Serverless Solutions Research Articles

Architecting Serverless Solutions for Cost - Effective and Scalable Applications

Architecting Serverless Solutions for Cost - Effective and Scalable Applications

A Seer knows best: Auto-tuned object storage shuffling for serverless analytics

Serverless platforms offer high resource elasticity and pay-as-you-go billing, making them a compelling choice for data analytics. To craft a “pure” serverless solution, the common practice is to transfer intermediate data between serverless functions via serverless object storage (IBM COS; AWS S3). However, prior works have led to inconclusive results about the performance of object storage systems, since they have left large margin for optimization. To verify that object storage has been underrated, we devise a novel shuffle manager for serverless data analytics called Seer. Specifically, Seer dynamically chooses between two shuffle algorithms to maximize performance. The algorithm choice is made online based on some predictive models, and very importantly, without end users having to specify intermediate shuffle data sizes at the time of the job submission. We integrate Seer with PyWren-IBM [31], a well-known serverless analytics framework, and evaluate it against both serverful (e.g., Spark) and serverless systems (e.g., Google BigQuery, Caerus [46] and SONIC [22]). Our results certify that our new shuffle manager can deliver performance improvements over them.

The uphill journey of FaaS in the open-source community

Since its introduction in 2014 by Amazon, the Function as a Service (FaaS) model of serverless computing has set the expectation to fulfill the promise of on-demand, pay-as-you-go, infrastructure-independent processing, originally formulated by cloud computing. Yet, serverless applications are fundamentally different than traditional service-oriented software in that they pose specific performance (e.g., cold start), design (e.g., stateless), and development challenges (e.g., debugging). A growing number of cloud solutions have been continuously attempting to address each of these challenges as a result of the increasing popularity of FaaS. Yet, the characteristics of this model have been poorly understood; therefore, the challenges are poorly tackled. In this paper, we assess the state of FaaS in open-source community with a study on almost 2K real-world serverless applications. Our results show a jeopardized ecosystem, where, despite the hype of serverless solutions in the last years, a number of challenges remain untackled, especially concerning component reuse, support for software development, and flexibility among different platforms — resulting in arguably slow adoption of the FaaS model. We believe that addressing the issues discussed in this paper may help researchers shaping the next generation of cloud computing models.

Serverless computing in omics data analysis and integration.

A comprehensive analysis of omics data can require vast computational resources and access to varied data sources that must be integrated into complex, multi-step analysis pipelines. Execution of many such analyses can be accelerated by applying the cloud computing paradigm, which provides scalable resources for storing data of different types and parallelizing data analysis computations. Moreover, these resources can be reused for different multi-omics analysis scenarios. Traditionally, developers are required to manage a cloud platform’s underlying infrastructure, configuration, maintenance and capacity planning. The serverless computing paradigm simplifies these operations by automatically allocating and maintaining both servers and virtual machines, as required for analysis tasks. This paradigm offers highly parallel execution and high scalability without manual management of the underlying infrastructure, freeing developers to focus on operational logic. This paper reviews serverless solutions in bioinformatics and evaluates their usage in omics data analysis and integration. We start by reviewing the application of the cloud computing model to a multi-omics data analysis and exposing some shortcomings of the early approaches. We then introduce the serverless computing paradigm and show its applicability for performing an integrative analysis of multiple omics data sources in the context of the COVID-19 pandemic.

Serverless Computing (Dagstuhl Seminar 21201)

In the backbone of our digital society, cloud computing enables an efficient, utility-like ecosystem of developing, composing, and providing software services. Responding to a trend to make cloud computing services more accessible, fine-grained, and affordable, serverless computing has gained rapid adoption in practice, and garnered much interest from both industry and academia. However successful, serverless computing manifests today the opportunities and challenges of emerging technology: a rapidly growing field but scattered vision, plenty of new technologies but no coherent approach to design solutions from them, many simple applications but no impressive advanced solution, the emergence of a cloud continuum (resources from datacenters to the edge) but no clear path to leverage it efficiently, and overall much need but also much technical complexity. Several related but disjoint fields, notably software and systems engineering, parallel and distributed systems, and system and performance analysis and modeling, aim to address these opportunities and challenges. Excellent collaboration between these fields in the next decade will be critical in establishing serverless computing as a viable technology. We organized this Dagstuhl seminar to bring together researchers, developers, and practitioners across disciplines in serverless computing, to develop a vision and detailed answers to the timely and relevant, open challenges related to the following topics: - Topic 1: design decisions for serverless systems, platforms, and ecosystems, - Topic 2: software engineering of serverless applications, but also systems, platforms, and ecosystems - Topic 3: applications and domain requirements for serverless computing, - Topic 4: evaluation of serverless solutions, and beyond (privacy, cyber-physical systems, etc.). In this document, we report on the outcomes of Dagstuhl Seminar 21201 Serverless Computing by integrating diverse views and synthesizing a shared vision for the next decade of serverless computing.

Serverless Computing-Where Are We Now, and Where Are We Heading?

Serverless systems are becoming more popular and being adopted by a large number of companies. Serverless computing is a new paradigm that provides a platform to efficiently develop and deploy applications to the market without having to manage any underlying infrastructure.1,2 However, serverless solutions introduce several new issues that companies need to tackle. In this article, we introduce serverless technology and its evolution, discuss its benefits and issues, and describe future challenges that researchers should address.

Toward Distributed Computing Environments with Serverless Solutions in Edge Systems

Computation offloading through stateless applications is gaining momentum thanks to the emergence of serverless frameworks with inherent scalability properties. However, adoption of a serverless framework in an edge computing system requires careful consideration to keep its advantages unscathed. In the cloud, micro-services are scaled automatically according to demands, but in edge computing this would incur a significantly higher cost than in a data center and cannot be as fluid. This is especially relevant in scenarios where edge nodes are spread across large areas and have relatively small computation capabilities. In this article we propose to overcome this issue by adapting the allocation of demands to the currently allocated micro-services at short timescales, with two alternative mechanisms designed for different target scenarios, both aimed at enabling distributed computing environments. The proposed solution can be integrated within the ETSI MEC standard, which specifies a reference architecture and open service interfaces. Our contribution is validated in a proof-of-concept scenario with a prototype implementation released as open source.

ОРГАНІЗАЦІЯ БЕЗСЕРВЕРНИХ ОБЧИСЛЕНЬ ДЛЯ РІЗНОПЛАТФОРМНИХ КЛІЄНТІВ

Зростання інтересу до безсерверних систем обчислень призвело до стрімкого розвитку можливостей, які надають хмарні провайдери. Але специфічність цієї сфери розробки програмного забезпечення поки що створює деякі проблеми, серед яких: • слідування вимогам архітектури серверного додатку не дозволяє розширювати sdk таким чином, щоб новий функціонал працював автоматично; • відсутність API для ефективної синхронізації пристроїв; • відсутність глобального тригеру для збереження історії змін об'єктів системи.Метою роботи є вирішення проблем з можливістю розширення існуючого sdk BaaS провайдеру Parse Server автоматичного запуску нового функціоналу.В ході роботи було використано системний та аналітичний методи дослідження. У роботі запропоновані алгоритми реалізації синхронізації даних різноплатформних пристроїв та принципи збереження знімків (коммітів) змін об'єктів в системах, у яких в ролі серверу використовується безсерверна система обчислення, реалізована на базі моделі BaaS. Для вирішення задачі були використані принципи прототипного об'єктно-орієнтованого принципу програмування, а також патерни проектування: декоратор, стратегія, event-sourcing, builder, factory. За BaaS провайдер було обрано Parse Server. Розроблені та програмно реалізовані наступні алгоритми: ChangeLogSpy – збереження знімків об'єктів системи без додавання додаткової логіки для класу об'єкта; SyncProvider - реалізація ефективної синхронізації сесій різноплатформних клієнтів; GetAllResultsForQuery - реалізація алгоритму асинхронного отримання всіх результатів запиту.

A Model Predictive Controller for Managing QoS Enforcements and Microarchitecture-Level Interferences in a Lambda Platform

Lambda paradigm, also known as Function as a Service (FaaS), is a novel event-driven concept that allows companies to build scalable and reliable enterprise applications in an off-premise computing data-center as a serverless solution. In practice, however, an important goal for the service provider of a Lambda platform is to devise an efficient way to consolidate multiple Lambda functions in a single host. While the majority of existing resource management solutions use only operating-system level metrics (e.g., average utilization of computing and I/O resources) to allocate the available resources among the submitted workloads in a balanced way, a resource allocation schema that is oblivious to the issue of shared-resource contention can result in a significant performance variability and degradation within the entire platform. This paper proposes a predictive controller scheme that dynamically allocates resources in a Lambda platform. This scheme uses a prediction tool to estimate the future rate of every event stream and takes into account the quality of service enforcements requested by the owner of each Lambda function. This is formulated as an optimization problem where a set of cost functions are introduced (i) to reduce the total QoS violation incidents; (ii) to keep the CPU utilization level within an accepted range; and (iii) to avoid the fierce contention among collocated applications for obtaining shared resources. Performance evaluation is carried out by comparing the proposed solution with an enhanced interference-aware version of three well-known heuristics, namely spread , binpack (the two native clustering solutions employed by Docker Swarm ) and best-effort resource allocation schema. Experimental results show that the proposed controller improves the overall performance (in terms of reducing the end-to-end response time) by 14.9 percent on average compared to the best result of the other heuristics. The proposed solution also increases the overall CPU utilization by 18 percent on average (for lightweight workloads), while achieves an average 87 percent (maximum 146 percent) improvement in preventing QoS violation incidents.

Developing a File System Structure to Solve Healthy Big Data Storage and Archiving Problems Using a Distributed File System

Recently, the use of internet has become widespread, increasing the use of mobile phones, tablets, computers, Internet of Things (IoT) devices and other digital sources. In the health sector with the help of new generation digital medical equipment, this digital world also has tended to grow in an unpredictable way in that it has nearly 10% of the global wide data itself and continues to keep grow beyond what the other sectors have. This progress has greatly enlarged the amount of produced data which cannot be resolved with conventional methods. In this work, an efficient model for the storage of medical images using a distributed file system structure has been developed. With this work, a robust, available, scalable, and serverless solution structure has been produced, especially for storing large amounts of data in the medical field. Furthermore, the security level of the system is extreme by use of static Internet protocol (IP), user credentials, and synchronously encrypted file contents. One of the most important key features of the system is high performance and easy scalability. In this way, the system can work with fewer hardware elements and be more robust than others that use name node architecture. According to the test results, it is seen that the performance of the designed system is better than 97% from a Not Only Structured Query Language (NoSQL) system, 80% from a relational database management system (RDBMS), and 74% from an operating system (OS).

On the performance of ACO-based methods in p2p resource discovery

Over the recent years peer-to-peer (p2p) systems have become increasingly popular. As of today most of the internet IP traffic is already transmitted in this format and still it is said to double in volume till 2014. Most p2p systems, however, are not pure serverless solutions, nor is the searching in those networks highly efficient, usually achieved by simple flooding. In order to confront with the growing traffic we must consider more elaborate search mechanisms and far less centralized environments. An effective proposal to this problem is to solve it in the domain of ant colony optimization metaheuristics. In this paper we present an overview of ACO algorithms that offer the best potential in this field, under the strict requirements and limitations of a pure p2p network. We design several experiments to serve as an evaluation platform for the mentioned algorithms to conclude the features of a high quality approach. Finally, we consider two hybrid extensions to the classical algorithms, in order to examine their contribution to the overall quality robustness.