Towards General-Purpose Resource Management in Shared Cloud Services pdfauthor=Jonathan Mace, Peter Bodik, Madan Musuvathi, Rodrigo Fonseca
نویسندگان
چکیده
In distributed services shared by multiple tenants, managing resource allocation is an important pre-requisite to providing dependability and quality of service guarantees. Many systems deployed today experience contention, slowdown, and even system outages due to aggressive tenants and a lack of resource management. Improperly throttled background tasks, such as data replication, can overwhelm a system; conversely, high-priority background tasks, such as heartbeats, can be subject to resource starvation. In this paper, we outline ve design principles necessary for eective and ecient resource management policies that could provide guaranteed performance, fairness, or isolation.We present Retro, a resource instrumentation framework that is guided by these principles. Retro instruments all system resources and exposes detailed, real-time statistics of pertenant resource consumption, and could serve as a base for the implementation of such policies.
منابع مشابه
Retro: Targeted Resource Management in Multi-tenant Distributed Systems pdfauthor=Jonathan Mace, Peter Bodik, Rodrigo Fonseca, Madanlal Musuvathi
In distributed systems shared by multiple tenants, effective resource management is an important pre-requisite to providing quality of service guarantees. Many systems deployed today lack performance isolation and experience contention, slowdown, and even outages caused by aggressive workloads or by improperly throttled maintenance tasks such as data replication. In this work we present Retro, ...
متن کاملRetro: Targeted Resource Management in Multi-tenant Distributed Systems
In distributed systems shared by multiple tenants, effective resource management is an important pre-requisite to providing quality of service guarantees. Many systems deployed today lack performance isolation and experience contention, slowdown, and even outages caused by aggressive workloads or by improperly throttled maintenance tasks such as data replication. In this work we present Retro, ...
متن کاملPivot Tracing: Dynamic Causal Monitoring for Distributed Systems pdfauthor=Jonathan Mace, Ryan Roelke, Rodrigo Fonseca
Monitoring and troubleshooting distributed systems is notoriously diõcult; potential problems are complex, varied, and unpredictable. _emonitoring and diagnosis tools commonly used today – logs, counters, andmetrics – have two important limitations: what gets recorded is deûned a priori, and the information is recorded in a componentor machine-centric way,making it extremely hard to correlate e...
متن کاملA Model based on Cloud Computing for the implementation and management IT services in Banks
In recent years, the banking industry has made significant changes in technology and communications. The expansion of electronic communications and a large number of people around the world access to the Internet, appropriate to establish trade and economic exchanges provided but high costs, lack of flexibility and agility in existing systems because of the large volume of information, confiden...
متن کاملA Genetic Based Resource Management Algorithm Considering Energy Efficiency in Cloud Computing Systems
Cloud computing is a result of the continuing progress made in the areas of hardware, technologies related to the Internet, distributed computing and automated management. The Increasing demand has led to an increase in services resulting in the establishment of large-scale computing and data centers, in addition to high operating costs and huge amounts of electrical power consumption. Insuffic...
متن کامل