft form: the infrequent failure of the mutual exclusion predicate is tolerated, without compromising the consistent operation of the overall system. For instance this occurs when the operation subject to mutual exclusion requires massive use of a shared resource. We introduce a scalable soft mutual exclusion algorithm, based on token passing: one distinguished feature of our algorithm is that instead of introducing an overlay topology we adopt a random walk approach. The consistency of our proposal is evaluated by simulation, and we exemplify its use in the coordination of large data transfers in a backbone based network. This algorithm is studied in the frame of the CoreGRID Institute of Grid Information, Resource and Workflow Monitoring Services, in cooperation with the FORTH Institute, in Greece.

Challenges Emerging from Future Cloud Application Scenarios

The cloud computing paradigm encompasses several key differentiating elements and technologies, tackling a number of inefficiencies, limitations and problems that have been identified in the distributed and virtualized computing domain. Nonetheless, and as it is the case for all emerging technologies, their adoption led to the presentation of new challenges and new complexities. In this paper we present key application areas and capabilities of future scenarios, which are not tackled by current advancements and highlight specific requirements and goals for advancements in the cloud computing domain. We discuss these requirements and goals across different focus areas of cloud computing, ranging from cloud service and application integration, development environments and abstractions, to interoperability and relevant to it aspects such as legislation. The future application areas and their requirements are also mapped to the aforementioned areas in order to highlight their dependencies and potential for moving cloud technologies forward and contributing towards their wider adoption.

Monitoring a virtual network infrastructure: an IaaS perspective

Infrastructure as a Service (IaaS) providers keep extending with new features the computing infrastructures they offer on a pay per use basis. In this paper we explore reasons and opportunities to include networking within such features, meeting the demand of users that need composite computing architectures similar to Grids. The introduction of networking capabilities within IaaSs would further increase the potential of this technology, and also foster an evolution of Grids towards a confluence, thus incorporating the experiences matured in this environment. Network monitoring emerges as a relevant feature of such virtual architectures, which must exhibit the distinguishing properties of the IaaS paradigm: scalability, dynamic configuration, accounting. Monitoring tools developed with the same purpose in Grids provide useful insights on problems and solutions.

Implementing a Probabilistic Clock Synchronization Algorithm

In this paper we present a new probabilistic clock synchronization algorithm, its prototype implementation and experimental results. The algorithm follows the client-server programming paradigm and is designed to work in a departmental environment with few servers and a number of clients connected through an arbitrary network topology.At the core of the algorithm is a remote clock reading method that mitigates the negative effects of message delay uncertainty. The implementation proves the effectiveness of this approach and corroborates the theoretical speculations.

Monitoring the connectivity of a grid

Grid computing is a new paradigm that enables the distributed coordination of resources and services which are geographically dispersed, span multiple trust domains and are heterogeneous. Network infrastructure monitoring, while vital for activities such as service selection, exhibits inherent scalability problems: in principle, in a Grid composed of <i>n</i> resources, we need to keep record of <i>n</i>² end-to-end paths. We introduce an approach to network monitoring that takes into account scalability: a Grid is partitioned into domains, and network monitoring is limited to the measurement of domain-to-domain connectivity. However, partitions must be consistent with network performance, since we expect that an observed network performance between domains is representative of the performance between the Grid Services included into domains.

Using simple diffusion to synchronize the clocks in a distributed system

Simple diffusion consists in propagating some data or computation to every unit in a network by spreading it from one unit to its neighbors, which in their turn spread it to the neighbors until every unit is reached. Simple diffusion is easy to implement, scalable, robust with respect to omission faults, and privileges local communication (i.e., uses only links whose cost is the lowest in terms of time needed and/or traffic generated). A solution to the problem of keeping synchronized all the clocks in a large distributed system is presented, that is based on simple diffusion. The traffic generated is O(N), where N is the number of units in the network. The attainable precision is O(d*/spl epsiv/) where /spl epsiv/ is the precision with which a unit can synchronize its clock with that of a neighbor, and d is the maximum distance between a unit and the nearest reference clock.<<ETX>>

Application level interface for a cloud monitoring service

Abstract We introduce a new paradigm, based on an extension of the Open Cloud Computing Interface (OCCI), for the on-demand monitoring of the cloud resources provided to a user. We have extended the OCCI with two new sub-types of core entities: one to collect the measurements and the other to process them. The user can request instances of such entities to implement a monitoring infrastructure. The paradigm does not target a specific cloud model, and is therefore applicable to any kind of resource provided as a service . The specifications include only the minimum needed to describe a monitoring infrastructure, thus making this standard extension simple and easily adoptable. Despite its simplicity, the model is able to describe complex solutions, including private/public clouds, and covers both infrastructure and application monitoring. To highlight the impact of our proposal in practice, we have designed an engine that deploys a monitoring infrastructure using its OCCI-compliant descriptions. The design is implemented in a prototype that is available as open source.

A Simple and Generic Interface for a Cloud Monitoring Service

The paper addresses the problem of defining an interface for the configuration of monitoring activities in a cloud. Such an interface must be extremely flexible to cope with a wide range of use cases, from the minimalist plug-and-play user, to that governing a complex infrastructure.