Provisioning Differentiated Last-Level Cache Allocations to VMs in Public Clouds

Due to the limitation of power and processor complexity on traditional single core processors, multi-core processors have become the mainstream. One key feature on commodity multi-cores is that the last level cache (LLC) is usually shared. However, the shared cache contention can affect the performance of applications significantly. Several existing proposals demonstrate that task co-scheduling has the potential to alleviate the contention, but it is challenging to make co-scheduling practical in commodity operating systems. In this paper, we propose two lightweight practical cache-aware co-scheduling methods, namely static SOA and dynamic SOA, to solve the cache contention problem on commodity multi-cores. The central idea of the two methods is that the cache contention can be reduced by adjusting the scheduling order properly. These two methods are different from each other mainly in the way of acquiring the process's cache requirement. The static SOA (static scheduling order adjustment) method acquires the cache requirement information statically by offline profiling, while the dynamic SOA (dynamic scheduling order adjustment) captures the cache requirement statistics by using performance counters. Experimental results using multi-programmed NAS workloads suggest that the proposed methods can greatly reduce the effect of cache contention on multi-core systems. Specifically, for the static SOA method, the execution time can be reduced by up to 15.7%, the number of cache misses can be reduced by up to 11.8%, and the performance improvement remains obvious across the cache size and the length of time slice. For the dynamic SOA method, the execution time reduction can achieve up to 7.09%.

Ideally, we could solve all memory performance problems by storing everything in high performance caches. However, the associated high-cost forces us to keep only a limited amount of working sets of data in these caches and the replacement of disposable blocks of data with the desired one is a fundamental property to caches. Ultimately, the performance of a cache is heavily influenced by the replacement policy it uses. In this paper, we conduct a comparative analysis of a few replacement policies currently in use from traditional strategies to the ones that attempt to emulate optimal replacement and simulate a few replacement strategies to test the performance of multilevel cache by configuring the cache memory subsystem in the Multi2sim simulator. We have focused on the shared nature of the last Level Cache (LLC) which includes many coherence problems. This is where directories and coherence protocols come into play. However, the blocks suffer when the directory entries are themselves evicted from the directory. In this paper, we look into a cache coherence mechanism that instead of evicting the block from the private caches when the directory entry is evicted, rather than storing the entry in the LLC slice. This guarantees freedom from the Directory Eviction Victim Blocks (DEVB). We analyze the performance of our proposed replacement technique, namely, coalesce private or spill shared (CPOSS) by varying the size and associativity of the sparse directory by providing the workload from Parsec, Splash2 and FFTW benchmarks and we observe that the CPOSS outperforms the existing least recently used (LRU) in slice LLC banks in Multicore processors.

Inclusive cache hierarchies are widely adopted in modern processors, since they can simplify the implementation of cache coherence. However, it sacrifices some performance to guarantee inclusion. Many recent intelligent management policies are proposed to improve the last-level cache (LLC) performance by evicting blocks with poor locality earlier. Unfortunately, they are inapplicable in inclusive LLCs. In this paper, we propose Two-level Eviction Priority (TEP) policy. Besides the eviction priority provided by the baseline replacement policy, TEP appends an additional high level of eviction priority to LLC blocks, which is decided at the insertion time and cannot be changed during their lifetime in the LLC. When blocks with high eviction priority are not in inner caches anymore, they get evicted from the LLC preferentially. Thus, the LLC can retain more useful blocks to improve performance. TEP can cooperate well with various baseline replacement policies. Our evaluation shows that TEP with NRU can improve the performance of inclusive LLCs significantly while requiring negligible extra storage. It also outperforms other recent proposals including QBS, DIP, and DRRIP.

To reduce the increasing cost of building and operating cloud data centers, cloud providers are seeking various mechanisms to achieve higher resource effectiveness. For example, cloud operators are leveraging dynamic resource management techniques to consolidate a higher density of application workloads into commodity physical servers to maximize server resource utilization. However, higher workload density is a major source of performance interference problems in multi-tenant clouds. Existing performance isolation techniques such as dedicated CPU cores for specific workloads are not enough as there are still common resource (e.g., last-level cache and memory bandwidth in memory subsystem) on the processor that are shared among all CPUs on the same NUMA node. While prior work has proposed a variety of resource partitioning techniques, it still remains unexplored to characterize the impact of memory subsystem resource partitioning for the consolidated workloads with different priorities and investigate software support to dynamically manage memory subsystem resource sharing in a real-time manner. To bridge the gap, we propose Themis, a feedback-based controller that enables a priority-aware and fairness-aware memory subsystem resource management strategy to guarantee the performance of high-priority workloads while maintaining fairness across all colocated workloads in high-density clouds. Themis is evaluated with multiple typical cloud applications in our data center environment. The results show that Themis improves the performance of various workloads by up to 3.15%, and fairness by more than 70% in memory subsystem resource allocation compared to existing state-of-the-art work.

On multi-core processors, contention on shared resources such as the last level cache (LLC) and memory bandwidth may cause serious performance degradation, which makes efficient resource allocation a critical issue in data centers. Intel recently introduces Memory Bandwidth Allocation (MBA) technology on its Xeon scalable processors, which makes it possible to allocate memory bandwidth in a real system. However, how to make the most of MBA to improve system performance remains an open question. In this work, (1) we formulate a quantitative relationship between a program's performance and its LLC occupancy and memory request rate on commodity processors. (2) Guided by the performance formula, we propose a heuristic bound-aware throttling algorithm to improve system performance and (3) we further develop a hierarchical clustering method to improve the algorithm's efficiency. (4) We implement these algorithms in EMBA, a low-overhead dynamic memory bandwidth scheduling system to improve performance on Intel commodity processors. The results show that, when multiple programs run simultaneously on a multi-core processor whose memory bandwidth is saturated, the programs with high memory bandwidth demand usually use bandwidth inefficiently compared with programs with medium memory bandwidth demand from the perspective of CPU performance. By slightly throttling the former's bandwidth, we can significantly improve the performance of the latter. On average, we improve system performance by 36.9% at the expense of 8.6% bandwidth utilization rate.

Bandwidth non-uniformity in multi-chip GPUs poses a major design challenge for its last-level cache (LLC) architecture. Whereas a memory-side LLC caches data from the local memory partition while being accessible by all chips, an SM-side LLC is private to a chip while caching data from all memory partitions. We find that some workloads prefer a memory-side LLC while others prefer an SM-side LLC, and this preference solely depends on which organization maximizes the effective LLC bandwidth. In contrast to prior work which optimizes bandwidth beyond the LLC, we make the observation that the effective bandwidth ahead of the LLC is critical to end-to-end application performance. We propose Sharing-Aware Caching (SAC) to adopt either a memory-side or SM-side LLC organization by dynamically reconfiguring the routing policies in the intra-chip interconnection network and LLC controllers. SAC is driven by a simple and lightweight analytical model that predicts the impact of data sharing across chips on the effective LLC bandwidth. SAC improves average performance by 76% and 12% (and up to 157% and 49%) compared to a memory-side and SM-side LLC, respectively. We demonstrate significant performance improvements across the design space and across workloads.

Cloud computing is a key technology for online service providers. However, current online service systems experience performance degradation due to the heterogeneous and time-variant incoming of user requests. To address this kind of diversity, we propose a hierarchical approach for resource management in hybrid clouds, where local private clouds handle routine requests and a powerful third-party public cloud is responsible for the burst of sudden incoming requests. Our goal is to answer (1) from the online service provider’s perspective, how to decide the local private cloud resource allocation, and how to distribute the incoming requests to private and/or public clouds; and (2) from the public cloud provider’s perspective, how to decide the optimal prices for these public cloud resources so as to maximize its profit. We use a Stackelberg game model to capture the complex interactions between users, online service providers and public cloud providers, based on which we analyze the resource allocation in private clouds and pricing strategy in public cloud. Furthermore, we design efficient online algorithms to determine the public cloud provider’s and the online service provider’s optimal decisions. Simulation results validate the effectiveness and efficiency of our proposed approach.

In a typical commercial multi-core processor, the last level cache (LLC) is shared by two or more cores. Existing studies have shown that the shared LLC is beneficial to concurrent query processes with commonly shared data sets. However, the shared LLC can also be a performance bottleneck to concurrent queries, each of which has private data structures, such as a hash table for the widely used hash join operator, causing serious cache conflicts. We show that cache conflicts on multi-core processors can significantly degrade overall database performance. In this paper, we propose a hybrid system method called MCC-DB for accelerating executions of warehouse-style queries, which relies on the DBMS knowledge of data access patterns to minimize LLC conflicts in multi-core systems through an enhanced OS facility of cache partitioning. MCC-DB consists of three components: (1) a cacheaware query optimizer carefully selects query plans in order to balance the numbers of cache-sensitive and cache-insensitive plans; (2) a query execution scheduler makes decisions to co-run queries with an objective of minimizing LLC conflicts; and (3) an enhanced OS kernel facility partitions the shared LLC according to each query's cache capacity need and locality strength. We have implemented MCC-DB by patching the three components in PostgreSQL and Linux kernel. Our intensive measurements on an Intel multi-core system with warehouse-style queries show that MCC-DB can reduce query execution times by up to 33%.

An increasing demand is seen from enterprises to host and dynamically manage middlebox services in public clouds in order to leverage the same benefits that network functions provide in traditional, in-house deployments. However, today's public clouds provide only a limited view and programmability for tenants that challenges flexible deployment of transparent, software-defined network functions. Moreover, current virtual network functions can't take full advantage of a virtualized cloud environment, limiting scalability and fault tolerance. In this paper we review and evaluate the current infrastructural limitations imposed by public cloud providers and present the design and implementation of GNFC, a cloud-based Network Function Virtualization (NFV) framework that gives tenants the ability to transparently attach stateless, container-based network functions to their services hosted in public clouds. We evaluate the proposed system over three public cloud providers (Amazon EC2, Microsoft Azure and Google Compute Engine) and show the effects on end-to-end latency and throughput using various instance types for NFV hosts.

To enjoy the full benefits of cloud computing, it is necessary to have built-in security, privacy, compliance and legal requirements for the cloud implementation and use. Each industry has its own risk levels that it could work within. A company, which is planning to use public cloud services, should be conscious of the regulations and industry risks and need to monitor it and abide by the same. This is due to multi-tenant and open to all nature of public cloud. A thorough risk analysis must be done initially with a public cloud provider. The main objective is to identify the existing vulnerabilities and to implement the measures to counter those threats. There are a variety of risks such as vendor lock-in, non-compliance, poor provisioning, unauthorized access, loss of control, Service Level Agreement (SLA) violations, Internet attacks, etc. To alleviate the risks, there are several measures that could be deployed. This paper deals with mitigation mechanism, security requirements and various risks associated with public cloud.

The advent of cloud computing has made access to computing infrastructure available to millions of users that face resource constraints. In the context of high performance computing (HPC), public cloud resources have emerged as a cost‐effective alternative to expensive on‐premises clusters. However, there are several challenges and limitations in adopting this approach. This paper proposes HPC@Cloud , a provider‐agnostic open‐source software toolkit that facilitates the migration, testing, and execution of HPC applications in public clouds. The toolkit takes advantage of various fault tolerance technologies to enable the use of inexpensive transient cloud infrastructure, commonly known as “spot” instances. Also, it features integration with singularity containers, allowing users to run complex applications on virtual HPC clusters in a portable and reproducible way. Finally, it provides a data‐based empirical approach to estimating cloud infrastructure costs for HPC workloads. The results obtained on two public cloud providers (AWS and Vultr) show that: (i) HPC@Cloud can efficiently build virtual HPC clusters on the cloud; (ii) the new adaptive fault tolerance strategy outperforms other existing strategies based on blocking restoration; (iii) the integration of singularity containers into HPC@Cloud improves the portability of HPC applications to public clouds with negligible performance penalty to the applications; (iv) the proposed cost prediction approach can estimate the cost of running the applications on AWS and Vultr with up to 93% accuracy on average.

The workloads of social computing are tremendous, bursty and unpredictable. A hybrid cloud, which combines a private cloud and a public cloud, is a promising solution to processing the workloads. For most corporations, they leverage one public cloud. However, with fierce competition among public cloud providers, public cloud services change frequently, which may lead to service unavailability and a less cost-effective hybrid cloud solution. As a result, leveraging multiple public clouds in the hybrid cloud is a potential solution. In this paper, we identify such a problem in current hybrid cloud and analyze the necessity of load balancing for hybrid cloud applications. Focusing on cost minimization and performance guarantee, we propose a Least Cost per Connection (LCC) algorithm so as to choose the most cost-effective clouds along with adapting changes among multiple public clouds. The simulation results show that our solution can significantly decrease the outsourcing cost as well as guarantee Quality of Service (QoS) of applications.

In recent years, features and applications of embedded systems have been increasing rapidly. Chip Multi-Processors (CMPs), have been used in these systems to meet the higher demand for performance and energy efficiency. In CMPs, the last level cache (LLC) and the memory bandwidth are usually shared by the cores. Despite the fact that CMPs improve performance of embedded systems, competition for the shared resources makes their performance unpredictable and suboptimal. In this paper, we propose PLSS: Phase-guided Locality Signature based Scheduler for arbitrating LLC requests in multi-core embedded processors. To achieve our goal, we perform phase-wise offline profiling to guide the runtime task scheduling scheme. Our approach can improve performance of dual core system by upto 11% over IPC based scheduler (5% on average) and 35% over LLC number-of-accesses based approach (6.5% on average).

Cloud computing technology is increasingly important in the industry with the promise of increased performance, dynamic resource allocation capabilities and provides a convincing opportunity for organizations to outsource their IT infrastructure under the pay-per-use model offered by many public cloud providers. But there are concerns when selecting between the private cloud and the public cloud that must be taken into account. Many SME enterprises with scarce in-house IT support and limited knowledge about different cloud technologies encounter difficulties in making the most appropriate choice on 'buying' a private cloud or 'renting' a public cloud service. Through a number of interviews with potential cloud adopters from the industry, this paper identifies the most relevant factors concerning cloud adoption and develops a cloud deployment model assessment method called Cloud Deployment Selection Model (CDSM). The model has been validated in real case studies, and recommendations derived have been compared with real adoption cases. Result shows it is able to accurately recommend a suitable cloud deployment model based on the factors identified from many SME organizations, an important tool for SMEs to decide between the private or the public cloud solution.

After years of development, cloud computing has developed into public cloud, private cloud, community cloud, hybrid cloud and other cloud forms. However, the government, financial and other industries have more concerns about security, operation and maintenance, cost and other issues in the process of cloud. In this regard, the industry, academia and other parties have launched the frontier exploration, give different solutions. To meet users’ new needs, we proposed a new model that combines the advantages of public and private clouds. It is called proprietary cloud. Proprietary cloud provides users with more efficient management by incorporating years of experience of public cloud providers. On the one hand, it meets the rigid needs of users’ independent control and exclusive resources, on the other hand, it reduces the maintenance cost and downtime probability of users, and achieves the desired effect of users. It mainly includes self-built model and exclusive model deployment modes. The main difference is whether to accept public cloud management. This paper introduces and analyzes proprietary cloud in detail from the aspects of deployment architecture, mode differences and common deployment patterns. It concludes that proprietary cloud is an indispensable cloud computing deployment pattern in the future.KeywordsCloud computingDigital transformationProprietary cloud