In this work, we examine the challenges that service providers encounter in managing complex service function graphs, while controlling service delivery latency. Based on the lessons we learn, we outline the design of a new system, Invenio, that empowers providers to effectively place microservices without prior knowledge of service functionality. Invenio correlates user actions with the messages they trigger seen in network traces, and computes procedural affinity for communication among microservices for each user action. The procedural affinity values can then be used to make placement decisions to meet latency constraints of individual user actions. Preliminary experiments with the Clearwater IP Multimedia Subsystem demonstrate that even a single high-latency link can result in significant performance degradation, and placement with Invenio can increase user quality of experience.
{"title":"Procedure-driven deployment support for the microservice era","authors":"Amit Sheoran, S. Fahmy, P. Sharma, Navin Modi","doi":"10.1145/3491084.3491429","DOIUrl":"https://doi.org/10.1145/3491084.3491429","url":null,"abstract":"In this work, we examine the challenges that service providers encounter in managing complex service function graphs, while controlling service delivery latency. Based on the lessons we learn, we outline the design of a new system, Invenio, that empowers providers to effectively place microservices without prior knowledge of service functionality. Invenio correlates user actions with the messages they trigger seen in network traces, and computes procedural affinity for communication among microservices for each user action. The procedural affinity values can then be used to make placement decisions to meet latency constraints of individual user actions. Preliminary experiments with the Clearwater IP Multimedia Subsystem demonstrate that even a single high-latency link can result in significant performance degradation, and placement with Invenio can increase user quality of experience.","PeriodicalId":180607,"journal":{"name":"Proceedings of the 22nd International Middleware Conference: Industrial Track","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123815515","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mengying Zhou, Tiancheng Guo, Yang Chen, Junjie Wan, Xin Wang
CDN is a critical Internet infrastructure that helps Internet users get contents with a short delay. With the development of CDN application scenarios, CDN requests will involve more than one resource type. Unfortunately, the existing CDN server selection schemes targeting one resource type cannot select the most suitable CDN server by considering different resource types together. In this paper, we propose Polygon, a CDN server selection system supporting multiple resource demands. The keystone of Polygon is the deployment of a set of dispatchers at strategic network locations, which can be accessed via anycast. The dispatchers are responsible for resource status collection, server assignment with resource demands, and forwarding requests to suitable CDN servers. Meanwhile, Polygon adopts the 0-RTT and connection migration features of the QUIC protocol to mitigate the extra delay for connection and forwarding. We conduct real-world experiments on the Google Cloud Platform to demonstrate the advantages of Polygon. The results show that the deployment of the dispatchers enables Polygon to provide a better CDN service with a median job completion time reduction of up to 54.8%. Also, Polygon improves resource utilization efficiency by 13% in terms of bandwidth and by 7% in terms of CPU.
{"title":"Polygon: a QUIC-based CDN server selection system supporting multiple resource demands","authors":"Mengying Zhou, Tiancheng Guo, Yang Chen, Junjie Wan, Xin Wang","doi":"10.1145/3491084.3491428","DOIUrl":"https://doi.org/10.1145/3491084.3491428","url":null,"abstract":"CDN is a critical Internet infrastructure that helps Internet users get contents with a short delay. With the development of CDN application scenarios, CDN requests will involve more than one resource type. Unfortunately, the existing CDN server selection schemes targeting one resource type cannot select the most suitable CDN server by considering different resource types together. In this paper, we propose Polygon, a CDN server selection system supporting multiple resource demands. The keystone of Polygon is the deployment of a set of dispatchers at strategic network locations, which can be accessed via anycast. The dispatchers are responsible for resource status collection, server assignment with resource demands, and forwarding requests to suitable CDN servers. Meanwhile, Polygon adopts the 0-RTT and connection migration features of the QUIC protocol to mitigate the extra delay for connection and forwarding. We conduct real-world experiments on the Google Cloud Platform to demonstrate the advantages of Polygon. The results show that the deployment of the dispatchers enables Polygon to provide a better CDN service with a median job completion time reduction of up to 54.8%. Also, Polygon improves resource utilization efficiency by 13% in terms of bandwidth and by 7% in terms of CPU.","PeriodicalId":180607,"journal":{"name":"Proceedings of the 22nd International Middleware Conference: Industrial Track","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130235324","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Proceedings of the 22nd International Middleware Conference: Industrial Track","authors":"","doi":"10.1145/3491084","DOIUrl":"https://doi.org/10.1145/3491084","url":null,"abstract":"","PeriodicalId":180607,"journal":{"name":"Proceedings of the 22nd International Middleware Conference: Industrial Track","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132832643","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Serverless computing is a recent paradigm to program the cloud. It allows to execute user-defined functions at scale, on demand, and in a pay-per-use manner. This paper reports on adapting the Unix shell for serverless. Our software, called the serverless shell (sshell), runs shell scripts on a serverless platform much like with a regular computer. It permits to reuse an existing code base while benefiting from the massive power of serverless and paying only for the resources used. sshell is built around a small set of components that includes a new inter-process communication layer for serverless. We evaluate it in AWS Lambda using several microbenchmarks and a large-scale application. Our results show that sshell achieves comparable or better performance than a high-end server. Moreover, it can be faster and more cost-efficient than a cluster-based solution to mine large datasets.
{"title":"The serverless shell","authors":"Aurèle Mahéo, P. Sutra, Tristan Tarrant","doi":"10.1145/3491084.3491426","DOIUrl":"https://doi.org/10.1145/3491084.3491426","url":null,"abstract":"Serverless computing is a recent paradigm to program the cloud. It allows to execute user-defined functions at scale, on demand, and in a pay-per-use manner. This paper reports on adapting the Unix shell for serverless. Our software, called the serverless shell (sshell), runs shell scripts on a serverless platform much like with a regular computer. It permits to reuse an existing code base while benefiting from the massive power of serverless and paying only for the resources used. sshell is built around a small set of components that includes a new inter-process communication layer for serverless. We evaluate it in AWS Lambda using several microbenchmarks and a large-scale application. Our results show that sshell achieves comparable or better performance than a high-end server. Moreover, it can be faster and more cost-efficient than a cluster-based solution to mine large datasets.","PeriodicalId":180607,"journal":{"name":"Proceedings of the 22nd International Middleware Conference: Industrial Track","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133066043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Authenticated data storage on an untrusted platform is an important computing paradigm for cloud applications ranging from data outsourcing, to cryptocurrency and general transparency logs. These modern applications increasingly feature update-intensive workloads, whereas existing authenticated data structures (ADSs) designed with in-place updates are inefficient to handle such workloads. This work addresses the issue and presents a novel authenticated log-structured merge tree (eLSM) based key-value store built on Intel SGX. We present a system design that runs the code of eLSM store inside enclave. To circumvent the limited enclave memory (128 MB with the latest Intel CPUs), we propose to place the memory buffer of the eLSM store outside the enclave and protect the buffer using a new authenticated data structure by digesting individual LSM-tree levels. We design protocols to support data integrity, (range) query completeness, and freshness. Our protocol causes small proofs by including the Merkle proofs at selected levels. We implement eLSM on top of Google LevelDB and Facebook RocksDB with minimal code change and performance interference. We evaluate the performance of eLSM under the YCSB workload benchmark and show a performance advantage of up to 4.5X speedup.
{"title":"Authenticated key-value stores with hardware enclaves","authors":"Y. Tang, Ju Chen, Kai Li, Jianliang Xu, Qi Zhang","doi":"10.1145/3491084.3491425","DOIUrl":"https://doi.org/10.1145/3491084.3491425","url":null,"abstract":"Authenticated data storage on an untrusted platform is an important computing paradigm for cloud applications ranging from data outsourcing, to cryptocurrency and general transparency logs. These modern applications increasingly feature update-intensive workloads, whereas existing authenticated data structures (ADSs) designed with in-place updates are inefficient to handle such workloads. This work addresses the issue and presents a novel authenticated log-structured merge tree (eLSM) based key-value store built on Intel SGX. We present a system design that runs the code of eLSM store inside enclave. To circumvent the limited enclave memory (128 MB with the latest Intel CPUs), we propose to place the memory buffer of the eLSM store outside the enclave and protect the buffer using a new authenticated data structure by digesting individual LSM-tree levels. We design protocols to support data integrity, (range) query completeness, and freshness. Our protocol causes small proofs by including the Merkle proofs at selected levels. We implement eLSM on top of Google LevelDB and Facebook RocksDB with minimal code change and performance interference. We evaluate the performance of eLSM under the YCSB workload benchmark and show a performance advantage of up to 4.5X speedup.","PeriodicalId":180607,"journal":{"name":"Proceedings of the 22nd International Middleware Conference: Industrial Track","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125728886","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：