Serverless Computing: One Step Forward, Two Steps Back

Hellerstein, Joseph M.; Faleiro, Jose M.; Sreekanti, Vikram; Tumanov, Alexey; Wu, Chenggang

doi:10.48550/arxiv.1812.03651

Cited by 44 publications

(58 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Functions are also usually not network-addressable and thus two functions could only communicate through slow and expensive storage. Current FaaS platforms lack access to any specialized hardware and functions are only provisioned with the processing power of a timeslice of a CPU hyperthread [Hellerstein et al 2018].…”

Section: Webmentioning

confidence: 99%

“…For Google Cloud Functions, the maximum memory that a function can use is 4096 MB, while the function duration is capped at 540 seconds [Functions 2021b]. Imposing limitations as discussed here, have detrimental effects for some application domains such as long running, compute intensive, data-driven applications [Hellerstein et al 2018].…”

Section: Runtime Resource Limitationsmentioning

confidence: 99%

“…Preliminary surveys on serverless computing identify opportunities of this new computing model, inherent challenges with the serverless concept and ideas for overcoming the said drawbacks [Shafiei et al 2019], [Jonas et al 2019] [Fox et al 2017. Hellerstein et al [Hellerstein et al 2018] pinpoint specific challenges of the serverless model with regard to applicability for data-driven distributed computing. Garcia et al [García-López et al 2019] analyse and discuss trade-offs of today's serverless platforms required for effective processing of big data analytics applications.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Holistic View on Resource Management in Serverless Computing Environments: Taxonomy and Future Directions

Mampage¹,

Karunasekera²,

Buyya³

2021

Preprint

View full text Add to dashboard Cite

Serverless computing has emerged as an attractive deployment option for cloud applications in recent times. The unique features of this computing model include, rapid auto-scaling, strong isolation, fine-grained billing options and access to a massive service ecosystem which autonomously handles resource management decisions. This model is increasingly being explored for deployments in geographically distributed edge and fog computing networks as well, due to these characteristics. Effective management of computing resources has always gained a lot of attention among researchers. The need to automate the entire process of resource provisioning, allocation, scheduling, monitoring and scaling, has resulted in the need for specialized focus on resource management under the serverless model. In this article, we identify the major aspects covering the broader concept of resource management in serverless environments and propose a taxonomy of elements which influence these aspects, encompassing characteristics of system design, workload attributes and stakeholder expectations. We take a holistic view on serverless environments deployed across edge, fog and cloud computing networks. We also analyse existing works discussing aspects of serverless resource management using this taxonomy. This article further identifies gaps in literature and highlights future research directions for improving capabilities of this computing model. CCS Concepts: • General and reference → Surveys and overviews; • Computer systems organization → Distributed architectures.

show abstract

Section: Webmentioning

confidence: 99%

Section: Runtime Resource Limitationsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Holistic View on Resource Management in Serverless Computing Environments: Taxonomy and Future Directions

Mampage¹,

Karunasekera²,

Buyya³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Serverless functions are stateless and its frameworks only deal with the runtime management of functions, completely separating it from the data management [11]. This separation simplifies serverless computing but has drawbacks for data-intensive and stream processing pipelines [12] and that can lead to issue of dealing intermediate data between the functions in pipeline.…”

Section: Introductionmentioning

confidence: 99%

Serverless data pipeline approaches for IoT data in fog and cloud computing

Poojara¹,

Dehury²,

Jakovits³

et al. 2021

Preprint

View full text Add to dashboard Cite

With the increasing number of Internet of Things (IoT) devices, massive amounts of raw data is being generated. The latency, cost, and other challenges in cloud-based IoT data processing have driven the adoption of Edge and Fog computing models, where some data processing tasks are moved closer to data sources. Properly dealing with the flow of such data requires building data pipelines, to control the complete life cycle of data streams from data acquisition at the data source, edge and fog processing, to Cloud side storage and analytics. Data analytics tasks need to be executed dynamically at different distances from the data sources and often on very heterogeneous hardware devices. This can be streamlined by the use of a Serverless (or FaaS) cloud computing model, where tasks are defined as virtual functions, which can be migrated from edge to cloud (and vice versa) and executed in an eventdriven manner on data streams. In this work, we investigate the benefits of building Serverless data pipelines (SDP) for IoT data analytics and evaluate three different approaches for designing SDPs: 1) Off-the-shelf data flow tool (DFT) based, 2) Object storage service (OSS) based and 3) MQTT based. Further, we applied these strategies on three fog applications (Aeneas, PocketSphinx, and custom Video processing application) and evaluated the performance by comparing their processing time (computation time, network communication and disk access time), and resource utilization. Results show that DFT is unsuitable for compute-intensive applications such as video or image processing, whereas OSS is best suitable for this task. However, DFT is nicely fit for bandwidthintensive applications due to the minimum use of network resources. On the other hand, MQTT-based SDP is observed with increase in CPU and Memory usage as the number of users rose, and experienced a drop in data units in the pipeline for PocketSphinx and custom video processing applications, however it performed well for Aeneas which had low size data units.

show abstract

“…The serverless computing model provides a great opportunity for edge computing, which consists of an event driven architecture and is subject to varying workloads [31] [25]. Processing data on the edge of the network provides a solution for limited network bandwidth and latency problems as well as data privacy concerns by executing code closer to devices and end users [12]. This compute model can provide support for latency-sensitive workloads [10] and help adhere to compliance and privacy regulations.…”

Section: Introductionmentioning

confidence: 99%

Serverless Platforms on the Edge: A Performance Analysis

Javed¹,

Toosi²,

Aslanpour³

2021

Preprint

View full text Add to dashboard Cite

The exponential growth of Internet of Things (IoT) has given rise to a new wave of edge computing due to the need to process data on the edge, closer to where it is being produced and attempting to move away from a cloud-centric architecture. This provides its own opportunity to decrease latency and address data privacy concerns along with the ability to reduce public cloud costs. The serverless computing model provides a potential solution with its event-driven architecture to reduce the need for ever-running servers and convert the backend services to an as-used model. This model is an attractive prospect in edge computing environments with varying workloads and limited resources. Furthermore, its setup on the edge of the network promises reduced latency to the edge devices communicating with it and eliminates the need to manage the underlying infrastructure. In this book chapter, first, we introduce the novel concept of serverless edge computing, then, we analyze the performance of multiple serverless platforms, namely, OpenFaaS, AWS Greengrass, Apache OpenWhisk, when set up on the single-board computers (SBCs) on the edge and compare it with public cloud serverless offerings, namely, AWS Lambda and Azure Functions, to deduce the suitability of serverless architectures on the network edge. These serverless platforms are set up on a cluster of Raspberry Pis and we evaluate their performance by simulating different types of edge workloads. The evaluation results show that OpenFaaS achieves the lowest response time on the SBC edge computing infrastructure while serverless cloud offerings are the most reliable with the highest success rate.

show abstract

Serverless Computing: One Step Forward, Two Steps Back

Cited by 44 publications

References 17 publications

A Holistic View on Resource Management in Serverless Computing Environments: Taxonomy and Future Directions

A Holistic View on Resource Management in Serverless Computing Environments: Taxonomy and Future Directions

Serverless data pipeline approaches for IoT data in fog and cloud computing

Serverless Platforms on the Edge: A Performance Analysis

Contact Info

Product

Resources

About