Ansar Rafique

Towards Portability and Interoperability Support in Middleware for Hybrid Clouds

The cloud computing paradigm promises increased flexibility and scalability for consumers and providers of software services. Service providers that exploit private cloud environments offer restricted flexibility and scalability because of the limited capacity. However, such organizations are often reluctant to migrate to public clouds because of business continuity threats and vendor lock-in. Hybrid clouds potentially combine the benefits of private and public (external) clouds. Vendor lock-in can be avoided when multiple external clouds are supported and effectively exploited. This paper presents a middleware platform for hybrid cloud applications. The middleware enables organizations to control the execution of their applications in hybrid cloud environments. Driven by policies, the middleware can dynamically decide which requests and tasks are executed on a particular part of the hybrid cloud. The core of the middleware, and the focus of this paper, is an abstraction layer. The abstraction layer enables portability over multiple services including data storage, blob storage, and asynchronous task execution of various PaaS platforms as well as interoperability between the PaaS platforms. We have validated the core concept by building a prototype implementation that runs on top of specific PaaS platforms as well as on a cloud-enabling middleware. A document processing SaaS application has been instantiated on the middleware. Performance results have been collected for JBoss AS cluster, Google App Engine, and Red Hat OpenShift.

Towards a container-based architecture for multi-tenant SaaS applications

SaaS providers continuously aim to optimize the cost-efficiency, scalability and trustworthiness of their offerings. Traditionally, these concerns have been addressed by application-level middleware platforms that implement a multi-tenant architecture. However, the recent uprise and industry adoption of container technology such as Docker and Kubernetes, exactly for the purpose of improving the cost-efficiency, elasticity and resilience of cloud native services, triggers the unanswered question whether and how container technology may affect such multi-tenant architectures. To answer this question, we outline our ideas on a container-based multi-tenant architecture for SaaS applications. Subsequently, we make an assessment of the technical Strengths, Weaknesses, Opportunities, and Threats (SWOT) which should be taken into account by a SaaS provider when considering the adoption of such container-based architecture.

Policy-driven middleware for heterogeneous, hybrid cloud platforms

The cloud computing paradigm promises increased flexibility and scalability. However, in private cloud environments this flexibility and scalability is constrained by the limited capacity. On the other hand, organizations are reluctant to migrate to public clouds because they lose control over their applications and data. The concept of a hybrid cloud tries to combine the benefits of private and public clouds, while also decreasing vendor lock-in. This paper presents PaaSHopper, a middleware platform for hybrid cloud applications that enables organizations to keep fine-grained control over the execution of their applications. Driven by policies, the middleware dynamically decides which requests and tasks are executed in a particular part of the hybrid cloud. We validated this work by means of a prototype on top of a hybrid cloud consisting of a local JBoss AS cluster, Google App Engine, and Red Hat OpenShift.

Object-NoSQL Database Mappers: a benchmark study on the performance overhead

In recent years, the hegemony of traditional relational database management systems (RDBMSs) has declined in favour of non-relational databases (NoSQL). These database technologies are better adapted to meet the requirements of large-scale (web) infrastructures handling Big Data by providing elastic and horizontal scalability. Each NoSQL technology however is suited for specific use cases and data models. As a consequence, NoSQL adopters are faced with tremendous heterogeneity in terms of data models, database capabilities and application programming interfaces (APIs). Opting for a specific NoSQL database poses the immediate problem of vendor or technology lock-in. A solution has been proposed in the shape of Object-NoSQL Database Mappers (ONDMs), which provide a uniform abstraction interface for different NoSQL technologies.Such ONDMs however come at a cost of increased performance overhead, which may have a significant economic impact, especially in large distributed setups involving massive volumes of data.In this paper, we present a benchmark study quantifying and comparing the performance overhead introduced by Object-NoSQL Database Mappers, for create, read, update and search operations. Our benchmarks involve five of the most promising and industry-ready ONDMs: Impetus Kundera, Apache Gora, EclipseLink, DataNucleus and Hibernate OGM, and are executed both on a single node and a 9-node cluster setup.Our main findings are summarised as follows: (i) the introduced overhead is substantial for database operations in-memory, however on-disk operations and high network latency result in a negligible overhead, (ii) we found fundamental mismatches between standardised ONDM APIs and the technical capabilities of the NoSQL database, (iii) search performance overhead increases linearly with the number of results, (iv) DataNucleus and Hibernate OGM’s search overhead is exceptionally high in comparison to the other ONDMs.

PaaSHopper: Policy-driven middleware for multi-PaaS environments

Offering Software-as-a-Service (SaaS) applications on top of a Platform-as-a-Service (PaaS) platform is a promising strategy as the SaaS provider does not need to acquire and maintain private cloud infrastructure, and it enables him/her to enjoy the benefits of cloud scalability and flexiblity as well. However, as this entails losing some control over the application and its data, SaaS providers are in practice reluctant to migrate to a PaaS platform entirely. To alleviate such concerns of vendor lock-in, the concept of a multi-cloud involves integrating and combining multiple cloud environments, private as well as public, but also involving multiple providers and different technologies. This has the added benefit that it further improves overall availability, flexibility and scalability. Current support for multi-cloud applications however is limited.This paper presents PaaSHopper, a middleware platform for developing and operating multi-tenant SaaS applications in a multi-PaaS environment. It enables the SaaS provider to have fine-grained control over the execution of applications and the storage of application data, while offering the tenant some degrees of customization and self-service as well. Driven by stakeholder-specific policies, the middleware dynamically decides which requests and tasks are executed in a particular part of the multi-PaaS environment. We validated this work in the context of four realistic SaaS application cases on top of a multi-cloud consisting of a local JBoss Application Server cluster, Google App Engine, and Red Hat OpenShift.

Towards an Adaptive Middleware for Efficient Multi-Cloud Data Storage

A multi-cloud storage architecture combines different storage technologies and resources from multiple clouds. As it allows application providers to manage the risks associated to technology or vendor lock-in, provider reliability, data security, privacy, it is an increasingly popular tactic for designing the storage tier of cloud-based applications. Multi-cloud storage architectures are however prone to run-time dynamicity: many dynamic properties impact the way such a setup is governed and evolved over time, e.g., storage providers enter or leave the market; pricing policies, QoS metrics and SLA guarantees change over time; etc. This paper provides an in-depth discussion of this complexity, and sketches our architectural vision of self-adaptive middleware solutions that monitor and change the storage architecture (semi-)autonomously. We highlight two areas of ongoing and future research that deserve specific attention: (i) systematically monitoring the storage systems, despite heterogeneity, and (ii) providing uniform methods and techniques to change and control the different storage resources dynamically.

Leveraging NoSQL for Scalable and Dynamic Data Encryption in Multi-tenant SaaS

In the context of multi-tenant SaaS applications, data confidentiality support is increasingly being offered from within the application layer instead of the database layer or the storage layer to accommodate continuously changing requirements of multiple tenants. Application-level data management middleware platforms are becoming increasingly compelling for dealing with the complexity of a multi-cloud or a federated cloud storage architecture as well as multi-tenant SaaS applications.However, these platforms typically support traditional data mapping strategies that are created under the assumption of a fixed and rigorous database schema. Thus, mapping data objects while supporting varying data confidentiality requirements, therefore, leads to fragmentation of data over distributed storage nodes. This introduces significant performance overhead at the level of individual database transactions (e.g., CRUD transactions) and negatively affects the overall scalability.To address these challenges, we present a dedicated data mapping strategy that leverages the data schema flexibility of columnar NoSQL databases to accomplish dynamic and fine-grained data encryption in a more efficient and scalable manner. We validate these solutions in the context of an industrial multi-tenant SaaS application and conduct a comprehensive performance evaluation. The results confirm that the proposed data mapping strategy indeed yields scalability and performance improvements.

Towards scalable and dynamic data encryption for multi-tenant SaaS

Application-level data management middleware solutions are becoming increasingly compelling to deal with the complexity of a multi-cloud or federated cloud storage and multitenant storage architecture. However, these systems typically support traditional data mapping strategies that are created under the assumption of a fixed and rigorous database schema, and mapping data objects while supporting varying data confidentiality requirements therefore leads to fragmentation of data over distributed storage nodes. This introduces performance over-head at the level of individual database transactions and negatively affects the overall scalability. This paper discusses these challenges and highlights the potential of leveraging the data schema flexibility of NoSQL databases to accomplish dynamic and fine-grained data encryption in a more efficient and scalable manner. We illustrate these ideas in the context of an industrial multi-tenant SaaS application.

On the State of NoSQL Benchmarks

The proliferation of Big Data systems and namely NoSQL databases has resulted in a tremendous heterogeneity in its offerings. It has become increasingly difficult to compare and select the most optimal NoSQL storage technology. Current benchmark efforts, such as the Yahoo! Cloud Serving Benchmark (YCSB), evaluate simple read and write operations on a primary key. However, while YCSB has become the de-facto benchmark solution for practitioners and NoSQL vendors, it is lacking in capabilities to extensively evaluate specific NoSQL solutions. In this paper, we present a systematic survey of current NoSQL benchmarks, in which we identify a clear gap in benchmarking more advanced workloads (e.g. nested document search) for features specific to NoSQL database families (e.g. document stores). Secondly, based on our survey, we discuss the strengths and weaknesses of the different benchmark design approaches, and argue in favor of a benchmark suite that targets specific families of NoSQL databases yet still allows overall comparison of databases in terms of their commonalities.

K8-scalar: a workbench to compare autoscalers for container-orchestrated database clusters

Although a considerable amount of research exists on auto-scaling of database clusters, the design of an effective auto-scaling strategy requires fine-grained tailoring towards the specific application scenario. This paper presents an easy-to-use and extensible workbench exemplar, named K8-Scalar (Kube-Scalar), which allows researchers to implement and evaluate different self-adaptive approaches to autoscaling container-orchestrated services. The workbench is based on Docker, a popular technology for easing the deployment of containerized software that also has been positioned as an enabler for reproducible research. The workbench also relies on a container orchestration framework: Kubernetes (K8s), the de-facto industry standard for orchestration and monitoring of elastically scalable container-based services. Finally, it integrates and extends Scalar, a generic testbed for evaluating the scalability of large-scale systems with support for evaluating the performance of autoscalers for database clusters. The paper discusses (i) the architecture and implementation of K8-Scalar and how a particular autoscaler can be plugged in, (ii) sketches the design of a Riemann-based autoscaler for database clusters, (iii) illustrates how to design, setup and analyze a series of experiments to configure and evaluate the performance of this autoscaler for a particular database (i.e., Cassandra) and a particular workload type, and (iv) validates the effectiveness of K8-Scalar as a workbench for accurately comparing the performance of different auto-scaling strategies.