Giovani Estrada

Knowledge-Defined Networking

The research community has considered in the past the application of Artificial Intelligence (AI) techniques to control and operate networks. A notable example is the Knowledge Plane proposed by D.Clark et al. However, such techniques have not been extensively prototyped or deployed in the field yet. In this paper, we explore the reasons for the lack of adoption and posit that the rise of two recent paradigms: Software-Defined Networking (SDN) and Network Analytics (NA), will facilitate the adoption of AI techniques in the context of network operation and control. We describe a new paradigm that accommodates and exploits SDN, NA and AI, and provide use-cases that illustrate its applicability and benefits. We also present simple experimental results that support, for some relevant use-cases, its feasibility. We refer to this new paradigm as Knowledge-Defined Networking (KDN).

Fuzzy Self-Learning Controllers for Elasticity Management in Dynamic Cloud Architectures

Cloud controllers support the operation and quality management of dynamic cloud architectures by automatically scaling the compute resources to meet performance guarantees and minimize resource costs. Existing cloud controllers often resort to scaling strategies that are codified as a set of architecture adaptation rules. However, for a cloud provider, deployed application architectures are black-boxes, making it difficult at design time to define optimal or pre-emptive adaptation rules. Thus, the burden of taking adaptation decisions often is delegated to the cloud application. We propose the dynamic learning of adaptation rules for deployed application architectures in the cloud. We introduce FQL4KE, a self-learning fuzzy controller that learns and modifies fuzzy rules at runtime. The benefit is that we do not have to rely solely on precise design-time knowledge, which may be difficult to acquire. FQL4KE empowers users to configure cloud controllers by simply adjusting weights representing priorities for architecture quality instead of defining complex rules. FQL4KE has been experimentally validated using the cloud application framework ElasticBench in Azure and OpenStack. The experimental results demonstrate that FQL4KE outperforms both a fuzzy controller without learning and the native Azure auto-scaling.

Self-Learning Cloud Controllers: Fuzzy Q-Learning for Knowledge Evolution

Auto-scaling features enable cloud applications to maintain enough resources to satisfy demand spikes, reduce costs and keep performance in check. Most auto-scaling strategies rely on a predefined set of rules to scale up/down the required resources depending on the application usage. Those rules are however difficult to devise and generalize, and users are often left alone tuning auto-scale parameters of essentially blackbox applications. In this paper, we propose a novel fuzzy reinforcement learning controller, FQL4KE, which automatically scales up or down resources to meet performance requirements. The Q-Learning technique, a model-free reinforcement learning strategy, frees users of most tuning parameters. FQL4KE has been successfully applied and we therefore think that a fuzzy controller with Q-Learning is indeed a promising combination for auto-scaling resources.

An Auto-Scaling Cloud Controller Using Fuzzy Q-Learning - Implementation in OpenStack

Auto-scaling, i.e., acquiring and releasing resources automatically, is a central feature of cloud platforms. The key problem is how and when to add/remove resources in order to meet agreed service-level agreements. Many commercial solutions use simple approaches such as threshold-based ones. However, providing good thresholds for auto-scaling is challenging. Recently, machine learning approaches have been used to complement and even replace expert knowledge. We propose a dynamic learning strategy based on a fuzzy logic algorithm, which learns and modifies fuzzy scaling rules at runtime without requiring prior knowledge. The proposed algorithm is implemented and evaluated as an extension to the OpenStack cloud platform, integrating it with the Heat and Ceilometer components for orchestration and monitoring, respectively, using Heat Orchestration Templates. We specifically focus on implementation and experimentation aspects here. Our auto-scaling approach can handle various load traffic situations, delivering resources on demand while reducing infrastructure and management costs. The experimentals show promising performance in terms of resource adjustment to optimize SLA compliance (response time) while reducing cloud provider’s costs.

A Comparison of Reinforcement Learning Techniques for Fuzzy Cloud Auto-Scaling

A goal of cloud service management is to design self-adaptable auto-scaler to react to workload fluctuations and changing the resources assigned. The key problem is how and when to add/remove resources in order to meet agreed service-level agreements. Reducing application cost and guaranteeing service-level agreements (SLAs) are two critical factors of dynamic controller design. In this paper, we compare two dynamic learning strategies based on a fuzzy logic system, which learns and modifies fuzzy scaling rules at runtime. A self-adaptive fuzzy logic controller is combined with two reinforcement learning (RL) approaches: (i) Fuzzy SARSA learning FSL and (ii) Fuzzy Q-learning FQL. As an off-policy approach, Q-learning learns independent of the policy currently followed, whereas SARSA as an on-policy always incorporates the actual agents behavior and leads to faster learning. Both approaches are implemented and compared in their advantages and disadvantages, here in the OpenStack cloud platform. We demonstrate that both auto-scaling approaches can handle various load traffic situations, sudden and periodic, and delivering resources on demand while reducing operating costs and preventing SLA violations. The experimental results demonstrate that FSL and FQL have acceptable performance in terms of adjusted number of virtual machine targeted to optimize SLA compliance and response time.

A Fuzzy Load Balancer for Adaptive Fault Tolerance Management in Cloud Platforms

To achieve high levels of reliability, availability and performance in cloud environments, a fault tolerance approach to handle failures effectively is needed. In most existing research, the primary focus has been on explicit specification-driven solutions which requires too much effort for application developers, and leads to inflexibility. We propose a fuzzy job distributor (load balancer) for fault tolerance management to reduce levels of management complexity for the user. The proposed approach aims to reduce the possibility of fault occurrences in the system by a fair distribution of user job requests among available resources. In our self-adaptive approach, the system manages anomalous situations that might lead to failure by distributing the incoming job request based on the reliability of processing nodes, i.e., virtual machines (VMs). The reliability of VMs is a variable parameter and changes during its lifetime. Our approach is implemented and comparatively analysed using OpenStack. The experimental results show a significant reduction in the occurrence of faults in comparison with other load balancing algorithms.

Anomaly Detection Guidelines for Data Streams in Big Data

Real time data analysis in data streams is a highly challenging area in big data. The surge in big data techniques has recently attracted considerable interest to the detection of significant changes or anomalies in data streams. There is a variety of literature across a number of fields relevant to anomaly detection. The growing number of techniques, from seemingly disconnected areas, prevents a comprehensive review. Many interesting techniques may therefore remain largely unknown to the anomaly detection community at large. The survey presents a compact, but comprehensive overview of diverse strategies for anomaly detection in evolving data streams. A number of recommendations based performance and applicability to use cases are provided. We expect that our classification and recommendations will provide useful guidelines to practitioners in this rapidly evolving field.