Content delivery networks (CDNs) provide efficient content distribution over the Internet. CDNs improve the connectivity and efficiency of global communications, but their caching mechanisms may be breached by cyber-attackers. Among the security mechanisms, effective anomaly detection forms an important part of CDN security enhancement. In this work, we propose a multi-perspective unsupervised learning framework for anomaly detection in CDNs. In the proposed framework, a multi-perspective feature engineering approach, an optimized unsupervised anomaly detection model that utilizes an isolation forest and a Gaussian mixture model, and a multi-perspective validation method, are developed to detect abnormal behaviors in CDNs mainly from the client Internet Protocol (IP) and node perspectives, therefore to identify the denial of service (DoS) and cache pollution attack (CPA) patterns. Experimental results are presented based on the analytics of eight days of real-world CDN log data provided by a major CDN operator. Through experiments, the abnormal contents, compromised nodes, malicious IPs, as well as their corresponding attack types, are identified effectively by the proposed framework and validated by multiple cybersecurity experts. This shows the effectiveness of the proposed method when applied to real-world CDN data.

Vehicle-to-everything (V2X) communication and services have been garnering significant interest from different stakeholders as part of future intelligent transportation systems (ITSs). This is due to the many benefits they offer. However, many of these services have stringent performance requirements, particularly in terms of the delay/latency. Multi-access/mobile edge computing (MEC) has been proposed as a potential solution for such services by bringing them closer to vehicles. Yet, this introduces a new set of challenges such as where to place these V2X services, especially given the limit computation resources available at edge nodes. To that end, this work formulates the problem of optimal V2X service placement (OVSP) in a hybrid core/edge environment as a binary integer linear programming problem. To the best of our knowledge, no previous work considered the V2X service placement problem while taking into consideration the computational resource availability at the nodes. Moreover, a low-complexity greedy-based heuristic algorithm named “Greedy V2X Service Placement Algorithm” (G-VSPA) was developed to solve this problem. Simulation results show that the OVSP model successfully guarantees and maintains the QoS requirements of all the different V2X services. Additionally, it is observed that the proposed G-VSPA algorithm achieves close to optimal performance while having lower complexity.

Different challenges are facing the adoption of cloud-based applications, including high availability (HA), energy, and other performance demands. Therefore, an integrated solution that addresses these issues is critical for cloud services. Cloud providers promise the HA of their infrastructure while cloud tenants are encouraged to deploy their applications across multiple availability zones. Moreover, the environmental and cost impacts of running applications in the cloud are integral parts of incorporated responsibility where the cloud providers and tenants intend to reduce. Hence, an analytical stochastic model is needed for the tenants and providers to quantify the expected availability offered by an application deployment. If multiple deployment options can satisfy the HA requirement, the question remains, how can we choose the deployment that satisfies the other providers and tenants requirements? Therefore, this paper proposes a cloud scoring system and integrates it with a Stochastic Petri Net model. While the Petri Net model evaluates the availability of cloud applications deployments, the scoring system selects the optimal HA-aware deployment in terms of energy, operational expenditure, and other norms. We illustrate our approach with a use case that shows how we can use the various deployment options to satisfy both the cloud tenant and provider needs.

Deploying V2X services has become a challenging task. This is mainly due to the fact that such services have strict latency requirements. To meet these requirements, one potential solution is adopting mobile edge computing (MEC). However, this presents new challenges including how to find a cost efficient placement that meets other requirements such as latency. In this work, the problem of cost-optimal V2X service placement (CO-VSP) in a distributed cloud/edge environment is formulated. Additionally, a cost-focused delay-aware V2X service placement (DA-VSP) heuristic algorithm is proposed. Simulation results show that both CO-VSP model and DA-VSP algorithm guarantee the QoS requirements of all such services and illustrates the trade-off between latency and deployment cost.

With the growing demand for data connectivity, network service providers are faced with the task of reducing their capital and operational expenses while simultaneously improving network performance and addressing the increased connectivity demand. Although Network Function Virtualization (NFV) has been identified as a solution, several challenges must be addressed to ensure its feasibility. In this paper, we address the Virtual Network Function (VNF) placement problem by developing a machine learning decision tree model that learns from the effective placement of the various VNF instances forming a Service Function Chain (SFC). The model takes several performance-related features from the network as an input and selects the placement of the various VNF instances on network servers with the objective of minimizing the delay between dependent VNF instances. The benefits of using machine learning are realized by moving away from a complex mathematical modelling of the system and towards a data-based understanding of the system. Using the Evolved Packet Core (EPC) as a use case, we evaluate our model on different data center networks and compare it to the BACON algorithm in terms of the delay between interconnected components and the total delay across the SFC. Furthermore, a time complexity analysis is performed to show the effectiveness of the model in NFV applications.

In this paper, we present our proof-of-concept of a workload aware storage platform. The POC demonstrates the feasibility of building a machine learning technics facilitated middleware for storage management. The middleware is capable of providing optimal assignments of storage workloads to backends as well as continuously on-the-fly optimization thereafter. Experiment indicates that the proposed middleware can efficiently and dynamically adapt the storage backend to satisfy the SLA requirements with minimum impact on the workloads.

Cloud computing offers the ability to use compute, network, and storage resources on demand in a virtualized environment. By virtualizing the physical infrastructure, the resources can be dimensioned at a finer grain allowing multiple tenants to share the same infrastructure while each uses its own share. Yet the question remains, how can we ensure that the resources we allocate to a given software application are enough to guarantee that it provides its functionality according to its service level agreement (SLA). The SLA can constrain the expected availability as well as the speed of handling requests. In this paper, first we define a model to profile the software application and the resources. Then, based on this model we derive a method to determine the needed resources to satisfy the SLA constraints. Our method is based on the low level analysis of resource utilization during the software component life cycle.

Linux containers provide a low overhead and lightweight virtualization solution that increases significantly deployment density of containers per host. Using containers, different applications are deployed as a set of micro-services. This in turn improves resource utilization and cost efficiency, system agility and continuous development, and also upgrade and runtime modification. Therefore, containers have attracted lots of attention as an alternative to virtual machines hosting the cloud applications. In spite of several advantages of containers versus virtual machines, assuring quality of the provided service is challenging. In this paper, we survey some of the available promising container orchestration solutions with particular attention to their quality of service assurance capability.

High availability (HA) is a main key performance indicator for cloud deployed services. Cloud providers offer different availability zones possibly located in different geographical regions. To protect cloud services against failures and natural disasters, it is recommended to deploy the applications on redundant resources across multiple zones and distribute the workload through a load-balancer. Different cloud infrastructure, located in different geographical zones with different energy source powering, hardware quality, etc., may have different reliability levels. Scheduling a cloud service on different zones while meeting the service level agreement availability requirements necessitate a solution to assess the expected availability of a given deployment. To quantify the expected availability offered by an application deployment, a formal stochastic model is required to capture the stochastic behavior of failures. This paper proposes a stochastic Petri Net model that captures the stochastic characteristics of cloud services and translates them into elements of an availability model. The model evaluates the availability of cloud services and their deployments in geographically distributed data centers (DCs). The results are useful to generate guidelines for an HA-aware scheduling.

Cloud computing is an attractive business model offering cost-efficiency and business agility. Recently, the trend is that small and large businesses are moving their services to cloud environments. The quality of service is always negotiated between the cloud users and the cloud providers and documented in the service level agreement (SLA). Yet assuring -- or even measuring -- the quality of the provided service can be challenging. This paper proposes a formal approach for quantifying the quality of service in the cloud systems as promised in the SLA. The proposed approach uses controller synthesis to find a system configuration that meets the SLA requirement. The formal approach suggested in this paper is based on, but not limited to, %the controller synthesis of Time Petri Nets (TPN). As a case study, we focus on service availability as a key performance indicator in the SLA and for a sample set of resources providing a service, we determine the system configuration satisfying the SLA.

Cloud computing is a popular paradigm for providing computational services. Cloud computing frameworks offer cost efficiency, better resource utilization and scalability. Availability remains as one of the main challenges for the cloud. Solutions, such as Heat, have been proposed and integrated with cloud controllers like Open Stack. These solutions protect services against application and infrastructure failure. However, the recovery for the application is tied to recovery of the underlying infrastructure, therefore, the service recovery and outage time may be substantial. In this paper we propose an architecture which integrates an existing middleware solution for high availability, OpenSAF, with OpenStack, for the management of the availability of the applications and the infrastructure. We show that in comparison with Heat, the proposed solution improves the service recovery and outage time and offers flexibility in terms of redundancy models and recovery actions.