Recently, flash-based solid-state drives (SSDs) have become standard options for laptop and desktop storage, but their impact on enterprise server storage has not been studied. Provisioning is challenging. It requires optimizing the performance, capacity, power reliability needs of expected workload, all while minimizing financial costs. In this paper we analyze a number workload traces from servers in both large small data centers, to decide whether how SSDs should be used support each. We...

This paper presents a method for admission control and request scheduling multiply-tiered e-commerce Web sites, achieving both stable behavior during overload improved response times. Our externally observes execution costs of requests online, distinguishing different types, performs protection preferential using relatively simple measurements straight forward mechanism. Unlike previous proposals, which require extensive changes to the server or operating system, our requires no...

Generalized snapshot isolation extends as used in Oracle and other databases a manner suitable for replicated databases. While (conventional) requires that transactions observe the "latest" of database, generalized allows use "older" snapshots, facilitating implementation. We show many desirable properties remain. In particular, read-only never block or abort they do not cause update to abort. Moreover, under certain assumptions on transaction workload execution is serializable. An...

We propose two protocols that provide scalable causal consistency for both partitioned and replicated data stores using dependency matrices (DM) physical clocks. The DM protocol supports basic read update operations uses two-dimensional to track dependencies in a client session. It utilizes the transitivity of causality sparse matrix encoding keep metadata small bounded. DM-Clock extends support read-only transactions loosely synchronized

Developers use Machine Learning (ML) platforms to train ML models and then deploy these as web services for inference (prediction). A key challenge platform providers is guarantee response-time Service Level Agreements (SLAs) workloads while maximizing resource efficiency. Swayam a fully distributed autoscaling framework that exploits characteristics of production deliver on the dual efficiency SLA compliance. Our contributions are (1) model-based takes into account SLAs workload...

Serverless computing is becoming increasingly popular due to its ease of programming, fast elasticity, and fine-grained billing. However, the serverless provider still needs provision, manage, pay IaaS for virtual machines (VMs) hosting platform. This ties cost platform underlying VMs. One way significantly reduce use spare resources, which cloud providers rent at a massive discount. Harvest VMs offer such cheap resources: they grow shrink harvest all unallocated CPU cores in their host...

Web search engines are optimized to reduce the high-percentile response time consistently provide fast responses almost all user queries. This is a challenging task because query workload exhibits large variability, consisting of many short-running queries and few long-running that significantly impact time. With modern multicore servers, parallelizing processing an individual promising solution execution time, but it gives limited benefits compared sequential since most see little or no...

Interactive services, such as Web search, recommendations, games, and finance, must respond quickly to satisfy customers. Achieving this goal requires optimizing tail (e.g., 99th+ percentile) latency. Although every server is multicore, parallelizing individual requests reduce latency challenging because (1) service demand unknown when arrive; (2) blindly all oversubscribes hardware resources; (3) the numerous short will not improve This paper introduces Few-to-Many (FM) incremental...

Bursts in data center workloads are a real problem for storage subsystems. Data volumes can experience peak I/O request rates that over an order of magnitude higher than average load. This requires significant overprovisioning, and often still results latency during peaks.In to address this we propose Everest, which allows written overloaded volume be temporarily off-loaded into short-term virtual store. Everest creates the store by opportunistically pooling underutilized resources either on...

Clock-SI is a fully distributed protocol that implements snapshot isolation (SI) for partitioned data stores. It derives and commit timestamps from loosely synchronized clocks, rather than centralized timestamp authority as used in current systems. A transaction obtains its by reading the clock at originating partition provides corresponding consistent across all partitions. In contrast to using authority, has availability performance benefits: avoids single point of failure potential...

This paper presents a scheduling model for class of interactive services in which requests are time bounded and lower result quality can be traded shorter execution time. These applications include web search engines, finance servers, other interactive, on-line services. We develop an efficient algorithm, Zeta, that allocates processor among service to maximize the minimize variance response.

This paper presents Mercury; a system for real-time support of top-k spatio-temporal queries on microblogs, where users are able to browse recent microblogs near their locations. With high arrival rates Mercury ensures query response within tight memory-constrained environment. bounds its search space include only those that have arrived certain spatial and temporal boundaries, in which the according ranking function, returned results. employs: (a) scalable dynamic in-memory index structure...

We propose a SPARQL-like language, G-SPARQL, for querying attributed graphs. The language expresses types of queries which large interest applications model their data as graphs such as: pattern matching, reachability and shortest path queries. Each query can combine both structural predicates value-based (on the attributes graph nodes edges). describe an algebraic compilation mechanism our proposed is extended from relational algebra based on basic construct building SPARQL queries, Triple...

Interactive service providers have strict requirements on high-percentile (tail) latency to meet user expectations. If tail targets with less energy, they increase profits, because energy is a significant operating expense. Unfortunately, optimizing and are typically conflicting goals. Our work resolves this conflict by exploiting servers per-core Dynamic Voltage Frequency Scaling (DVFS) Asymmetric Multicore Processors (AMPs). We introduce the Adaptive Slow-to-Fast scheduling framework,...

We can increase the efficiency of public cloud datacenters by harvesting allocated but temporarily idling CPU cores from customer virtual machines (VMs) to run batch or analytics workloads. Even small gains translate into substantial savings, since provisioning and operating a datacenter costs hundreds millions dollars per year. The main challenge is harvest idle with little no impact on VMs, which could be running latency-sensitive services are essentially black-boxes provider.

In stand-alone databases, the functions of ordering transaction commits and making effects transactions durable are performed in one single action, namely writing commit record to disk. For efficiency many these writes grouped into a disk operation. replicated databases which all replicas agree on order update transactions, two typically separated. Specifically, replication middleware determines global order, while database make durable.The contribution this paper is demonstrate that...

This paper presents a method that allows replicated database system to provide global isolation level stronger than the provided on each individual replica. We propose new multi-version concurrency control algorithm called, serializable generalized snapshot (SGSI), targets middleware systems. Each replica runs locally and replication guarantees one-copy serializability. introduce novel techniques level, namely readset extraction enhanced certification prevents read-write write-write...

A web search query made to Microsoft Bing is currently parallelized by distributing the processing across many servers. Within each of these servers, is, however, processed sequentially. Although server may be multiple queries concurrently, with modern multicore parallelizing an individual within nonetheless improve user's experience reducing response time. In this paper, we describe issues that make parallelization a challenging, and present approach effectively addresses challenges. Since...

Graphs are used in many large-scale applications, such as social networking. The management of these graphs poses new challenges too large for a single server to manage efficiently. Current distributed techniques map-reduce and Pregel not well-suited processing interactive ad-hoc queries against graphs. In this paper we demonstrate Horton, query execution engine Horton defines language that allows the expression regular reach ability provides with optimizer on parallel. demo, show...

Horton+ is a graph query processing system that executes declarative reachability queries on partitioned attributed multi-graph. It employs language, optimizer, and distributed execution engine. The language expresses queries, supports closures predicates node edge attributes to match paths. We introduce three algebraic operators, select, traverse, join, compiled into an plan containing these operators. As access the elements in random pattern, therefore maintained main memory of cluster...

A commercial web search engine shards its index among many servers, and therefore the response time of a query is dominated by slowest server that processes query. Prior approaches target improving responsiveness reducing tail latency an individual server. They predict execution time, if predicted to be long-running, it runs in parallel, otherwise sequentially. These are, however, not accurate enough for high when responses are aggregated from servers because this requires each reduce...

Since the mid-90s there has been a widely-held belief that signature files are inferior to inverted for text indexing. In recent years Bing search engine developed and deployed an index based on bit-sliced signatures. This index, known as BitFunnel, replaced existing production system index. The driving factor behind shift away from was operational cost savings. paper describes algorithmic innovations changes in cloud computing landscape led us reconsider eventually field technology once...