Sameh Elnikety

A method for transparent admission control and request scheduling in e-commerce web sites

This paper presents a method for admission control and request scheduling for multiply-tiered e-commerce Web sites, achieving both stable behavior during overload and improved response times. Our method externally observes execution costs of requests online, distinguishing different request types, and performs overload protection and preferential scheduling using relatively simple measurements and a straight forward control mechanism. Unlike previous proposals, which require extensive changes to the server or operating system, our method requires no modifications to the host O.S., Web server, application server or database. Since our method is external, it can be implemented in a proxy. We present such an implementation, called Gatekeeper, using it with standard software components on the Linux operating system. We evaluate the proxy using the industry standard TPC-W workload generator in a typical three-tiered e-commerce environment. We show consistent performance during overload and throughput increases of up to 10 percent. Response time improves by up to a factor of 14, with only a 15 percent penalty to large jobs.

Migrating server storage to SSDs: analysis of tradeoffs

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 server storage is challenging. It requires optimizing for the performance, capacity, power and reliability needs of the expected workload, all while minimizing financial costs. In this paper we analyze a number of workload traces from servers in both large and small data centers, to decide whether and how SSDs should be used to support each. We analyze both complete replacement of disks by SSDs, as well as use of SSDs as an intermediate tier between disks and DRAM. We describe an automated tool that, given device models and a block-level trace of a workload, determines the least-cost storage configuration that will support the workloads performance, capacity, and fault-tolerance requirements. We found that replacing disks by SSDs is not a costeffective option for any of our workloads, due to the low capacity per dollar of SSDs. Depending on the workload, the capacity per dollar of SSDs needs to increase by a factor of 3-3000 for an SSD-based solution to break even with a diskbased solution. Thus, without a large increase in SSD capacity per dollar, only the smallest volumes, such as system boot volumes, can be cost-effectively migrated to SSDs. The benefit of using SSDs as an intermediate caching tier is also limited: fewer than 10% of our workloads can reduce provisioning costs by using an SSD tier at todays capacity per dollar, and fewer than 20% can do so at any SSD capacity per dollar. Although SSDs are much more energy-efficient than enterprise disks, the energy savings are outweighed by the hardware costs, and comparable energy savings are achievable with low-power SATA disks.

Orbe: scalable causal consistency using dependency matrices and physical clocks

We propose two protocols that provide scalable causal consistency for both partitioned and replicated data stores using dependency matrices (DM) and physical clocks. The DM protocol supports basic read and update operations and uses two-dimensional dependency matrices to track dependencies in a client session. It utilizes the transitivity of causality and sparse matrix encoding to keep dependency metadata small and bounded. The DM-Clock protocol extends the DM protocol to support read-only transactions using loosely synchronized physical clocks. We implement the two protocols in Orbe, a distributed key-value store, and evaluate them experimentally. Orbe scales out well, incurs relatively small overhead over an eventually consistent key-value store, and outperforms an existing system that uses explicit dependency tracking to provide scalable causal consistency.

Tashkent: uniting durability with transaction ordering for high-performance scalable database replication

In stand-alone databases, the functions of ordering the transaction commits and making the effects of transactions durable are performed in one single action, namely the writing of the commit record to disk. For efficiency many of these writes are grouped into a single disk operation. In replicated databases in which all replicas agree on the commit order of update transactions, these two functions are typically separated. Specifically, the replication middleware determines the global commit order, while the database replicas make the transactions durable.The contribution of this paper is to demonstrate that this separation causes a significant scalability bottleneck. It forces some of the commit records to be written to disk serially, where in a standalone system they could have been grouped together in a single disk write. Two solutions are possible: (1) move durability from the database to the replication middleware, or (2) keep durability in the database and pass the global commit order from the replication middleware to the database.We implement these two solutions. Tashkent-MW is a pure middleware solution that combines durability and ordering in the middleware, and treats an unmodified database as a black box. In Tashkent-API, we modify the database API so that the middleware can specify the commit order to the database, thus, combining ordering and durability inside the database. We compare both Tashkent systems to an otherwise identical replicated system, called Base, in which ordering and durability remain separated. Under high update transaction loads both Tashkent systems greatly outperform Base in throughput and response time.

Tashkent+: memory-aware load balancing and update filtering in replicated databases

We present a memory-aware load balancing (MALB) technique to dispatch transactions to replicas in a replicated database. Our MALB algorithm exploits knowledge of the working sets of transactions to assign them to replicas in such a way that they execute in main memory, thereby reducing disk I/O. In support of MALB, we introduce a method to estimate the size and the contents of transaction working sets. We also present an optimization called update filtering that reduces the overhead of update propagation between replicas. We show that MALB greatly improves performance over other load balancing techniques -- such as round robin, least connections, and locality-aware request distribution (LARD) -- that do not use explicit information on how transactions use memory. In particular, LARD demonstrates good performance for read-only static content Web workloads, but it gives performance inferior to MALB for database workloads as it does not efficiently handle large requests. MALB combined with update filtering further boosts performance over LARD. We build a prototype replicated system, called Tashkent+, with which we demonstrate that MALB and update filtering techniques improve performance of the TPC-W and RUBiS benchmarks. In particular, in a 16-replica cluster and using the ordering mix of TPC-W, MALB doubles the throughput over least connections and improves throughput 52% over LARD. MALB with update filtering further improves throughput to triple that of least connections and more than double that of LARD. Our techniques exhibit super-linear speedup; the throughput of the 16-replica cluster is 37 times the peak throughput of a standalone database due to better use of the clusters memory.

G-SPARQL: a hybrid engine for querying large attributed graphs

We propose a SPARQL-like language, G-SPARQL, for querying attributed graphs. The language expresses types of queries which of large interest for applications which model their data as large graphs such as: pattern matching, reachability and shortest path queries. Each query can combine both of structural predicates and value-based predicates (on the attributes of the graph nodes and edges). We describe an algebraic compilation mechanism for our proposed query language which is extended from the relational algebra and based on the basic construct of building SPARQL queries, the Triple Pattern. We describe a hybrid Memory/Disk representation of large attributed graphs where only the topology of the graph is maintained in memory while the data of the graph is stored in a relational database. The execution engine of our proposed query language splits parts of the query plan to be pushed inside the relational database while the execution of other parts of the query plan are processed using memory-based algorithms, as necessary. Experimental results on real datasets demonstrate the efficiency and the scalability of our approach and show that our approach outperforms native graph databases by several factors.

Predictive parallelization: taming tail latencies in web search

Web search engines are optimized to reduce the high-percentile response time to consistently provide fast responses to almost all user queries. This is a challenging task because the query workload exhibits large variability, consisting of many short-running queries and a few long-running queries that significantly impact the high-percentile response time. With modern multicore servers, parallelizing the processing of an individual query is a promising solution to reduce query execution time, but it gives limited benefits compared to sequential execution since most queries see little or no speedup when parallelized. The root of this problem is that short-running queries, which dominate the workload, do not benefit from parallelization. They incur a large parallelization overhead, taking scarce resources from long-running queries. On the other hand, parallelization substantially reduces the execution time of long-running queries with low overhead and high parallelization efficiency. Motivated by these observations, we propose a predictive parallelization framework with two parts: (1) predicting long-running queries, and (2) selectively parallelizing them. For the first part, prediction should be accurate and efficient. For accuracy, we study a comprehensive feature set covering both term features (reflecting dynamic pruning efficiency) and query features (reflecting query complexity). For efficiency, to keep overhead low, we avoid expensive features that have excessive requirements such as large memory footprints. For the second part, we use the predicted query execution time to parallelize long-running queries and process short-running queries sequentially. We implement and evaluate the predictive parallelization framework in Microsoft Bing search. Our measurements show that under moderate to heavy load, the predictive strategy reduces the 99th-percentile response time by 50% (from 200 ms to 100 ms) compared with prior approaches that parallelize all queries.

Zeta: scheduling interactive services with partial execution

This paper presents a scheduling model for a class of interactive services in which requests are time bounded and lower result quality can be traded for shorter execution time. These applications include web search engines, finance servers, and other interactive, on-line services. We develop an efficient scheduling algorithm, Zeta, that allocates processor time among service requests to maximize the quality and minimize the variance of the response. Zeta exploits the concavity of the request quality profile to distribute processing time among outstanding requests. By executing some requests partially (and obtaining much or most benefit of a full execution), Zeta frees resources for other requests, which might have timed out and produced no results. Compared to scheduling algorithms that consider only deadline or quality profile information, Zeta improves overall response quality and reduces response quality variance, yielding significant improvement in the high-percentile response quality. We implemented and deployed Zeta in the Microsoft Bing web search engine and evaluated its performance in a production environment with realistic workloads. Measurements show that at the same response quality and latency as the production system, Zeta increases system capacity by 29% by improving both average and high percentile response quality. We also implemented Zeta in a finance server that computes option prices. In this application, Zeta improves average response quality by 28% and the 99-percentile quality by 80%. Using a simulation, we also compared Zeta to the offline optimal schedule and other scheduling algorithms. Although Zeta is only close to optimal, it provides better performance than prior algorithms under a wide variety of operating conditions.

Clock-SI: Snapshot Isolation for Partitioned Data Stores Using Loosely Synchronized Clocks

Clock-SI is a fully distributed protocol that implements snapshot isolation (SI) for partitioned data stores. It derives snapshot and commit timestamps from loosely synchronized clocks, rather than from a centralized timestamp authority as used in current systems. A transaction obtains its snapshot timestamp by reading the clock at its originating partition and Clock-SI provides the corresponding consistent snapshot across all the partitions. In contrast to using a centralized timestamp authority, Clock-SI has availability and performance benefits: It avoids a single point of failure and a potential performance bottleneck, and improves transaction latency and throughput. We develop an analytical model to study the trade-offs introduced by Clock-SI among snapshot age, delay probabilities of transactions, and abort rates of update transactions. We verify the model predictions using a system implementation. Furthermore, we demonstrate the performance benefits of Clock-SI experimentally using a micro-benchmark and an application-level benchmark on a partitioned key-value store. For short read-only transactions, Clock-SI improves latency and throughput by 50% by avoiding communications with a centralized timestamp authority. With a geographically partitioned data store, Clock-SI reduces transaction latency by more than 100 milliseconds. Moreover, the performance benefits of Clock-SI come with higher availability.

Horton: Online Query Execution Engine for Large Distributed Graphs

Graphs are used in many large-scale applications, such as social networking. The management of these graphs poses new challenges as such graphs are too large for a single server to manage efficiently. Current distributed techniques such as map-reduce and Pregel are not well-suited to processing interactive ad-hoc queries against large graphs. In this paper we demonstrate Horton, a distributed interactive query execution engine for large graphs. Horton defines a query language that allows the expression of regular language reach ability queries and provides a query execution engine with a query optimizer that allows interactive execution of queries on large distributed graphs in parallel. In the demo, we show the functionality of Horton managing a large graph for a social networking application called Codebook, whose graph represents data on software components, developers, development artifacts such as bug reports, and their interactions in large software projects.