Constraint-Based Anycasting Over Optical Burst Switched Networks

2017

J. Opt. Commun. Netw.

Abstract-In cloud-like scenarios, demand is served at one of multiple possible data center (DC) destinations. Usually, which DC exactly is used can be freely chosen, which leads to an anycast routing problem. Furthermore, the demand volume is expected to change over time, e.g., following a diurnal pattern. Given that virtually all application domains today heavily rely on cloud-like services, it is important that the backbone networks connecting users to the DCs is resilient against failures. In this paper, we consider the problem of resiliently routing multi-period traffic: we need to find routes to both a primary DC and a backup DC (to be used in case of failure of the primary one, or the network connection towards it), and also account for synchronization traffic between primary and backup DCs. We formulate this as an optimization problem and adopt column generation, using a path formulation in two sub-problems: the (restricted) master problem (RMP) selects "configurations" to use for each demand in each of the time epochs it lasts, while the pricing problem (PP) constructs a new "configuration" that can lead to lower overall costs (which we express as the number of network resources, i.e., bandwidth, required to serve the demand). Here, a "configuration" is defined by the network paths followed from the demand source to each of the two selected DCs, as well as that of the synchronization traffic in between the DCs. Our decomposition allows for PPs to be solved in parallel, for which we quantitatively explore the reduction of the time required to solve the overall routing problem. The key question that we address with our model is an exploration of the potential benefits in rerouting traffic from one time epoch to the next: we compare several (re)routing strategies, allowing traffic that spans multiple time periods to (i) not be rerouted in different periods, (ii) only change the backup DC and routes, or (iii) freely change both primary and backup DC choices and routes towards them.

Section: A Related Workmentioning

confidence: 99%

Exploring the Benefit of Rerouting Multi-Period Traffic to Multi-Site Data Centers [Invited]

Wang

Jaumard

2017

J. Opt. Commun. Netw.

“…(For works on optical burst switching (OBS) anycast, we refer to [14], [15].) The anycast routing problem in OCS WDM networks amounts to anycast routing and wavelength assignment (ARWA), finding routes for each anycast request while, e.g., minimizing the total number of wavelengths used, and/or the load on the links [16].…”

Section: Related Workmentioning

confidence: 99%

“…The PP's decision variables and its auxiliary variables have the same definitions as before, but we drop the index. The first set of the following equation represents the flow conservation equations, expressing that the net flow going into a node should be either 1 (for the source node), 1 (for a destination node), or 0 otherwise: if if otherwise and (15) Next, constraints (16) assure that there will be no loops, and exactly one working and backup path will be constructed. Additionally, (17) enforces that a single working and backup destination will be chosen.…”

Section: ) Ppmentioning

confidence: 99%

Joint Dimensioning of Server and Network Infrastructure for Resilient Optical Grids/Clouds

IEEE/ACM Trans. Networking

Buysse

Dhoedt

et al. 2014

We address the dimensioning of infrastructure, comprising both network and server resources, for large-scale decentralized distributed systems such as grids or clouds. We design the resulting grid/cloud to be resilient against network link or server failures. To this end, we exploit relocation: Under failure conditions, a grid job or cloud virtual machine may be served at an alternate destination (i.e., different from the one under failurefree conditions). We thus consider grid/cloud requests to have a known origin, but assume a degree of freedom as to where they end up being served, which is the case for grid applications of the bag-of-tasks (BoT) type or hosted virtual machines in the cloud case. We present a generic methodology based on integer linear programming (ILP) that: 1) chooses a given number of sites in a given network topology where to install server infrastructure; and 2) determines the amount of both network and server capacity to cater for both the failure-free scenario and failures of links or nodes. For the latter, we consider either failure-independent (FID) or failure-dependent (FD) recovery. Case studies on Europeanscale networks show that relocation allows considerable reduction of the total amount of network and server resources, especially in sparse topologies and for higher numbers of server sites. Adopting a failure-dependent backup routing strategy does lead to lower resource dimensions, but only when we adopt relocation (especially for a high number of server sites): Without exploiting relocation, potential savings of FD versus FID are not meaningful.

“…This has been considered in packet switched IP networks (Tim Stevens et al, 2007), or even optical burst switched (OBS) networks (Bathula & Elmirghani, 2009;De Leenheer et al, 2006). As indicated above, we consider anycast routing in optical circuit-switched (OCS) networks, where it amounts to the so-called anycast routing and wavelength assignment (ARWA) problem.…”

Section: Anycast Routing and Wavelength Assignment (Arwa)mentioning

confidence: 99%

“…An online routing problem (while also briefly raised by Walkowiak, 2010) is studied specifically by Bhaskaran et al, who assume the number of anycast sites varies over time, according to a time-varying load (Bhaskaran, Triay, & Vokkarane, 2011). In (Bathula & Elmirghani, 2009), the authors propose anycast routing methods to improve the performance of reconfigurable WDM networks under the variations in the IP traffic. In (Tim Stevens, De Leenheer, De Turck, Dhoedt, & Demeester, 2006), authors show that the anycast routing problem can be reduced to unicast routing.…”

Section: Anycast Routing and Wavelength Assignment (Arwa)mentioning

confidence: 99%

Dimensioning Resilient Optical Grid/Cloud Networks

Jaumard

Advances in Systems Analysis, Software Engineering, and High Performance Computing

Optical networks play a crucial role in the provisioning of grid and cloud computing services. Their high bandwidth and low latency characteristics effectively enable universal users' access to computational and storage resources that thus can be fully exploited without limiting performance penalties. Given the rising importance of such cloud/grid services hosted in (remote) data centers, the various users (ranging from academics, over enterprises, to non-professional consumers) are increasingly dependent on the network connecting these data centers, that must be designed to ensure maximal service availability, i.e., minimizing interruptions. In this chapter we will outline the challenges encompassing the design, i.e., dimensioning, of large-scale backbone (optical) networks interconnecting data centers. This amounts to extensions of the classical routing and wavelength assignment algorithms (RWA) to so-called anycast RWA, but also pertains to jointly dimensioning not just the network but also the data center resources (i.e., servers). We specifically focus on resiliency, given the criticality of the grid/cloud infrastructure in today's businesses, and, for highly critical services, we also include specific design approaches to achieve disaster resiliency. INTRODUCTIONBack in the 1960s, John McCarthy envisioned the concept of "computation as a public utility", making computing power equally easily accessible as the classical utilities that provide users with water, gas, and electricity. That seminal idea reappeared in the 1990s under the form of grid computing, borrowing its name from the power grid, where "the grid" was aimed to be a highly powerful computing resource that scientists could easily tap into for performing challenging tasks. Similarly, today's cloud computing paradigm is built on the idea of relieving the user from worrying about the resources required to run applications and to store data, as well as on the idea of enabling access to such applications and data from basically any device. Clearly, such concept can be made possible only through a high capacity and low latency network that connects the user to "the cloud", i.e., the distributed computing/storage resources. Undeniably, development of optical network technology has been a major driver that enabled the realization of such grids/clouds. The rise of broadband access networks, and high speed optical networking in wide area networks (WAN) has increased the geographical scale of distributed computing paradigms, extending their range from on-site computing facilities to the cost-efficient aggregation of IT resources for both processing and storage in large scale data centers. These now can supply a broad spectrum of applications, serving a wide audience ranging from end consumers, over business users, to scientists requiring high performance computing (HPC) facilities. Basic concepts underlying so-called grid technology, originating in the e-Science domain (e.g., to process massive data flows from the large hadron collider (LHC) at CE...