An In-Depth Study of Microservice Call Graph and Runtime Performance

Luo, Shutian; Xu, Huanle; Lu, Chengzhi; Ye, Kejiang; Xu, Guoyao; Zhang, Liping; He, Junjia; Xu, Cheng‐Zhong

doi:10.1109/tpds.2022.3174631

Cited by 37 publications

(11 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In summary, there is a relatively large jump between the scaling groups' scaling factors in a setting that is optimized for scaling cost. The authors of [19], [20] found that a small percentage of microservices are hot-spots in call graphs, specifically, about 5% of microservices are multiplexed by more than 90% of online services in Alibaba clusters, which creates such large differences between scaling factor values, ideal for marking the borders of scaling groups. For an analytically tractable model, in the next statement the modules are assumed to be infinitesimally small, and the scaling factor is interpreted as a differentiable continuous function over the variable that depicts the cumulative resource demand of the modules sorted in the increasing order of their scaling factors.…”

Section: Proposed Method: the Cost Model Of Running A Polylithmentioning

confidence: 99%

“…In order to translate the presented cost values into a realworld example, let us take the microservice statistics of eBay, reported in [25], and that of Alibaba [19], [20]. The microservice system in eBay [25] includes around 3000 services (called as modules in this paper); these services work together to serve more than 10 business domains and form thousands of call paths.…”

Section: Large Scale Example Based On Ebay and Alibaba Tracesmentioning

confidence: 99%

“…The microservice system processes around 26 billion requests per day, i.e., 300 thousand requests per second. Luo et al [19], [20] put more emphasis on the interconnection of the separate microservices, and presented an in-depth study of call graphs within the large-scale deployments of microservices at Alibaba clusters. Their main findings are that i) the size of microservice call graphs follows a heavy-tail distribution: around 10% of call graphs consist of more than 40 unique microservices (they have found that the largest call graph can even consist of 1500+ microservices), ii) a small percentage of microservices are hot-spots in call graphs, specifically, about 5% of microservices are multiplexed by more than 90% of online services and handle 95% of total invocations in Alibaba traces.…”

Section: Large Scale Example Based On Ebay and Alibaba Tracesmentioning

confidence: 99%

“…Based on the reported numbers, let us take a considerably large application, consisting of 100 microservices (the 95th percentile in the cumulative distribution of the number of microservices in each call graph reported in [20]). Furthermore, let us assume an exponential skewness in the invocation rate by combining the findings in [25] and [20]: a 400-fold difference between the most and least frequently called microservice, averaging to 100 calls per second. For simplicity, we assume a balanced memory footprint over the modules, each microservice instance taking 256MB memory.…”

Section: Large Scale Example Based On Ebay and Alibaba Tracesmentioning

confidence: 99%

“…Contrarily, the rise of communication costs is observable as we move from a monolith towards a fully polylithic scenario in which software modules are all separately placed in the cloud, potentially resulting in additional delays during service invocations. As [20] reports, end-to-end latency of online services increases linearly in the length of critical path, which is usually proportional to the number of microservices of the cloud-native application. The reason behind this phenomenon is that invocation between a pair of microservices is usually performed via HTTP REST API, RPC calls or Message Queues, and this can lead to a large communication overhead when many instances of these dependent microservices are located far away from each other.…”

Section: Large Scale Example Based On Ebay and Alibaba Tracesmentioning

confidence: 99%

See 4 more Smart Citations

The Shape of Your Cloud: How to Design and Run Polylithic Cloud Applications

Toka

2022

IEEE Access

View full text Add to dashboard Cite

Nowadays the major trend in IT dictates deploying applications in the cloud, cutting the monolithic software into small, easily manageable and developable components, and running them in a microservice scheme. With these choices come the questions: which cloud service types to choose from the several available options, and how to distribute the monolith in order to best resonate with the selected cloud features. We propose a model that presents monolithic applications in a novel way and focuses on key properties that are crucial in the development of cloud-native applications. The model focuses on the organization of scaling units, and it accounts for the cost of provisioned resources in scale-out periods and invocation delays among the application components. We analyze dis-aggregated monolithic applications that are deployed in the cloud, offering both Container-as-a-Service (CaaS) and Function-as-a-Service (FaaS) platforms. We showcase the efficiency of our proposed optimization solution by presenting the reduction in operation costs as an illustrative example. We propose to group similarly low scale components together in CaaS, while running dynamically scaled components in FaaS. By doing so, the price is decreased as unnecessary memory provisioning is eliminated, while application response time does not show any degradation.

show abstract

Section: Proposed Method: the Cost Model Of Running A Polylithmentioning

confidence: 99%

Section: Large Scale Example Based On Ebay and Alibaba Tracesmentioning

confidence: 99%

Section: Large Scale Example Based On Ebay and Alibaba Tracesmentioning

confidence: 99%

Section: Large Scale Example Based On Ebay and Alibaba Tracesmentioning

confidence: 99%

Section: Large Scale Example Based On Ebay and Alibaba Tracesmentioning

confidence: 99%

See 3 more Smart Citations

The Shape of Your Cloud: How to Design and Run Polylithic Cloud Applications

Toka

2022

IEEE Access

View full text Add to dashboard Cite

show abstract

Practice of Alibaba cloud on elastic resource provisioning for large‐scale microservices cluster

Xu,

Yang,

Wang

et al. 2023

Softw Pract Exp

Self Cite

View full text Add to dashboard Cite

SummaryCloud‐native architecture is becoming increasingly crucial for today's cloud computing environments due to the need for speed and flexibility in developing applications. It utilizes microservice technology to break down traditional monolithic applications into light‐weight and self‐contained microservice components. However, as microservices grow in scale and have dynamic inter‐dependencies, they also pose new challenges in resource provisioning that cannot be fully addressed by traditional resource scheduling approaches. The various microservices with different resource demands and latency requirements can create complex calling chains, making it difficult to provide fine‐grained and accurate resource allocation to each component while maintaining the overall quality of service in the chain. Alibaba Cloud has fully embraced cloud‐native and microservice technologies to drive its key business and scenarios, including Double 11 Shopping Festival. In this work, we aim to address the research problem on how to efficiently provision resources for the growing scale of microservice platform and ensure the performance of latency‐critical microservices. To address the problem, we present in‐depth analyses of Alibaba's microservice cluster and propose optimized resource provisioning algorithms to enhance resource utilization while ensuring the latency requirement. First, we analyze the distinct features of microservices in Alibaba's cluster compared to traditional applications. Then we present Alibaba's resource capacity provisioning workflow and framework to address challenges in resource provisioning for large‐scale and latency‐critical microservice clusters. Finally, we propose enhanced resource provisioning algorithms over Alibaba's current practice by making both proactive and reactive scheduling decisions based on different workloads patterns, which can improve resource usage by 10%–15% in Alibaba's clusters, while maintaining the necessary latency for microservices.

show abstract

ChainsFormer: A Chain Latency-Aware Resource Provisioning Approach for Microservices Cluster

Song,

Xu,

et al. 2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

An In-Depth Study of Microservice Call Graph and Runtime Performance

Cited by 37 publications

References 29 publications

The Shape of Your Cloud: How to Design and Run Polylithic Cloud Applications

The Shape of Your Cloud: How to Design and Run Polylithic Cloud Applications

Practice of Alibaba cloud on elastic resource provisioning for large‐scale microservices cluster

ChainsFormer: A Chain Latency-Aware Resource Provisioning Approach for Microservices Cluster

Contact Info

Product

Resources

About