Workload Migration Research Articles

Enhanced IPFIX flow monitoring for VXLAN based cloud overlay networks

<span lang="EN-US">The demands for cloud computing services is rapidly growing due to its fast adoption and the migration of workloads from private data centers to cloud data centers. Many companies, small and large, prefer switching their data to the enterprise cloud environment rather than expanding their own data centers. As a result, the network traffic in cloud data centers is increasing rapidly. However, due to the dynamic resource provisioning and high-speed virtualized cloud networks, the traditional flow-monitoring systems is unable to provide detail visibility and information of traffic traversing the cloud overlay network environment. Hence, it does not fulfill the monitoring requirement of cloud overlay traffic. As the growth of cloud network traffic causes difficulties for the service providers and end-users to manage the traffic efficiently, an enhanced IPFIX flow monitoring mechanism for cloud overlay networks was proposed to address this problem. The monitoring mechanism provided detail visibility and information of overlay network traffic that traversed the cloud environment, which is not available in the current network monitoring systems. The experimental results showed that the proposed monitoring system able to capture overlay network traffic and segregated the tenant traffic based on virtual machines as compare to the standard monitoring system.</span>

Workload forecasting based elastic resource management in edge cloud

Cloud services are provided at the edge of the network so that data from users can be processed and calculated at the edges. The user's irregular access triggers the fluctuations of the edge cloud workload. Therefore, an elastic resource management method based on workload forecasting in edge clouds is proposed in this paper. When the resource demand is large, more resources are requested from the cloud service provider so that the task can be completed before the deadline. When the resource demand is small, the idle resource is released to meet the cost constraint. The resource demand is judged based on the workload forecasting. In order to improve the accuracy of workload forecasting, a workload forecasting model based on error correction is proposed in this paper. Neither overload nor the light-load status of edge cloud nodes can make full use of the resources. To improve the node processing performance and reduce migration times, a workload migration model for minimizing migration times is proposed in this paper. The experimental results show that the proposed methods can effectively forecast the workload and improve the processing performance of the entire cluster.

Cost and green aware workload migration on geo-distributed datacentres

With the development of the inter-datacentre (inter-DC) virtual machine migration technology, it is possible to reduce the cost of electricity and the environment by using the workload migration across the datacentre. This paper presents a solution - cost and green aware workload migration algorithm (CGWM) that utilising the difference of electricity prices, CO2 emissions and water consumption between different geographical locations to manage the workload. CGWM attempts to reduce electricity costs, carbon emissions and water consumption. When the three optimisation goals conflict, CGWM first to ensure the reduction of electricity cost, and then by adjusting the weight factor to make CGWM more biased to optimise the carbon dioxide or water consumption. Simulation results show CGWM can reduce electricity costs while controlling carbon dioxide emissions and water consumption.

Cost and green aware workload migration on geo-distributed datacentres

With the development of the inter-datacentre (inter-DC) virtual machine migration technology, it is possible to reduce the cost of electricity and the environment by using the workload migration across the datacentre. This paper presents a solution - cost and green aware workload migration algorithm (CGWM) that utilising the difference of electricity prices, CO2 emissions and water consumption between different geographical locations to manage the workload. CGWM attempts to reduce electricity costs, carbon emissions and water consumption. When the three optimisation goals conflict, CGWM first to ensure the reduction of electricity cost, and then by adjusting the weight factor to make CGWM more biased to optimise the carbon dioxide or water consumption. Simulation results show CGWM can reduce electricity costs while controlling carbon dioxide emissions and water consumption.

Flow-Aware Workload Migration in Data Centers

In data centers, subject to workloads with heterogeneous (and sometimes short) lifetimes, workload migration is a way of attaining a more efficient utilization of the underlying physical machines. To not introduce performance degradation, such workload migration must take into account not only machine resources, and per-task resource requirements, but also application dependencies in terms of network communication. This paper presents a workload migration model capturing all of these constraints. A linear programming framework is developed allowing accurate representation of per-task resources requirements and inter-task network demands. Using this, a multi-objective problem is formulated to compute a re-allocation of tasks that (1) maximizes the total inter-task throughput, while (2) minimizing the cost incurred by migration and (3) allocating the maximum number of new tasks. A baseline algorithm, solving this multi-objective problem using the $$\varepsilon$$ -constraint method is proposed, in order to generate the set of Pareto-optimal solutions. As this algorithm is compute-intensive for large topologies, a heuristic, which computes an approximation of the Pareto front, is then developed, and evaluated on different topologies and with different machine load factors. These evaluations show that the heuristic can provide close-to-optimal solutions, while reducing the solving time by one to two order of magnitudes.

Towards efficiently supporting database as a service with QoS guarantees

Quality of Service (QoS) is at the core of the vision of Database as a Service (DBaaS), which provides guarantees to database users on the usability of database services, even when the underlying database infrastructure is shared by multiple users. When QoS guarantees are necessitated, traditional approaches in DBaaS often have to reserve computation resources (e.g. CPU and memory) for tenants according to their performance Service Level Objectives (SLOs), so that the database engine always possesses sufficient resources to accomplish the expected workloads under any circumstance. Such resource reservation schemes inevitably result in poor resource utilization, as the actual workloads of tenants are usually below their maximal workload expectation described in their SLOs.In this paper, we propose a novel scheme called FrugalDB to further improve resource utilization and thus reduce operational cost for DBaaS systems with QoS guarantees. FrugalDB accommodates two independent database engines, an in-memory engine for heavy workloads with tight SLOs, and a disk-based engine for light workloads with loose SLOs. The in-memory database is leveraged to migrate temporary heavy workloads from the disk-based database, when the latter itself does not suffice to handle all active tenants’ workloads, and thus it could be relieved of the tremendous data serving pressure yielded by those heavy workloads, and focus on processing massive numbers of light workloads. When the heavy workloads fade out, FrugalDB reversely migrates a tenant’s workload from the in-memory database into the disk-based database, so that the occupied memory resources could be recollected for subsequent workload migrations. By an effective workload estimation method and an efficient migration schedule algorithm, FrugalDB tries to minimize workload migration cost incurred in moving workloads between the two engines. By allocating each tenant’s workload to an appropriate engine via workload migration, this dual-engine scheme can substantially save computation resources, and thus consolidate more tenants on a single database server. We evaluate FrugalDB with extensive experiments, which show that it has a higher tenant consolidation rate with performance SLO guarantees and fewer performance SLO violations than the existing systems, as well as with acceptable response latency.

Real-World Experience with a Multicloud Exchange

In the past decade, cloud service provider offerings and customer awareness of how to use them have evolved significantly. One of the major developments has been the evolution of customers from using a single service from a single provider to using multiple clouds in a true, integrated multicloud. Due to industry characteristics such as differential and dynamic pricing, emerging technologies such as containers and standards that facilitate workload migration, and the needs of customers to minimize spend yet maximize performance and availability, cloud brokers such as UCX are becoming an essential element of the entire ecosystem.

HWM: a hybrid workload migration mechanism of metadata server cluster in data center

In data center, applications of big data analytics pose a big challenge to massive storage systems. It is significant to achieve high availability, high performance and high scalability for PB-scale or EB-scale storage systems. Metadata server (MDS) cluster architecture is one of the most effective solutions to meet the requirements of applications in data center. Workload migration can achieve load balance and energy saving of cluster systems. In this paper, a hybrid workload migration mechanism of MDS cluster is proposed and named as HWM. In HWM, workload of MDS is classified into two categories: metadata service and state service, and they can be migrated rapidly from a source MDS to a target MDS in different ways. Firstly, in metadata service migration, all the dirty metadata of one sub file system is flushed to a shared storage pool by the source MDS, and then is loaded by the target MDS. Secondly, in state service migration, all the states of that sub file system are migrated from source MDS to target MDS through network at file granularity, and then all of the related structures of these states are reconstructed in targetMDS. Thirdly, in the process of workload migration, instead of blocking client requests, the source MDS can decide which MDS will respond to each request according to the operation type and the migration stage. The proposed mechanismis implemented in the BlueWhaleMDS cluster. The performance measurements show that the HWM mechanism is efficient to migrate the workload of a MDS cluster system and provides low-latency access to metadata and states.

Toward Beneficial Transformation of Enterprise Workloads to Hybrid Clouds

With the promise of providing flexible and elastic computing resources on demand, the cloud computing has been attracting enterprises and individuals to migrate workloads in the legacy environment to the public/private/hybrid clouds. However, the workload migration is often interpreted as an image migration or re-installation/data copying as the exact snapshot of the source machine, and the various cloud platforms and service models are rarely taken into consideration during migration planning. Thus, the cloud migration techniques have not provided enough options that can satisfy the various migration requirements. In this paper, we propose a model to tackle the migration challenges that transforms one resource into the same or another resource in hybrid clouds. We formulate the problem as a constraint satisfaction problem, and iteratively decompose the server components and consolidate the servers. Furthermore, we propose a compute-network mapping algorithm to match computing resources with network resources to guarantee network affinity. The ultimate goal is to recommend the optimal target cloud platform with network affinity and the minimum cost. Through the evaluation of the proposed model using real enterprise datasets (up to 2012 machines), we prove that the proposed model satisfies the goal. We show that when migrating into virtualized cloud environments, thorough resource planning can reduce 16% of current resources, 5%–10% servers can be consolidated, and more than 60% servers are possible candidates for server decomposition.

Thermal Prediction Models for Virtualized Data center Servers by using Thermal-profiles

The energy dissipated as heat for each utilization level of a data center server is empirically measured and stored as the thermal-profile. These thermal-profiles are used to predict the outlet temperatures of the related servers for current and future utilization. The predicted outlet temperature is an important parameter for energy efficient thermal-aware workload scheduling and workload migration in green data centers. This paper presents three models for outlet temperature prediction on virtualized data center servers based on thermal-profile. The best case scenario managed to predict the outlet temperature with a negligible error of 0.3 degree Celsius.

Towards an Applicability of Current Network Forensics for Cloud Networks: A SWOT Analysis

In recent years, the migration of the computational workload to computational clouds has attracted intruders to target and exploit cloud networks internally and externally. The investigation of such hazardous network attacks in the cloud network requires comprehensive network forensics methods (NFM) to identify the source of the attack. However, cloud computing lacks NFM to identify the network attacks that affect various cloud resources by disseminating through cloud networks. In this paper, the study is motivated by the need to find the applicability of current (C-NFMs) for cloud networks of the cloud computing. The applicability is evaluated based on strengths, weaknesses, opportunities, and threats (SWOT) to outlook the cloud network. To the best of our knowledge, no research to date has been conducted to assist network forensics investigators and cloud service providers in finding an optimal method for investigation of network vulnerabilities found in cloud networks. To this end and in this paper, the state-of-the-art C-NFMs are classified and analyzed based on the cloud network perspective using SWOT analysis. It implies that C-NFMs have a suitable impact on cloud network, which further requires for reformation to ensure its applicability in cloud networks.

ICIF: an inter-cloud interoperability framework for computing resource cloud providers in factories of the future

Factories of the future (FoF) is a relatively new topic focusing on developing competitive manufacturing. Exploiting cloud services is a vital area for FoF. The collaboration between different cloud vendors can provide better quality of service at a lower price. This paper discusses a framework which supports inter-cloud interoperability in a heterogeneous computing resource cloud environment with the goal of dispatching the workload to the most effective clouds available at runtime. Moreover, a genetic algorithm (GA)–based job-scheduler is proposed as a part of the interoperability framework offering workload migration with the best performance at the least cost. The resource selection model is evaluated using agent-based simulation approach.

Research on CPU Workload Prediction and Balancing in Cloud Environment

Servers workload in the cloud environment should be balanced in order to achieve high efficiency and reduce resources consuming. One of the solutions is based on workload prediction, and design a proper load migration and balancing strategy. For the ease of discussion, we focus on CPU workload only in this paper. Specifically, considering the characteristics of workload, such as the strong correlation with time, we employ a time-series based two-step method to predict the CPU workload for both individual physical server and the cluster. Then, with the knowledge of the cluster workload, we design a strategy for workload migration and load balancing. Besides, we conduct extensive experiments to evaluate our method.

Cooling aware job migration for reducing cost in cloud environment

With the growth in computing needs, energy cost includes a large portion of operating cost of cloud data centers. Electricity prices vary in different times and geographical places. Such diversity provides opportunity for diminishing total cost via migration of jobs to places with lower energy prices. Most of the previous studies only focus on computing cost of data centers and disregard other significant parameters such as cooling cost of data centers. These approaches prefer data centers which are located in states with cheaper computing cost. Nonetheless, inappropriate workload migration may lead to a remarkable increase in the total cost because of ignoring the cooling cost of data centers. To address this challenge, we show that minimization of the total cost must cover both the computing and cooling cost while considering delay requirements of jobs. Moreover, we propose an analytical approach which captures the interaction between migration decisions and cooling cost in cloud data centers. Features that make our approach distinct from other similar approaches are the following: first, we consider that cooling cost increases in a nonlinear way with respect to the data center utilization; second, we model cooling cost without any assumption about how the data center cooling system works. In order to achieve energy saving, we determine how much workload should be migrated to other data centers and also the number of servers allocated to each data center for executing the workload. We accomplish migration of workload between data centers by utilizing variety in electricity prices in different locations and times and achieve lower total cost compared with previous schemes. Eventually, using MapReduce traces, we validate our method and indicate that remarkable cost saving, around 37 % can be obtained by cooling-aware job migration.

DS2: A DHT-based substrate for distributed services

DHT (Distributed Hash Table) algorithms are very efficient for distributed data management. As one kind of P2P overlay, DHT overlay also has the advantages of high reliability, high scalability and low cost. Because of these advantages, DHT has been proposed to form server farms such as DHT-based NoSQL databases, SIP server farms, IMS server farms, OpenFlow controller farms, etc. This paper presents DS2, a DHT-based substrate designed for the application server farms providing distributed services. DS2 facilitates the deployment of DHT-based distributed services in three aspects. First, DS2 offers a powerful data model to manage complex data. Second, DS2 provides application message routing function and workload migration function, which help application server farm to achieve load balance, failover and service continuity. Third, DS2 allows deploying application server farms across data centers. DS2 deploying across data centers is optimized to provide better performance and service continuity. We have implemented a DS2 prototype and used it in ZTE’s P2P CSCFprototype and service routing prototype to enable IMS services and service routing services successfully.

Heavy-Traffic Limits for a Many-Server Queueing Network with Switchover

We study a multiclass Markovian queueing network with switchover across a set of many-server stations. New arrivals to each station follow a nonstationary Poisson process. Each job waiting in queue may, after some exponentially distributed patience time, switch over to another station or leave the network following a probabilistic and state-dependent mechanism. We analyze the performance of such networks under the many-server heavy-traffic limiting regimes, including the critically loaded quality-and-efficiency-driven (QED) regime, and the overloaded efficiency-driven (ED) regime. We also study the limits corresponding to mixing the underloaded quality-driven (QD) regime with the QED and ED regimes. We establish fluid and diffusion limits of the queue-length processes in all regimes. The fluid limits are characterized by ordinary differential equations. The diffusion limits are characterized by stochastic differential equations, with a piecewise-linear drift term and a constant (QED) or time-varying (ED) covariance matrix. We investigate the load balancing effect of switchover in the mixed regimes, demonstrating the migration of workload from overloaded stations to underloaded stations and quantifying the load balancing impact of switchover probabilities.

Heavy-Traffic Limits for a Many-Server Queueing Network with Switchover

We study a multiclass Markovian queueing network with switchover across a set of many-server stations. New arrivals to each station follow a nonstationary Poisson process. Each job waiting in queue may, after some exponentially distributed patience time, switch over to another station or leave the network following a probabilistic and state-dependent mechanism. We analyze the performance of such networks under the many-server heavy-traffic limiting regimes, including the critically loaded quality-and-efficiency-driven (QED) regime, and the overloaded efficiency-driven (ED) regime. We also study the limits corresponding to mixing the underloaded quality-driven (QD) regime with the QED and ED regimes. We establish fluid and diffusion limits of the queue-length processes in all regimes. The fluid limits are characterized by ordinary differential equations. The diffusion limits are characterized by stochastic differential equations, with a piecewise-linear drift term and a constant (QED) or time-varying (ED) covariance matrix. We investigate the load balancing effect of switchover in the mixed regimes, demonstrating the migration of workload from overloaded stations to underloaded stations and quantifying the load balancing impact of switchover probabilities.

Virtualization

With virtualization, a resource's consumers are provided with a virtual rather than physical version of that resource. This layer of indirection has helped address myriad problems, including efficiency, security, high availability, elasticity, fault containment, mobility, and scalability. This special issue of IEEE Internet Computing surveys some examples of virtualization enabling technologies such as workload migration and gaming consoles and introduces three articles on the subjects of language-level virtualization exploited to automatically parallelize applications; network virtualization used to enable tenant-specific network customization; and a virtualization layer employed above existing clouds to enable a multicloud grid.

Resource-aware architectures for adaptive particle filter based visual target tracking

There are a growing number of visual tracking applications now being envisioned for mobile devices. However, since computer vision algorithms such as particle filtering have large computational demands, they can result in high energy consumption and temperatures in mobile devices. Conventional approaches for distributed target tracking with a camera node and a receiver node are either sender-based (SB) or receiver-based (RB). The SB approach uses little energy and bandwidth, but requires a sender with large computational resources. The RB approach fits applications where computational resources are completely unavailable to the sender, but requires very large energy and bandwidth. In this article, we propose three architectures for distributed particle filtering that (i) reduce particle filtering workload and (ii) allow for dynamic migration of workload between nodes participating in tracking. We also discuss an adaptive particle filtering extension that adapts particle filter computational complexity and can be applied to both the conventional and proposed architectures for improved energy efficiency. Results show that the proposed solutions require low additional overhead, improve on tracking system lifetime, balance node temperatures, maintain track of the desired target, and are more effective than conventional approaches in many scenarios.

Core frequency adjustment to optimize Time Warp on many-core processors

Time Warp synchronized parallel discrete event simulators are organized to operate asynchronously and aggressively without explicit synchronization between the concurrently executing simulation objects. In place of an explicit synchronization mechanism, the concurrent simulators implement an independent but common virtual clock model and a rollback/recovery mechanism to restore causal order when out-of-order events are detected. When the critical path of execution of the simulation is balanced across this parallel threads of execution, this can result in a highly effective, lightweight synchronization mechanism to implement parallel simulation. However, imbalances in the workload across the threads can result in excessive rollback in some threads and slowed progress of the critical path. On small shared memory multi-core systems, a lowest time-stamp scheduling policy can effectively balance the workload. However, on larger many-core chips, conventional load balancing and workload migration will once again become necessary. Fortunately, emerging many-core chips contain some interesting features that can potentially be exploited to improve the performance of parallel simulations. In particular, the recently developed Intel Single-chip Cloud Computer (SCC) provides mechanisms for the runtime control of the frequency and voltage settings of the chip. Furthermore, the frequency and voltage settings are independently set within different regions (called islands) of the chip. Thus, in a Time Warp simulation, one could increase the frequency of the cores executing threads on the critical path (those experiencing infrequent rollback) and decrease the frequency of the cores executing threads off the critical path (those experiencing excessive rollback). This paper investigates the run-time control and adjustment of core frequency in some contemporary x86 multi-core processors to identify the platforms that can support the exploration of dynamic run-time control of core frequency settings. The results show that while all multi-core processors have software controllable core frequency modulation capabilities, they are generally not fully independent as the system comes under load and are therefore unsuitable for these studies. Fortunately, one processor, the AMD X6 line, provides software control for core frequencies that can be fixed (by software) even as the system operates under load.