The challenge of how to make networking technology "greener", i.e. more environmentally sustainable, has recently been gaining a lot of attention. One key to potential solutions concerns the ability to have visibility into metrics that can be used to assess and optimize networks’ environmental footprint. While past attention has focused on individual device metrics, such as power consumption, certain use cases would also benefit from more holistic metrics that apply to networking paths.This paper describes an approach for instrumenting a network with metrics that indicate the environmental efficiency of paths, obtained by aggregating a set of configurable device efficiency metrics along networking paths. The aggregation occurs using a new extension to In-situ Operations Administration and Maintenance (IOAM) as a packet traverses hops in the data plane, without depending on controllers or needing to collect individual data records that might lead to Maximum Transmission Unit (MTU) issues associated with other versions of IOAM. Particular emphasis is given to the efficiency of aggregation operations to minimize performance impact. A proof of concept has been developed using P4.

Data center traffic management challenges increase with the complexity and variety of new Internet and Web applications. Efficient network management systems are often needed to thwart delays and minimize failures. In this regard, it seems helpful to identify in advance the different classes of flows that (co)exist in the network, characterizing them into different types based on different latency/bandwidth requirements. In this paper, we propose Howdah, a traffic identification and profiling mechanism that uses Machine Learning and a load-aware forwarding strategy to offer adaptation to different classes of traffic with the support of programmable data planes. With Howdah, the sender and gateway elements inject in-band traffic information obtained by a supervised learning algorithm. When a switch or router receives a packet, it exploits this host-based traffic classification to adapt to a desirable traffic profile, for example, to balance the traffic load. We compare our solution against recent traffic engineering proposals and demonstrate the effectiveness of the cooperation between host traffic classification and P4-based switch forwarding policies, reducing packet transmission time in data center scenarios.

Autonomic Networking has long been a topic for standardization at the IETF. The first phase of standardization was mostly concerned with providing infrastructural foundations: Autonomic and secure enrollment of devices into a network domain and establishment of an autonomic control plane to support secure communication between autonomic service agents, on top of which actual autonomic functionality could subsequently be built. At the same time, Intent-Based Networking has been emerging as a new management paradigm, concerned with managing networks by defining outcomes without specifying how they might be achieved, which could in turn require autonomic functionality. In this position paper we outline ways in which Autonomic and Intent-Based Networking can complement one another and explore the role that interoperability and standards can play in this space. From this, we outline gaps and identify opportunities in current IETF standardization and offer our conclusions regarding which directions IETF standardization may take in the future.

One of the next waves of networking innovation is expected to be driven by the needs of applications that require communication services with stringent service level guarantees, for example near-deterministic and ultra-low end-to-end latency. Providing such services is riddled with challenges due to limitations with existing QoS mechanisms and application characteristics that are not tolerant of even slight degradation. In addition, to be deployable in practice, advances in accounting technology to be able to verify compliance of services with their guarantees will also be sorely needed, resulting in a separate set of challenges. This industry vision paper provides an overview of the problems, highlights some solution approaches, and articulates a set of very practical research challenges that, when successfully addressed, will have significant practical impact.

The challenges of managing datacenter traffic increase with the complexity and variety of new Internet and Web applications. Efficient network management systems are often required to thwart delays and minimize failures. In this regard, it appears helpful to identify in advance the different classes of flows that (co)exist in the network, characterizing them into different types according to the different latency/bandwidth requirements. In this paper, we propose Howdah, a traffic identification and profiling mechanism that uses Machine Learning and a congestion-aware forwarding strategy to offer adaptation to different traffic classes with the support of programmable data-planes. With Howdah, sender and gateway elements inject in-band traffic information obtained using supervised learning. When a switch or a router receives a packet, it exploits such host-based traffic classification to adapt to a desirable traffic profile, for example, balancing the load. We compare our solutions against recent traffic engineering solutions and show the efficacy of cooperation between host traffic classification and P4-based switch forwarding policies, reducing packet transmission time in datacenter scenarios.

Translating high-level policies to lower-level network rules is one of the main goals of control or data plane network programmability. To further abstract requirements and propel automation in networking, several industries have proposed the paradigm of “network intent”. However, the translation from intents to low-level policies is considered critical to program data planes and other network elements, especially when dealing with P4-enabled switches. In this paper, we present NLP4, an architecture that helps translate intents, in the form of human language, into data-plane programs, in the form of P4 rules. In particular, NLP4 uses Natural Language Processing (NLP) techniques to translate high-level human-language intents, a MultiLayer Perceptron (MLP) model for processing the NLP output and converting it into mid-level policy. An API then uses this information, which separates the intent from the network to generate commands readable by P4-enabled switches. Our initial prototype on a network emulator validates our architecture for a specific case: load profiling, demonstrating how even users with limited P4 expertise may customize their networks by merely specifying intents.

Reducing mankind’s carbon footprint and becoming more sustainable is rapidly becoming one of the main challenges of our age. The networking industry needs to respond to those challenges not only by enabling applications that allow users to save energy (such as, replacing travel with teleconferencing), but also by making networking technology itself significantly more energy-efficient. As a result, the aspect of minimizing network power consumption is rapidly increasing in importance. This position paper aims to articulate opportunities where energy efficiency of networks can be improved, to point out the associated technical challenges, and to identify corresponding research challenges.

We present Token Cell Routing (TCR), a new sub-IP layer protocol that provides a powerful yet hardware-friendly method of constructing data plane packets to meet the needs of new applications. Token cells are a special kind of lightly structured token which can be programmed with different semantics to serve different purposes. A TCR packet consists of a series of token cells, each associated with a certain functionality. Using longest prefix matching, code points are invoked that implement logic specific to the token cell's type. Packets also define token cell processing workflows and allow token cells to cross-reference data. This results in the ability to program very powerful functionality. At the same time, processing is kept simple, allowing for straightforward implementations able to operate at line rate. TCR is currently still in the concept stage while proofs-of-concepts and venues for standardization are being explored.

This paper contributes gLBF (guaranteed Latency-Based Forwarding), a new algorithm for leaky bucket defined traffic flows, feasible for low-cost, high-speed forwarding planes in LAN/RAN/WAN networks to support simply computed, per-hop bounded latency for high-precision communication services. gLBF combines benefits of earlier algorithms such as Urgency-Based Scheduling (UBS) and Cyclic Queuing and Forwarding (CQF) while avoiding their downsides. gLBF does not require per-hop, per-flow state such as shapers or (interleaved) regulators, thus avoiding control complexity and limiting hardware cost. It does not require strict network-wide time synchronization and therefore eliminates the opex and capex of deploying IEEE1588. It minimizes end-to-end jitter, thus eliminating the need for playout buffers in endpoints to compensate for network jitter.

The networking landscape is expected to undergo profound changes over the coming years, with new applications emerging that depend on the ability of networks to provide high-precision services with stringent service-level guarantees. To enable and incentivize network providers to deploy and deliver such services will also require advances in the way such services can be accounted for. At its core, this will require support for high-precision metering that allows to generate custom accounting records that can be custom tailored towards specific service offerings and service level agreements. This paper presents a framework to enable such metering.

This paper describes experiences gained during the development of a Proof-of-Concept implementation of NewIP/BPP, the protocol at the core of a novel packet-programmable networking framework, using P4, a popular SDN technology for the implementation of networking protocols using protocol-independent packet processors. NewIP/BPP introduces a number of novel requirements whose implementation encountered a number of P4 limitations that proved very challenging to overcome. We hope that the resulting insights will be useful for future implementations of NewIP/BPP as well as for its and P4's evolution.

In a companion paper [1], we introduced Latency-Based Forwarding (LBF) as a new approach to achieving stringent end-to-end latency objectives with precisely specified lower and upper bounds. In order to accomplish this, a new forwarding protocol is used that can carry latency objectives as part of packet metadata and take differentiated actions at network nodes depending on a packet’s remaining latency budget, its destination, and the latency encountered so far. LBF is intended to enable new emerging network applications such as Industry 4.0, haptic applications, and holographic-type communications which require support for corresponding high-precision network latency services that are not possible with current best-effort Internet or QoS mechanisms.This demo shows a near real-time, lightweight network emulator for LBF. Individual network nodes are emulated by separate cores on a host. As part of the demo, we vary the end-to-end latency objectives of multiple competing simultaneous flows and compare minimum, maximum, average packet latencies across different scenarios. This demonstrates the behavior and effectiveness of LBF and how it affects flows and packets.

New emerging network applications such as Industry 4.0, haptic applications, and holographic-type communications will require support for stringent end-to-end latency objectives with precisely specified lower and upper bounds. These requirements will push current best-effort Internet technology and QoS mechanisms to their limit. This paper presents a new approach to achieve high-precision latency objectives, carrying latency objectives as part of packet metadata and taking differentiated actions at network nodes depending on a packet’s remaining latency budget, its destination, and the latency encountered so far. A Proof-of-Concept has been developed using BPP, a novel programmable networking framework.

Next generation application domains such as Virtual Reality (VR), Augmented Reality (AR) together with the Tactile Internet paradigm impose ultra-low latency requirements on the networks (1 to 5 ms end-to-end latency). Towards this objective, networks are undergoing a tremendous transformation from the current packet switching models to Software Defined Networking (SDN) architectures, which provide programmability to configure the network. In its simplest variant, one single centralized controller orchestrates the whole SDN infrastructure. However, the fully centralized architecture (one single controller) can become a performance bottleneck, especially in terms of response throughput and flow setup latency. Furthermore, it suffers from massive scalability issues. In this direction, a number of more sophisticated SDN architectures are currently under research. While their theoretical advantages have been thoroughly discussed in the state-of-the-art, a comparative experimental analysis of these architectures is still missing. This study aims at providing such experimental performance comparison. Herein, we put to test a set of SDN architectures ranging from a fully centralized to a completely distributed control plane, comparing them in terms of flow setup latency. Overall results show that completely distributed architectures provide significantly better performance with almost 31% gain in terms of average flow setup latency over the centralized case.

Segment Routing (SR) reintroduces source routing to networking. While SR has been defined for MPLS and IPv6 data planes, there are considerable problems with respect to increased path overhead in various deployments. This paper presents a new framework that is designed to overcome the SR challenges with a new routing paradigm, PPR (Preferred Path Routing). PPR is intended as enabler for next generation source routing by minimizing the data plane overhead caused in SR, which includes the network layer tax and processing overhead that is imposed on packets, critical specifically for small packets that are characteristic to many 5G applications. PPR extends SR for IP data planes without needing to replace existing hardware or even to upgrade the data plane. In addition, PPR allows dynamic path QoS reservations viz., bandwidth, and resources for providing deterministic queuing latency.

Analytics of network telemetry data is useful for addressing many important network operational problems. While Big Data techniques have been pushing scale boundaries for processing data ever further, in many cases the real bottleneck for analytics is the acquisition, i.e. the ability to generate and export the data on which analytics depends. To address this issue, we have earlier introduced DNA, a framework for Distributed Network Analytics that pushes analytics processing into the network and dynamically sets up data sources as needed. One of the challenges of such a framework concerns providing users with simple ways to articulate network analytics queries and instruct the network which data to generate and provide. We have addressed this issue using a model-driven approach that is presented in this paper. Using YANG as a way to model network analytics tasks, our system lets users articulate network analytics tasks at a very high level of abstraction that is subsequently broken down by the framework into lower-level analytics tasks which are deployed across the network.

Information in networked systems often has spatial properties: routers, sensors, or virtual machines have coordinates in a geographical or virtual space, for instance. In this paper, we propose a peer-to-peer design for a spatial search system that processes queries, such as range or nearest-neighbor queries, on spatial information cached on nodes inside a networked system. Key to our design is a protocol that creates a distributed index of object locations and adapts to object and node churn. The index builds upon the concept of the minimum bounding rectangle, to efficiently encode a large set of locations. We present a search protocol, which is based on an echo protocol and performs query routing. Simulations show the efficiency of the protocol in pruning the search space, thereby reducing the protocol overhead. For many queries, the protocol efficiency increases with the network size and approaches that of an optimal protocol for large systems. The protocol overhead depends on the network topology and is lower if neighboring nodes are spatially close. As a key difference to works in spatial databases, our design is bottom-up, which makes query routing network-aware and thus efficient in networked systems.

Analytics of network telemetry data helps address many important operational problems. Traditional Big Data approaches run into limitations even as they push scale boundaries for processing data further. One reason for this is the fact that in many cases, the bottleneck for analytics is not analytics processing itself but the generation and export of the data on which analytics depends. The amount of data that can be reasonably collected from the network runs into inherent limitations due to bandwidth and processing constraints in the network itself. In addition, management tasks related to determining and configuring which data to generate lead to significant deployment challenges. In order to address these issues, we propose a novel distributed solution to network analytics that we have implemented as a proof of concept, called DNA (Distributed Network Analytics). In DNA, analytics processing is performed at the source of the data by specialized agents embedded within network devices, which also dynamically set up and reconfigure telemetry data sources as required by an analytics task. An SDN controller application orchestrates network analytics tasks across the network to allow users to interact with the network as a whole instead of individual devices one at a time. Our demonstration of DNA includes a GUI front end used to specify and monitor analytics tasks as well as visualize analytics results.

Analytics of network telemetry data helps address many important operational problems. Traditional Big Data approaches run into limitations even as they push scale boundaries for processing data further. One reason for this is the fact that in many cases, the bottleneck for analytics is not analytics processing itself but the generation and export of the data on which analytics depends. The amount of data that can be reasonably collected from the network runs into inherent limitations due to bandwidth and processing constraints in the network itself. In addition, management tasks related to determining and configuring which data to generate lead to significant deployment challenges. In order to address these issues, we propose a novel distributed solution to network analytics. Analytics processing is performed at the source of the data by specialized agents embedded within network devices, which also dynamically set up and reconfigure telemetry data sources as required by an analytics task. An SDN controller application orchestrates network analytics tasks across the network to allow users to interact with the network as a whole instead of individual devices one at a time. The solution has been implemented as a proof-of-concept, called DNA (Distributed Network Analytics)1.

Critical networked services are usually regulated by Service Level Agreements (SLAs). In order to ensure SLAs are being met, it is necessary to monitor Service Level Objectives (SLOs). Active measurement mechanisms are usually chosen to perform this monitoring task, which requires measurement probes to be activated in network devices. However, these probes are expensive in terms of computational resources consumption, thus, active measurement mechanisms usually can cover only a fraction of what could be measured, which can lead to SLA violations being missed. Besides that, highly dynamic networking patterns require the ongoing selection of the candidate network destinations for probing and their respective prioritization, a practice that is not well suited for human administrators because configuring the probes is labor-intensive and error-prone. A possibility to improve the detection of SLA violations is the employment of Peer-to-Peer (P2P) technology in order to steer tasks related to a distributed decision making process for probe activation. In this context, a P2P management overlay can be used to coordinate the probe activation and to share measurement results among the network devices. For a node to rely on measurement data from a peer to determine which probes to configure, it needs to know which of the peers are best correlated with itself, i.e., which nodes have the most significance in terms of being indicative of service level violations that might be observed by the node itself. We propose an autonomic P2P solution to coordinate the placement of active measurement probes in large-scale networks. The edge nodes of the network cooperate via a P2P management overlay to determine what destinations should be monitored. Each edge node determines autonomously what destinations to probe considering local measurements and measurement data from other edge nodes. The measurements considered are traffic information from passive measurement results and past service level measurement results from active measurement results. The proposed solution is evaluated using simulation and the results show its feasibility and interesting features.

Network search makes operational data available in real-time to management applications. In contrast to traditional monitoring, neither the data location nor the data format needs to be known to the invoking process, which simplifies application development, but requires an efficient search plane inside the managed system. The search plane is realized as a network of search nodes that process search queries in a distributed fashion. This paper introduces matching and ranking for network search queries. We are proposing a semantic for matching and ranking, which is configurable to support different types of management applications-from exact matching for database-style queries to loose, approximate matching, which is appropriate for exploratory purposes. We describe an echo protocol for efficient distributed query processing that supports matching and ranking. Further, we present the design of a search node, which maintains a real-time database of operational information and allows for parallel processing of search queries. A prototype implementation on a cloud testbed shows that the network search system, on a 9-node cluster with 24 core servers, executes 200 global search queries/sec with the 75th percetile latency below 100 milliseconds and with a CPU utilization below 5%. The performance measurements, together with our design, suggest that a system of 100,000 servers processing the same load would exhibit the same overhead per server and a query latency of below 1 sec.

Critical networked services enable significant revenue for network operators and, in turn, are regulated by Service Level Agreements (SLAs). In order to ensure SLAs are being met, service levels need to be monitored. One technique for this involves active measurement mechanisms which employ measurement probes along the network to inject synthetic traffic and compute the network performance. However, these mechanisms are expensive in terms of resources consumption. Thus, active measurement mechanisms usually can cover only a fraction of what could be measured, which can lead to SLA violations being missed. Besides that, the definition of this fraction is a practice done by human administrators, which does not scale well and does not adapt to highly dynamic networking patterns. The contribution of the present work is the proposal of a solution to increase the potential number of detected SLA violations in which network devices autonomously and dynamically share service level measurement results. The sharing of these measurement results is based on the utilization of a Peer-to-Peer (P2P) management overlay built using past service level measurement results and a coordination strategy characterized by a high degree of decentralized decision making. The solution is evaluated using simulation and the results show its feasibility and interesting features.

A Cloud is a very dynamic environment where resources offered by a Cloud Service Provider (CSP), out of one or more Cloud Data Centers (DCs) are acquired or released (by an enterprise (tenant) on-demand and at any scale. Typically a tenant will use Cloud service interfaces to acquire or release resources directly. This process can be automated by a CSP by providing auto-scaling capability where a tenant sets policies indicating under what condition resources should be auto-scaled. This is specially needed in a Cloud environment because of the huge scale at which a Cloud operates. Typical solutions are naïve causing spurious auto-scaling decisions. For example, they are based on only thresholding triggers and the thresholding mechanisms themselves are not Cloud-ready. In a Cloud, resources from three separate domains, compute, storage and network, are acquired or released on-demand. But in typical solutions resources from these three domains are not auto-scaled in an integrated fashion. Integrated auto-scaling prevents further spurious scaling and reduces the number of auto-scaling systems to be supported in a Cloud management system. In addition, network resources typically are not auto-scaled. In this paper we describe a Cloud resource auto-scaling system that addresses and overcomes above limitations.

While networked systems hold and generate vast amounts of configuration and operational data, this data is not accessible through a simple, uniform mechanism. Rather, it must be gathered using a range of different protocols and interfaces. Our vision is to make all this data available in a simple format through a realtime search process which runs within the network and aggregates the data into a form needed by applications - a concept we call network search. We believe that such an approach, though challenging, is technically feasible and will enable rapid development of new management applications and advanced network functions. This paper motivates and formulates the concept of network search, compares it to related concepts like web search, outlines a search architecture, describes the design space and research challenges, and reports on a testbed implementation with management applications built for exploratory purposes of this new paradigm.

Proactive service assurance depends on the ability to anticipate problems in a network. This paper presents a system to forecast high-severity system event messages before they occur. Forecasts are based on a stream of syslog messages from devices, which is compared and scored against a set of codes that are representative of known problem event patterns. The system includes a component to mine relevant event patterns from past problem occurrences. In addition to describing the techniques and algorithms that the system is based on, the paper also provides an assessment of the effectiveness of those techniques.

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.

We examine the use of gossip protocols for continuous monitoring of network-wide aggregates. Aggregates are computed from local management variables using functions such as AVERAGE, MIN, MAX, or SUM. A particular challenge is to develop a gossip-based aggregation protocol that is robust against node failures. In this paper, we present G-GAP, a gossip protocol for continuous monitoring of aggregates, which is robust against discontiguous failures (i.e., under the constraint that neighboring nodes do not fail within a short period of each other). We formally prove this property, 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 G-GAP 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.

This paper presents a service on demand management framework to support dynamic service creation, and a system which instantiates this framework for an Ethernet-To-The-Home/Business (ETTx) service environment. The service-on-demand framework delivers two major functions: (1) it allows service providers to create new types of service dynamically through a policy-driven "service building" process; (2) it allows end users to subscribe to a service and have their service orders fulfilled in a fully automated, near real time, and "zero touch" fashion. We have applied this framework to build an industrial strength service fulfillment system geared towards ETTx access networks. Those networks represent one of the most dynamic service creation environments today.

Summary form only given. This tutorial presents a view of the IP network management problem from the perspective of the network itself. It focuses on how to address enhanced manageability requirements in the network infrastructure. This includes what can and should be instrumented inside a network element, and what can and should be instrumented with a scope that goes beyond the individual network element to result in more general "manageability services" provided by the network. We discuss the relevance of manageability as a business driver and provide an in-depth analysis of the various factors that contribute to manageability. After this, techniques are presented that can be applied to enhance manageability, also in light of increased focus on applications such as on-demand services, managed services, and plug-n-play networking. We also discuss the role of certain technologies such as XML and Netflow in the emergence of a new generation of management interfaces. Finally, scenarios are used to illustrate how the resulting advances in manageability facilitate not only the development of management applications, but ultimately result in networks that are more autonomous and "self managing", and less complex and costly to operate than the norm today.

With large-scale deployments of open packet telephony (OPT) in service provider networks, effective OPT management solutions become a necessity. The current state of the art is characterized by solutions that are fairly element-centric. In part, this comes from the fact that management information models in this domain have not been established and are in their infancy, in particular when it comes to manage OPT as a domain in a network context, beyond the network elements. This paper therefore presents an OPT information model. It addresses the need to model the network elements as well as the OPT application domain as a whole, providing a level of abstraction that allows us to hide much of the underlying distribution of voice functionality. How the model can be leveraged by management applications is also discussed. The presented concepts have been realized as part of an OPT management system.

Provisioning of residential voice and data subscriber services is a repetitive task and at the same time constitutes a key differentiation area for service providers. This paper presents an approach to subscriber service provisioning that is built around a concept of service and provision profiles. They allow services to be described in a network-independent manner and at the same time provide a formal method to define and derive the necessary network provisioning steps. Because they are machine interpretable, they can be leveraged to build provisioning systems that automate the subscriber provisioning process and are very easy to customize.

Open packet telephony (OPT) networks based on MGCP/H.248 control protocols present an excellent value proposition for service providers. Management of such networks, whose architecture significantly departs from that of traditional TDM networks, poses new challenges, such as the adaptation of existing operations support infrastructure. This paper explores these challenges in detail and presents an experimental management solution. One of its key aspects is the introduction of the concept of a virtual switch, which hides much of the complexity in dealing with the distributed nature of OPT networks.

In this presentation, the issues involved in integrating an existing non-TMN management system into a TMN environment are investigated. We draw on practical experiences gained in a project to integrate iRIDES, a system for the management of optical access networks, with a TMN service manager. Challenges encountered in the realization are exposed as well as proposals for solutions discussed. Major issues included the replication of inventory, the integration of GUIs, but most importantly the way in which management is being modeled: TMN's ambition is generally to provide generic models that will be applicable to any management application. As is the case with access networks, this can lead to models that are overly complex and therefore impractical for most applications. We argue therefore to follow a "TMN lite" approach which avoids many of these problems and, in our opinion, achieves a reasonable balance between TMN compliance and practicality. The proposal presents accordingly the attempt to preserve and indeed bring out TMN's true value while avoiding the unnecessary overhead and cost associated with it.