Masum Z. Hasan

Integrated and autonomic cloud resource scaling

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.

Hybrid networking: evolution toward combined IP and dynamic circuit services [Guest Editorial]

There is a wide consensus among service providers, operators, equipment vendors, and academics that the next decade will see highly evolving Internet and network transport architectures. Both circuit and packet switching technologies will be used in hybrid networks that will offer both IP packet and dynamic circuit services. A variety of switching technologies, such as multiprotocol label switching (MPLS), carrier Ethernet based switching, and wavelength-division multiplexed (WDM) optical switching, will be used to offer dynamic (virtual) circuit services, while IP routers will continue to be used for IP packet services. Integrated systems that offer both IP-layer forwarding and MPLS switching, or other combinations, will be used in these hybrid networks.

IP/MPLS OAM: challenges and directions

This paper reflects different understandings and positions on IP/MPLS OAM functions and management applications, as perceived by representatives from industry and academia. There is a clear effort to improve the state-of-the-art and to apply lessons learnt from the Ethernet/ATM approaches. A series of questions on the short-term, mid-term and long-term directions definitively arises, especially with regard to the separation of control and management functions. The questions mentioned below do not exhaustively cover the topic. Instead, the questions serve as a starting point for a discussion on methodologies to apply, on engineering requirements to consider, and on the technical challenges to resolve harnessing this next stage of evolution of heterogeneous networking. The panel guests present their perception on these issues. What are the lessons learnt from ATM and Ethernet OAM techniques that are oriented to a single network technology? What is equivalent of OAM appropriate for IP/[G]MPLS (it has to be more at link level rather than connection level)? What MPLS MIB information is intended to support OAM functions? What are the challenges on OAM in hybrid and multi-layer networks? How management systems of next generation networks (NGN) and GRID networks make use of the OAM functions and what else is needed to smoothen the control-management interactions? What is the impact of convergent IP networking and autonomic networking on OAM functions and their distribution? How much proactiveness can management applications achieve with and without OAM functionalities.

BLOMERS: Balanced Load Multi-Constrained Resource Scheduler

The process of resource management in local and wide-area distributed systems involves three main phases: resource discovery, resource scheduling and resource allocation. This paper presents a heuristic resource scheduler to cover the scheduling phase in large-scale distributed systems such as Grids. This approach is supported by BLOMERS (balanced load multi-constrain resource scheduler). Our heuristic scheduler implements a Genetic Algorithm (GA) in order to improve the scalability of the system, in large-scale Grids the amount of resource is massive but their computational capacity is low. Therefore, selection of computational resources to allocate customers services becomes an NP-hard problem. The paper outlines a performance evaluation by means of experimental tests. The results highlight two aspects: The scheduling system is reliable in real large-scale Grids and BLOMERS heuristic scheduler offers better performance than common selection algorithms when the number of resource requests is increasing.

Application and Network Resource Access Control

Controlling access to enterprise resources is of outmost importance for effective and secure functioning of an enterprise. Access control is provided in terms of authentication and authorization. The former verifies a user or an entity’s identity and credentials when access is attempted, whereas the latter deals with what actions are allowed on the resources to which access has been granted. A modern enterprise has to provide resource access control (RAC) to wide varieties of resources, from (ISO/OSI) layer 1 to layer 7 (L1–L7) resources. Typically, accesses to application, server, and storage (or L7) level resources are controlled by an application RAC (ARAC) system and that of network resources controlled by a network RAC (NRAC) system.

Intelligent healthcare computing and networking

There is need for providing proper quality of service (QoS) and security to healthcare traffic in the network in a healthcare environment. In addition, there is need for global visibility of healthcare related communications or transactions between various entities (applications, information systems and devices). There are three major factors that make this challenging: 1) healthcare network traffic is shared by other non-healthcare traffic, 2) there are plethora of healthcare information systems, devices and modalities with respective plethora of protocols and message formats, 3) healthcare network with its wide varieties of information systems, devices and modalities may be widely distributed (across various departments and wide-area network segments). To provide proper QoS and security in the midst of these challenges we bring healthcare application protocol and message related intelligence in the network, thus creating a logical overlay network (we call IHON: Intelligent Healthcare Overlay Network).

Guest editorial: Advances in IP-optical networking for IP quad-play traffic and services

As quad-play (data, voice, video and mobile communications) and high-performance computing (HPC) traffic are proliferating and converging on to public (Internet), private, enterprise or service providers’ IP networking infrastructure, there is need for new generation of networking infrastructure and relevant management of the enhanced infrastructure. This new generation network will include converged architectures of IP and Optical networking (IP–Optical). Optical networks are not to be anymore considered “peddlers of terabytes” but an intelligent and integral part of the overall network infrastructure. In this intelligent IP–Optical architecture, the traditional client–server model will transition to the controlplane based network with support for, among other features, automated end-to-end traffic engineering, comprehensive QoS and IP-over-DWDM. For the advanced capabilities of optical networks to flourish, the research community has to fundamentally rethink current architectures and protocols. This special issue covering invited and peer-reviewed papers, addresses this rethinking precisely, with promise to bring fundamental changes in optical networking architecture, protocol and management. We are happy to introduce four peer-reviewed and six invited articles that represent recent advances in the area of intelligent optical control planes for future IP–optical networking, including new research topics and the development of testbeds and applications. The first four articles address novel concepts and architectures from access to core topology. The paper “Light-Mesh—A Pragmatic Optical Access Network Architecture for IP-Centric Service Oriented Communication”, proposes a new light-mesh access network topology and related methods. The paper “Compact Photonic Gateway for Dynamic Path Control Using Acousto-optic Tunable Filter” proposes

SNMP-based monitoring agents and heuristic scheduling for large-scale grids

This paper presents both, SNMP-based resource monitoring and heuristic resource scheduling systems targeted to manage large-scale Grids. This approach involves two phases: resource monitoring and resource scheduling. Resource monitoring (even discovery) phase is supported by the SNMP-based Balanced Load Monitoring Agents for Resource Scheduling (SBLOMARS). This resource monitoring and discovery approach is different from current distributed monitoring systems in three main areas. Firstly, it reaches a high level of generality by the integration of SNMP technology and thus, it is offering an alternative solution to handle heterogeneous operating platforms. Secondly, it solves the flexibility problem by the implementation of complex dynamic software structures, which are used to monitor from simple personal computers to robust multi-processor systems or clusters with even multiple hard disks and storage partitions. Finally, the scalability problem is covered by the distribution of the monitoring system into a set of submonitoring instances which are specific per each kind of computational resource to monitor (processor, memory, software, network and storage). Resource scheduling phase is supported by the Balanced Load Multi-Constrain Resource Scheduler (BLOMERS). This resource scheduler is implemented based on a Genetic Algorithm, as an alternative to solve the inherent NP-hard problem for resource scheduling in large-scale Grids. We show some graphical and textual snapshots of resource availability reports as well as a scheduling scenario in the Grid5000 platform. We have obtained a scalable scheduler with an extraordinary load balanced between all nodes participating in the Grid.

Guest Editorial: Internet Traffic Engineering and Management

The best-effort Internet is evolving into the ubiquitous next-generation network through the support of rich (multimedia, multiservice, multiclass) traffic classes requiring appropriate quality of services, and optimal utilization of global network resources through optimized routing and traffic handling. The support of rich traffic classes and optimized utilization of network resources require the introduction of sophisticated traffic engineering features and mechanisms in next-generation network architectures, control, signaling, data, and management planes. Some of the major functions of traffic engineering are route computation, connection admission control (CAC), capacity planning, and monitoring of networks and services. Route computation and CAC may require satisfaction of various constraints, such as QoS preservation (bandwidth, delay, and jitter), optimized network resource utilization, cost minimization, or revenue maximization. The result of monitoring may be used as feedback for other traffic engineering functions. Traffic engineering also involves designing network architectures that are optimized for certain traffic classes requiring various degrees of QoS. For example, TDM architecture is suitable for circuit-switched voice traffic, but a different architecture may be well suited for multimedia traffic. The papers featured in this special issue of Journal of Network and Systems Management covers most of the above issues providing possible solutions. The first paper by J. Jun and S. Papavassiliou presents a routing method for computing QoS and network resource-constrained routes. The proposed method is called Optimal Least Weight Routing, where the weight is a function of a number of parameters, including shortest and alternate path trade-off, effect of low priority traffic on higher priority traffic, link utilization or load, and revenue. A route with the smallest weight is chosen. The computation, while computing a route for a connection with a particular class of service, tries to achieve the following: estimate the effect of a connection admission on the network, other classes of service, and future requests; minimize impact on higher priority traffic when connections with

Guest Editorial: IP Operations and Management

The classical IP best-effort network technology is evolving into a ubiquitous allservice networking infrastructure through the introduction of quality of service (QoS) and traffic engineering. Networks providing a rich set of complex features and supporting a relevant rich set of services, such as Web access with guaranteed QoS, voice over IP (VoIP), real-time video, multi-media multi-party conferencing, etc., require careful planning, operations, and management. There has to be a paradigm shift in managing this new breed of IP network. In such a network, QoS support will be provided through Differentiated Services (DiffServ), with Multi-Protocol Label Switching (MPLS) potentially deployed for traffic engineering. Both DiffServ and MPLS-enabled network elements need to be configured and monitored carefully, with end-to-end QoS path provisioning and monitoring becoming an important challenge. Service level agreements negotiated and customized through electronic interfaces will need to be automatically mapped to appropriate traffic engineering mechanisms, with QoS policy parameters set at various levels. Provisioning and configuration needs to take place together with dynamic control and reconfiguration through feedback from monitoring mechanisms. With more sophisticated services being provided, accounting and billing based on state-of-the-art usage metering and pricing mechanisms will become necessary. Simple monitoring of network health, which is mostly the case today, will not be adequate for providing quality services to customers. Operators running a complex network require tools that can make sense out of large amounts of monitoring data. Sophisticated health monitoring and analysis tools should be able to predict impending faults by analyzing and correlating traffic patterns. IP management systems should evolve out of passive remote (central) poll-oriented