Self-∗ through self-learning: Overload control for distributed web systems

Bartolini, Novella; Bongiovanni, G.; Silvestri, Simone

doi:10.1016/j.comnet.2008.11.015

Cited by 20 publications

(7 citation statements)

References 21 publications

(24 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our work on an adaptively self-scaling web server finds points of similarity with several projects that attempt to optimize the QoS of a server within a static allocation of resources, be they physical machines or VMs. Some of the methods used include SnowFlock: Virtual Machine Cloning as a First-Class Cloud Primitive 2:39 VM migration or scheduling [Ranjan et al 2003], workload management and admission control to preserve the QoS through optimal resource use [Zhang et al 2009;Elnikety et al 2004;Bartolini et al 2009], and allowing applications to barter resources [Norris et al 2004]. The Emulab [White et al 2002] testbed predates many of the ideas behind virtual clusters.…”

Section: Related Workmentioning

confidence: 99%

SnowFlock

Lagar-Cavilla

Whitney

Bryant

et al. 2011

ACM Trans. Comput. Syst.

View full text Add to dashboard Cite

A basic building block of cloud computing is virtualization. Virtual machines (VMs) encapsulate a user's computing environment and efficiently isolate it from that of other users. VMs, however, are large entities, and no clear APIs exist yet to provide users with programatic, fine-grained control on short time scales.We present SnowFlock, a paradigm and system for cloud computing that introduces VM cloning as a first-class cloud abstraction. VM cloning exploits the well-understood and effective semantics of UNIX fork. We demonstrate multiple usage models of VM cloning: users can incorporate the primitive in their code, can wrap around existing toolchains via scripting, can encapsulate the API within a parallel programming framework, or can use it to load-balance and self-scale clustered servers.VM cloning needs to be efficient to be usable. It must efficiently transmit VM state in order to avoid cloud I/O bottlenecks. We demonstrate how the semantics of cloning aid us in realizing its efficiency: state is propagated in parallel to multiple VM clones, and is transmitted during runtime, allowing for optimizations that substantially reduce the I/O load. We show detailed microbenchmark results highlighting the efficiency of our optimizations, and macrobenchmark numbers demonstrating the effectiveness of the different usage models of SnowFlock.

show abstract

Section: Related Workmentioning

confidence: 99%

SnowFlock

Lagar-Cavilla

Whitney

Bryant

et al. 2011

ACM Trans. Comput. Syst.

View full text Add to dashboard Cite

show abstract

“…In order to avoid unexpected load changes because of the interactions between irregular parallelism of database applications and scheduling, several admission control techniques [14], [15], [18] are employed. These techniques monitor system statistics regarding CPU, memory, lock contention and tune the amount of work allowed into the system [1].…”

Section: ) Limitationsmentioning

confidence: 99%

No More Backstabbing... A Faithful Scheduling Policy for Multithreaded Programs

Pusukuri

Gupta

Bhuyan

2011

2011 International Conference on Parallel Architectures and Compilation Techniques

View full text Add to dashboard Cite

Efficient contention management is the key to achieving scalable performance for multithreaded applications running on multicore systems. However, contention management policies provided by modern operating systems increase context-switches and lead to performance degradation for multithreaded applications under high loads. Moreover, this problem is exacerbated by the interaction between contention management policies and OS scheduling polices. Time Share (TS) is the default scheduling policy in a modern OS such as OpenSolaris and with TS policy, priorities of threads change very frequently for balancing load and providing fairness in scheduling. Due to the frequent ping-ponging of priorities, threads of an application are often preempted by the threads of the same application. This increases the frequency of involuntary context-switches as wells as lock-holder thread preemptions and leads to poor performance. This problem becomes very serious under high loads.To alleviate this problem, in this paper, we present a scheduling policy called Faithful Scheduling (FF), which dramatically reduces context-switches as well as lock-holder thread preemptions. We implemented FF on a 24-core Dell PowerEdge R905 server running OpenSolaris.2009.06 and evaluated it using 22 programs including the TATP database application, SPECjbb2005, programs from PARSEC, SPEC OMP, and some microbenchmarks. The experimental results show that FF policy achieves high performance for both lightly and heavily loaded systems. Moreover it does not require any changes to the application source code or the OS kernel.

show abstract

“…Moreover, simple over-provisioning cannot handle typical events like flash crowds that often overload Web sites [7]. A number of approaches for overload protection in E-Commerce Web sites have been proposed and developed by researchers.…”

Section: Summary Of Findingsmentioning

confidence: 99%

Admission Control and Request Scheduling for Secured-concurrent-available Architecture

Harini¹,

Padmanabhan²

2013

IJCA

View full text Add to dashboard Cite

The Internet society is continuously growing and the need of performance studies in this field is essential in order to obtain better throughput for the Web users. Moreover, Internet applications and clients have varied service expectations and demand provisioning of different levels of QoS to multiple traffic classes on the Internet. Meeting client QoS expectations prove to be a difficult task for E-Commerce service providers, especially when web servers experience overload conditions that cause increased response time, request rejections, user frustration, lowered usage of the service and reduced revenues. A recently proposed MLF (Multi Layered Filtering) framework manages potential workload; it also guaranties data freshness for all transactions that complete within their deadlines without service differentiation using request filtering and admission control schemes. In this paper, an improved request classifier and class-based admission control scheme called Multi-phase Admission Control (MPAC) that can be integrated with the MLF framework to prioritize user requests into several service classes according to their importance, and offer differentiated quality of service in terms of minimizing the frustration of premium users is presented and evaluated. The performance evaluations carried out confirm that the scheme can significantly boost the reward for having serviced a request while providing better QoS to clients.

show abstract

Self-∗ through self-learning: Overload control for distributed web systems

Cited by 20 publications

References 21 publications

SnowFlock

SnowFlock

No More Backstabbing... A Faithful Scheduling Policy for Multithreaded Programs

Admission Control and Request Scheduling for Secured-concurrent-available Architecture

Contact Info

Product

Resources

About