O-RAN introduces intelligent and flexible network control in all parts of the network. The use of controllers with open interfaces allow us to gather real time network measurements and make intelligent/informed decision. The work in this paper focuses on developing a use-case for open and reconfigurable networks to investigate the possibility to predict handover events and understand the value of such predictions for all stakeholders that rely on the communication network to conduct their business. We propose a Long-Short Term Memory Machine Learning approach that takes standard Radio Access Network measurements to predict handover events. The models were trained on real network data collected from a commercial O-RAN setup deployed in our OpenIreland testbed. Our results show that the proposed approach can be optimized for either recall or precision, depending on the defined application level objective. We also link the performance of the Machine Learning (ML) algorithm to the network operation cost. Our results show that ML-based matching between the required and available resources can reduce operational cost by more than 80%, compared to long term resource purchases.

The widespread adoption of Machine Learning (ML) is leading to an increase in processing demands. Dealing with the growing volume of data poses a significant challenge in providing accurate classification services using ML models. Offloading ML tasks to network switches presents an opportunity to tackle this challenge, offering high throughput and low latency. Nonetheless, network devices encounter limitations in resources, and programmable languages like P4 lack support for basic operations, necessary for the ML methods, including floating-point arithmetic and native repetition structures. In this paper, we investigate the use of drift detection ML models to enhance the accuracy of in-network traffic classification. The novelty lies in designing drift detection based on bitwise operations, which are well-suited for implementation within the data plane. As a proof-of-concept, we implement drift detection using the P4 language on BMv2 switches, validated with a dataset of over 2 million samples. Our results demonstrate a significant increase in classification accuracy with drift detection, while maintaining line-speed operation and quickly adapting to changes in traffic patterns.

In the fight against cyber attacks, Network Softwarization (NS) is a flexible and adaptable shield, using advanced software to spot malicious activity in regular network traffic. However, the availability of comprehensive datasets for mobile networks, which are fundamental for the development of Machine Learning (ML) solutions for attack detection near their source, is still limited. Cross-Domain Artificial Intelligence (AI) can be the key to address this, although its application in Open Radio Access Network (O-RAN) is still at its infancy. To address these challenges, we deployed an end-to-end O-RAN network, that was used to collect data from the RAN and the transport network. These datasets allow us to combine the knowledge from an in-network ML traffic classifier for attack detection to bolster the training of an ML-based traffic classifier specifically tailored for the RAN. Our results demonstrate the potential of the proposed approach, achieving an accuracy rate of 93%. This approach not only bridges critical gaps in mobile network security but also showcases the potential of cross-domain AI in enhancing the efficacy of network security measures.

This paper presents a scalable and efficient solution for secure network design that involves the selection and verification of network paths. The proposal addresses the challenges related to compliance policies by introducing a Proof-of-Transit (PoT) feasible implementation for path-aware programmable networks. Our approach relies on i) a source routing mechanism based on a fixed routeID representing a unique identifier per path, which serves as a key for PoT lookup tables; ii) the "in situ" that allows to collect telemetry information in the packet while the packet traverses a path. The former enables path selection with policy at the edge, while the later allows to perform path verification without extra probe-traffic. A P4 programmable language prototype demonstrates the effectiveness of this approach to protect against deviation attacks with low overhead. The results show its scalability considering the protocol overhead as the path length increases; a significant reduction in network’s forwarding state for fat-tree topologies depending on the workload per path (flows/path). Finally, experimental results show a RTT comparison evaluation, the impact of PoT computation, protection to path deviation and seamless path migration keeping flow protection.

We design and demonstrate a method for early detection of Denial-of-Service attacks. The proposed approach takes advantage of the OpenRAN framework to collect measurements from the air interface (for attack detection) and to dynamically control the operation of the Radio Access Network (RAN). For that purpose, we developed our near-Real Time (RT) RAN Intelligent Controller (RIC) interface. We apply and analyze a wide range of Machine Learning algorithms to data traffic analysis that satisfy the accuracy and latency requirements set by the near-RT RIC. Our results show that the proposed framework is able to correctly classify genuine vs. malicious traffic with high accuracy (i.e., 95%) in a realistic testbed environment, allowing us to detect attacks already at the Distributed Unit (DU), before malicious traffic even enters the Centralized Unit (CU).

This paper presents a scalable and efficient solution for secure network design that involves the selection and verification of network paths. The proposed approach addresses the challenge of extending compliance policies to cover path-aware programmable networks by decoupling the routing/forwarding mechanisms from the Proof-of-Transit (PoT) implementation. Thus, two concepts are bounded: i) a source routing mechanism based on a fixed routeID representing a unique identifier per path, which serves as a key for the PoT lookup table; ii) the “in situ” that allows to collect telemetry information in the packet while the packet traverses a path. The former enables path selection with policy at the edge, while the later allows to perform path verification without extra probe-traffic. A P4 programmable language prototype demonstrates the effectiveness of this approach to protect against deviation attacks with low overhead. The results show a significant reduction in network’s forwarding state for fat-tree topologies depending on the workload per path (flows/path).

Cryptographic hash functions are widely used to provide from digital time stamping to authenticity and digital signatures, mapping an extensive collection of messages into a small set of message digests and help to secure network connection and data, consequently consuming CPU resources. P4 enables data plane customisation using a high-level programming language to facilitate in-network computing development across diverse hardware targets, including Network Interface Cards (NICs). Currently, most P4 targets do not implement secure hash functions due to a lack of hardware instructions or the absence of formal functions to expose their native hardware-based implementation. Moreover, many applications and protocols cannot be instantiated using in-network computing due to stringent requirements based on these hash functions. In order to empower the security and other hash-based applications, in this paper we propose and implement a P4 shared object library for a secure hash algorithm 2 (SHA-2). Our goal is to enable SHA-2 to be used as an embedded Network Function (eNF), overcoming the lack of support in a SmartNIC architecture, in order to address the latency and throughput requirements of Service Function Chain (SFC) forwarding performance within the Network Function Virtualization (NFV) paradigm. Thus, our prototype is evaluated against kernel-level Open vSwitch (OvS) and user-space Data Plane Development Kit (DPDK) implementations. The outcomes demonstrate different tradeoffs over each platform, from the randomness added by the OS to the high cost of executing the aforesaid function using a network programmable device, leading us to highlight the best choice for each specific application.

Current Service Function Chaining (SFC) architectures are tailor-made for specific environments and platforms, often relying on SFC support on network devices or specialized frameworks. Thus, the service plane and the data plane are tightly coupled, which hinders innovation. For example, prototyping new SFC protocols usually requires re-implementing service functions (SFs) or modifying network devices. To address these issues, we propose Chaining-Box, a new SFC architecture based on a simple idea: implementing all the SFC functionality as a sequence of stages. This is done in a fully transparent manner without changing neither SFs nor network devices. Stages are implemented using BPF, a technology that allows user-defined programs to run inside the Linux kernel. The stages run as packets traverse the kernel stack and implement all SFC actions to provide the chaining. A proof-of-concept of Chaining-Box is implemented as a prototype, which demonstrates a decrease of 20%-40% in latency in comparison with similar proposals. The results also show equivalent performance when compared with OVS-based SFC, but allowing the bridges to be SFC-agnostic.

Innovative traffic engineering functions and services require disrupting routing and forwarding mechanisms to be performed with low overhead over complex network topologies. Source routing (SR) is a prominent alternative to table-based routing for providing the needed expressiveness and agility by reducing the number of network states. This work proposes the M-PolKA, a topology-agnostic multipath source routing scheme and orchestration architecture for reliable communications, which explores special properties from the Residue Number System (RNS) polynomial arithmetic. A P4-based proof-of-concept is experimentally demonstrated using emulated and hardware prototypes. Also, use cases for revealing M-PolKA’s functionalities are tested in different scenarios in order to address problems, such as communication reliability improvement, agile path migration and fast failure reaction. Finally, low overhead for extra functionalities is observed when RNS-based SR is compared to traditional routing approaches.

Self-driving networks guided by machine-learning (ML) algorithms are the driving force for building networks of the future. ML is effective at making inferences about data that is too complex or too unpredictable for humans. The network softwarization enabled by a deep programmability approach opens up new opportunities to deploy ML at the programmable data plane. In this paper, we introduce the MAP4 as a framework that explores the feasibility of mapping ML models in programmable network devices. To achieve this, we rely on the P4 language to deploy a pre-trained model into a programmable switch, utilizing the ML model to accurately classify flows at line rate. Our approach demonstrates that ML models working as classifiers can better fit the data by using the new levels of network programmability from the P4 language. The results showed that with few packets, most of the flows are properly classified. In some use cases, with two packets in the flow, 97% of traffic can be correctly classified, and all classes are properly labeled with a maximum of four packets.

In-Network Computing is a promising field that can be explored to leverage programmable network devices to offload computing towards the edge of the network. This has created great interest in supporting a wide range of network functionality in the data plane. Considering a networked robotics domain, this brings new opportunities to tackle the communication latency challenges. However, this approach opens a room for hardware-level exploits, with the possibility to add a malicious code to the network device in a hidden fashion, compromising the entire communication in the robotic facilities. In this work, we expose vulnerabilities that are exploitable in the most widely used flexible framework for writing robot software, Robot Operating System (ROS). We focus on ROS protocol crossing a programmable SmartNIC as a use case for In-Network Hijacking and In-Network Replay attacks, that can be easily implemented using the P4 language, exposing security vulnerabilities for hackers to take control of the robots or simply breaking the entire system.

Deploying accurate machine learning algorithms into a high-throughput networking environment is a challenging task. On the one hand, machine learning has proved itself useful for traffic classification in many contexts (e.g., intrusion detection, application classification, and early heavy hitter identification). On the other hand, most of the work in the area is related to post-processing (i.e., training and testing are performed offline on previously collected samples) or to scenarios where the traffic has to leave the data plane to be classified (i.e., high latency). In this work, we tackle the problem of creating simple and reasonably accurate machine learning models that can be deployed into the data plane in a way that performance degradation is acceptable. To that purpose, we introduce a framework and discuss issues related to the translation of simple models, for handling individual packets or flows, into the P4 language. We validate our framework with an intrusion detection use case and by deploying a single decision tree into a Netronome SmartNIC (Agilio CX 2x10GbE). Our results show that high-accuracy is achievable (above 95%) with minor performance degradation, even for a large number of flows.

5G relies on its pervasive and convergent cloud-based architecture to accomplish its futuristic challenge of being the next-generation communication platform. However, the new perspectives opened by 5G networks do not go unnoticed. Regardless of their motivation or objectives, cyberattackers find in the new 5G ecosystem, including its tenancy-driven control plane, an attractive greenfield to create new types of denial of services attacks. In this article, we leverage on the virtualised environment of 5G to propose REPEL – an intelligent resource scaling strategy to mitigate DDoS signalling attacks preserving legitimate traffic. Our prevention-based approach uses games theory to build up a defence front line, able to keep services availability and discourage the attacker. To demonstrate the effectiveness and feasibility of our approach, we feed a queuing model with parameters obtained from a testbed, where simulated subscribers connect to a virtualised evolved packet core prototype. The final results show a dramatic signalling losses reduction, which can ensure the appropriate control plane availability under a DDoS attack.

Due to its ubiquitous use, WiFi may play an essential role in providing indoor connectivity for the new real-time services in several 5G verticals. However, there are still pressing issues to be addressed, requiring new mobility management schemes to guarantee reliable and low latency communications. In this paper, we propose a novel SDN-NFV based architecture with a low cost off-the-shelf WiFi that explores a multiconnectivity scheme at the user devices. In order to demonstrate the feasibility of our approach, we developed a prototype and performed experiments on seamless handover using: i) an SDNNFV based packet duplication solution; and ii) a source-routing solution for end-to-end communication. Results show that the proposed architecture can provide an efficient seamless handover, increasing the likelihood of delivering packets with minimal effects on latency.

Network Function Virtualization (NFV) ambitiously aims at shifting functionalities from proprietary hardware to commodity servers as well as providing a common platform for the consolidation of network management. In practice, however, network architects have not massively subscribed to NFV so far. One possible reason for the slow adoption of NFV is the trade-off between commoditization and end-to-end performance in core networks. On the other hand, the ever-increasing disparity between bandwidth requirements and computing power is already pushing network architects toward NFV. By using innetworking processors to co-execute parts of tenant's applications in the cloud, it is possible to offload packet processing tasks. We can also save end-host server CPU cores, while achieving lower latency for service requests. However, applications with an in-networking processor bring two main challenges: flexible programmability and manageable offloading constraints. This work proposes PIaFFE: a framework that uses P4 language for decomposing and deploying Virtual Network Functions (VNFs) into small embedded Network Functions (eNFs) on in-network processors. Our key contribution is allowing the correct balance between hardware capabilities from a programmable NIC, and the flexibility of traditional VNFs on vCPUs. It is only made possible by a novel multi-level chaining scheme in PIaFFE. An experimental use case focuses on the chaining of three eNFs and multi-tenant ready hypervisor, where the percentage of packet processing sharing with a VNF is varied for one of them. As a result, the appropriate offload factor can be found for different traffic conditions for either maximum throughput or minimum latency.

Source routing (SR) is a prominent alternative to table-based routing for reducing the number of network states. However, traditional SR approaches, based on Port Switching, still maintain a state in the packet by using a header rewrite operation. The residue number system (RNS) is a promising way of executing fully stateless SR, in which forwarding decisions at core nodes rely on a simple modulo operation over a route label. Nevertheless, such operation over integer arithmetic is not natively supported by commodity network hardware. Thus, we propose a novel RNS-based SR scheme, named PolKA, that explores binary polynomial arithmetic using Galois field (GF) of order 2. We evaluate PolKA in comparison to Port Switching by implementing emulated and hardware prototypes using P4 architecture. Results show that PolKA can achieve equivalent performance, while providing advanced routing features, such as fast failure reaction and agile path migration.

Machine Learning as a Service (MLaaS) platforms enables access to Machine Learning (ML) processing with scalable infrastructure, from anywhere, and at any time, but requires sending large amounts of data to the cloud. ML on the edge is emerging as an option to reduce latency and bandwidth usage, maintaining data privacy. However, the existing edge approaches are not aware of the current network state for orchestrating the tasks. Network- aware orchestration services are supported by the Network Function Virtualization (NFV) architecture, making it a promising approach to manage and place ML tasks. In this paper, we propose Machine Learning Function Virtualization (MLFV), a fully network-aware framework that explores the NFV environment to virtualize ML tasks as virtual network functions. We describe a novel model for placing chains of ML tasks, considering constraints on CPU, memory, the existence of ML libraries, and the network overload, aiming to reduce the overall execution time of a chain. The results showed that MLFV outperformed existing cloud and edge approaches, particularly when network connections present instabilities. MLFV was able to identify the irregularities, allocating the ML tasks on hosts with normal connections, and thus, reducing the time for classifying single and multiple concurrent requests.

The introduction of Next Generation Networks (NGN) has changed significantly the network service model. Vertical structures that are designed to each type of access have been replaced by horizontal structures. However, the service architecture has to be rebuilt to support the required agility within the horizontal structures. Our proposal, named SANGN, is based on a full integration of Open Services Gateway Initiative (OSGi) with Java Agent DEvelopment (JADE) framework for providing multimedia services of IP Multimedia Subsystem (IMS). We have introduced a Televote service as a use case in order to validate the SANGN architecture ensuring the massive scalability of these services to meet the stringent multimedia requirements. Our implementation is able to adapt with a high dynamic load of the system, making use of a minimal set of computational resources seamlessly for the clients, providing IMS core components in the network infrastructure.

Data analytics and scientific computing are two modern applications that in recent years have substantially changed their computation and communication needs, requiring additional processing capability and bandwidth to be able to keep pace with current demands. These applications are commonly processed within data centers, exchanging enormous volumes of data, rapidly stressing existing network infrastructures. Thus, it is crucial for data center operations and management to be able to understand and classify the communication demands of these applications. The traditional approaches for classifying application traffic are port-based and Deep Packet Inspection, both presenting issues with current network technology. Some recent works propose using machine learning plus statistical information collected from application flows to classify traffic. Applications running in data centers present communication patterns which can be recognized through their traffic matrices. So, the main contribution of this paper is a method that explores the textural information extracted from these matrices to classify the data center traffic using machine learning techniques. As a proof-of-concept, we implemented this method in a system named DCTraCS. The experimental dataset was gathered from two real data centers, collecting the traffic matrices of MapReduce and a set of scientific applications every second for a period of 30 minutes. For assessing our proposal, we compared it with other machine learning techniques for classifying application traffic found in current literature. Results show that our approach achieved the highest accuracy, classifying correctly over 99% of our data center applications.

Service Providers (SP) have deployed virtualized network functions (VNFs) as key elements for handling traffic in the SP domain. Packets are steered to follow a certain order through a set of VNFs before reaching their destination. The provision of such chaining is called Service Function Chaining (SFC). However, current SFC implementations require complex management and are tailored to specific platforms or devices. In this paper, we propose PhantomSFC, a fully virtualized SFC architecture that aims to clearly decouple the service plane from the underlying network. The approach enables SFC to be network agnostic, allowing it to be deployed in multiple scenarios. Besides, SPs can scale resources allocated to SFC execution according to service demands. A proof-of-concept (PoC) prototype of the PhantomSFC architecture is built as a DPDK application to demonstrate the feasibility of our approach. Results show it can handle SFC operation with reasonable performance, considering latency, jitter, and throughput.

Wireless networks have become in the last years a key enabling technology for cloud-enabled robots. Among those, the usage of WiFi is a first choice due to its almost ubiquitous use nowadays. However, WiFi suffers from crucial issues like spectrum interference, connectivity losses, long delay for client association and high latency handover. This work proposes a novel architectural split of the WiFi functionalities based on an enhanced software-defined wireless architecture. Cloud-enabled robots scenarios are addressed to derive results showing that the proposed architecture allows uninterrupted communication during handovers, and a quicker failover management.

Datacenter (DC) design has been moved toward the edge computing paradigm motivated by the need of bringing cloud resources closer to end users. However, the software defined networking (SDN) architecture offers no clue to the design of micro DCs (MDCs) for meeting complex and stringent requirements from next generation 5G networks. This is because canonical SDN lacks a clear distinction between functional network parts, such as core and edge elements. Besides, there is no decoupling between the routing and the network policy. In this paper, we introduce residue defined networking architecture (RDNA) as a new approach for enabling key features like ultra-reliable and low-latency communication in MDC networks. RDNA explores the programmability of residues number system as a fundamental concept to define a minimalist forwarding model for core nodes. Instead of forwarding packets based on classical table lookup operations, core nodes are tableless switches that forward packets using merely remainder of the division (modulo) operations. By solving a residue congruence system representing a network topology, we found out the algorithms and their mathematical properties to design RDNA's routing system that: 1) supports unicast and multicast communication; 2) provides resilient routes with protection for the entire route; and 3) is scalable for 2-tier Clos topologies. Experimental implementations on Mininet and NetFPGA SUME show that RDNA achieves 600 ns switching latency per hop with virtually no jitter at core nodes and sub-millisecond failure recovery time.

Edge Data Centres (EDC) are often managed by a single administrative entity with logically centralized control. The architectural split of control and data planes and the new control plane abstractions have been touted as Software-Defined Networking (SDN), where the OpenFlow protocol is one common choice for the standardized programmatic interface to data plane devices. However, in the design of an SDN architecture, there is no clear distinction between functional network parts such as core and edge elements. It means that all switches require to support lookups over hundreds of bits with complex actions that have to be specified by multiple tables. In this paper, we propose a new programmable architecture for EDC networks, named Residues Defined Networks (RDN). In RDN, a controller defines a network policy (e.g. connectivity protection) setting flow entries at the edges. Based on these entries, the edge switches assign routeIDs to flows. A route is defined as the remainder of the division (Residue) between a route-ID and a set of switch-IDs within RDN core. In case of failures, emergency routes are compactly encoded as programmable residues forwarding paths written into the packets. RDN scalability is evaluated considering 2-tier Clos topologies which cover mostly EDC deployments supporting up to 2304 servers. A RDN proof-of-concept prototype is implemented in Mininet for network emulation. Also, to increase the accuracy on latency measures, we implement RDN in NetFPGA that is validated in a testbed with 10Gbps Ethernet boards. RDN offers ultra-fast failure recovery (sub-milliseconds carrier grade), achieves low latency with RDN switching time per hop (≈0.6μs) and no jitter within the RDN core.

Scientific applications demand huge computational power connected through fast networks. They are developed using parallel kernel methods, usually implemented with the Message Passing Interface (MPI), presenting well-behaved communication patterns across computing nodes. The current network technologies do not allow defining traffic forwarding policies considering the different application traffic, resulting in an unbalanced load on the network links. Moreover, the devices are not concerned if the traffic is latency-sensitive or bandwidth-intensive. To handle this, we present NetSA, a framework exploiting the communication patterns of scientific applications, considering latency and bandwidth constraints, as the key logic for evenly placing the application flows on the network available paths. Through NetSA, the scientific application developer can easily modify the network behavior to best fit the application communication requirements. We have performed experiments for optimizing the MPI communication primitives and applied our solution to speed up scientific applications, obtaining an execution time reduction up to 27%.

Emerging paradigms, such as edge computing, require the geographical distribution of small-scale data centers that will use network functions virtualization (NFV) to provide new services with stringent demands related to throughput, latency, cost, innovation, and efficient orchestration. To tackle these demands, this paper proposes VirtPhy, a fully programmable architecture for NFV orchestration in edge data centers, based on server-centric topologies, software-defined networking, software switches, distributed service chaining, and source routing. A proof-of-concept is implemented in a data center testbed using OpenStack cloud platform in a hypercube topology. The results show its orchestration mechanisms can efficiently provision NFV service requests by steering traffic through a sequence of virtualized network functions in a server-centric topology.

Content Centric Networking (CCN) represents an important change in the current operation of the Internet, prioritizing content over the communication between end nodes. Routers play an essential role in CCN, since they receive the requests for a given content and provide content caching for the most popular ones. They have their own forwarding strategies and caching policies for the most popular contents. Despite the number of works on this field, experimental evaluation of different forwarding algorithms and caching policies yet demands a huge effort in routers programming. In this paper we propose SDCCN, a SDN approach to CCN that provides programmable forwarding strategy and caching policies. SDCCN allows fast prototyping and experimentation in CCN. Proofs of concept were performed to demonstrate the programmability of the cache replacement algorithms and the Strategy Layer. Experimental results, obtained through implementation in the Mininet environment, are presented and evaluated.

Scientific applications (SciApps) are broadly used in all science domains. For more accurate results, they have been increasingly demanding computational power and extremely agile networks. These applications are usually implemented using numerical methods presenting well-behaved patterns to exchange data across its computing nodes. This paper presents SpateN, a tool that exploits the spatial communication patterns of SciApps as the fundamental logic to drive the network programming. SpateN classifies the SciApps nodes communications and balances the elephant flows across the available network paths. As a proof of concept, we carried out a set of experiments in real testbeds, demonstrating that network programming may affect the performance of SciApps significantly. Also, a balanced flow allocation can speed up SciApps to near-optimal execution times.

Software Defined Networking (SDN) is a promising architectural approach based on a programmatic separation of the control and data planes. For high availability purposes, logically centralized SDN controllers follow a distributed implementation. While controller role features in the OpenFlow protocol allow switches to communicate with multiple controllers, these mechanisms alone are not sufficient to guarantee a resilient control plane, leaving the actual implementation as open challenge for SDN designers. This paper explores OpenFlow roles for the design of resilient SDN control plane and proposes AR2C2 as an actively replicated multi-controller strategy. As proof of concept, AR2C2 is implemented based on the Ryu controller and relying on OpenReplica to ensure consistent state among the distributed controllers. Our prototype is experimentally evaluated using real commodity switches and Mininet emulated environment. Results of the measured times to recover from failures for different workloads shed some light on the practical trade-offs on replication overhead and latency as a step forward towards SDN resiliency.

Recent proposals on Data Center Networks (DCN) are based on centralized control and a logical network fabric following a well-controlled baseline topology. The architectural split of control and data planes and the new control plane abstractions have been touted as Software-Defined Networking (SDN), where the OpenFlow protocol is one common choice for the standardized programmatic interface to data plane devices. In this context, source routing has been proposed as a way to provide scalability, forwarding flexibility and simplicity in the data plane. One major caveat of source routing is network failure events, which require informing the source node and can take at least on the order of one RTT to the controller. This paper presents SlickFlow, a resilient source routing approach implemented with OpenFlow that allows fast failure recovery by combining source routing with alternative path information carried in the packet header. A primary and alternative paths are compactly encoded as a sequence of segments written in packet header fields. Under the presence of failures along a primary path, packets can be rerouted to alternative paths by the switches themselves without involving the controller. We evaluate SlickFlow on a prototype implementation based on Open vSwitch and demonstrate its effectiveness in a Mininet emulated scenario for fat-tree, BCube, and DCell topologies.

As network topologies have grown in size and complexity, it is becoming a daunting task for network administrators to keep track the capacity dimensioning of newly installed web-servers within a single or multiple providers. In fact, monitor capacity dimensioning is not a trivial activity since network state changes rather frequently, in particular, in academic environments. In this paper, we describe estimation algorithms and the software architecture of an efficient network management suite to automatically mine path capacity and minimum delays from a venture point to a set of observed web servers. The principle of the suite is based on packet dispersion techniques and repetitive non-intrusive measurements. We provide analytical insights, simulation results and some real case studies where we argument about the correctness, accuracy and usefulness of the suite in the context of management and operation of complex IP based networks.

In this paper, we introduce a simple analytical modeling approach for computing service unavailability due to long response time, for infinite and finite single-server systems as well as for multi-server systems. Closed-form equations of system unavailability based on the conditional response time distributions are derived and sensitivity analyses are carried out to analyze the impact of long response time on service unavailability. The evaluation provides practical quantitative results that can help distributed system developers in design decisions

International audience