Data centers are often under-utilized due to over-provisioning as well time-varying resource demands of typical enterprise applications. One approach increase utilization is consolidate applications in a shared infrastructure using virtualization. Meeting application-level quality service (QoS) goals becomes challenge consolidated environment application needs differ. Furthermore, for multi-tier applications, the amount resources needed achieve their QoS might be different at each tier and...

Main memory database systems (MMDBs) store their data in main physical and provide very high-speed access. Conventional are optimized for the particular characteristics of disk storage mechanisms. Memory resident systems, on other hand, use different optimizations to structure organize data, as well make it reliable. The authors survey major residence briefly discuss some MMDBs that have been designed or implemented.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML"...

Long lived transactions (LLTs) hold on to database resources for relatively long periods of time, significantly delaying the termination shorter and more common transactions. To alleviate these problems we propose notion a saga. A LLT is saga if it can be written as sequence that interleaved with other The management system guarantees either all in are successfully completed or compensating run amend partial execution. Both concept its implementation simple, but they have potential improve...

Just like parallel processing elements can substantially speed up computationally intensive tasks, concurrent transfer of data in and out memory tasks. In this paper we study one general purpose facility for achieving motion: disk striping. A group disks is striped if each block multiplexed across all the disks. Since subblock a different device, input output proceed parallel. With help an analytical model, investigate effect striping on service times its advantages limitations set...

Long lived transactions (LLTs) hold on to database resources for relatively long periods of time, significantly delaying the termination shorter and more common transactions. To alleviate these problems we propose notion a saga. A LLT is saga if it can be written as sequence that interleaved with other The management system guarantees either all in are successfully completed or compensating run amend partial execution. Both concept its implementation simple, but they have potential improve...

Snapshot isolation is a popular transactional level in database systems. Several replication techniques based on snapshot have recently been proposed. These proposals, however, do not fully leverage the local concurrency controls that provide isolation. Furthermore, guaranteeing lazy replicated systems may result transaction inversions, which happen when transactions see stale data. Strong isolation, provided centralized servers, avoids inversions but expensive to system. In this paper, we...

Data centers are often under-utilized due to over-provisioning as well time-varying resource demands of typical enterprise applications. One approach increase utilization is consolidate applications in a shared infrastructure using virtualization. Meeting application-level quality service (QoS) goals becomes challenge consolidated environment application needs differ. Furthermore, for multi-tier applications, the amount resources needed achieve their QoS might be different at each tier and...

Virtual machine monitors are becoming popular tools for the deployment of database management systems and other enterprise software applications. In this paper, we consider a common resource consolidation scenario, in which several system instances, each running virtual machine, sharing pool physical computing resources. We address problem optimizing performance these by controlling configurations machines they run. These determine how shared resources will be allocated to different...

Database design is critical for high performance in relational databases and a myriad of tools exist to aid application designers selecting an appropriate schema. While the problem schema optimization also highly relevant NoSQL databases, existing are inadequate that setting. Application wishing use database instead rely on rules thumb select We present system recommending schemas applications. Our cost-based approach uses novel binary integer programming formulation guide mapping from...

An adaptive technique for reducing disk seek times is described. The copies frequently referenced blocks from their original locations to reserved space near the middle of disk. Reference frequencies need not be known in advance. Instead, they are estimated by monitoring stream arriving requests. Trace-driven simulations show that can cut substantially copying only a small number using this technique. has been implemented modifying UNIX device driver. No modifications required file system uses

In this paper we present a technique for building high-availability (HA) database management system (DBMS). The proposed can be applied to any DBMS with little or no customization, and reasonable performance overhead. Our approach is based on Remus, commodity HA solution implemented in the virtualization layer, that uses asynchronous virtual machine (VM) state replication provide transparent failover capabilities. We show while Remus similar systems protect DBMS, workloads incur overhead of...

Providing the ability to elastically use more or fewer servers on demand (scale out and scale in) as load varies is essential for database management systems (DBMSes) deployed today's distributed computing platforms, such cloud. This requires solving problem of dynamic (online) data placement, which has so far been addressed only workloads where all transactions are local one sever. In DBMSes ACID can access than partition, represent a major performance bottleneck. Scaling spreading across...

Incremental refresh of a materialized join view is often less expensive than full, non-incremental refresh. However, it still potentially costly atomic operation. This paper presents an algorithm that performs incremental maintenance as series small, asynchronous steps. The size each step can be controlled to limit contention between the process and concurrent operations access or underlying relations. supports point-in-time refresh, which allows refreshed any time last present.

Long-lived transactions (LLTs) hold on to database resources for relatively long periods of time, significantly delaying the completion shorter and more common transactions. To alleviate this problem we propose an extension two-phase locking, called altruistic whereby LLTs can release their locks early. Transactions that access released data are said run in wake LLT must follow special locking rules. Like is easy implement guarantees serializability.

A database system is considered in which a main-memory holds all data semiconductor memory, and for recovery purposes backup copy of the maintained secondary storage. The checkpointer component crash manager responsible maintaining copy. Ideally, should maintain an almost-up-to-date while interfering as little possible with system's transaction processing activities. Several algorithms such are presented compared using analytic model. results show some significant performance differences...

Lazy replication is a popular technique for improving the performance and availability of database systems. Although there are concurrency control techniques, which guarantee serializability in lazy systems, these techniques result undesirable transaction orderings. Since transactions may see stale data, they be serialized an order different from one were submitted. Strong avoids such problems, but it very costly to implement. We propose generalized form strong that suitable use with...

System M is an experimental transaction processing testbed that runs on top of the Mach operating system. Its database stored in primary memory. The structure and algorithms used are described. checkpointer component periodically sweeps memory propagates updates to a backup copy disk. Several different checkpointing (and logging) were implemented, their performance was experimentally evaluated.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Virtual machine monitors are becoming popular tools for the deployment of database management systems and other enterprise software. In this article, we consider a common resource consolidation scenario in which several system instances, each running separate virtual machine, sharing pool physical computing resources. We address problem optimizing performance these by controlling configurations machines they run. These determine how shared resources will be allocated to different instances....

The use of flash-based solid state drives (SSDs) in storage systems is growing. Adding SSDs to a system not only raises the question how manage SSDs, but also whether current buffer pool algorithms will still work effectively. We are interested hybrid systems, consisting and hard disk (HDDs), for database management. present cost-aware replacement algorithms, which aware difference performance between HDDs, both DBMS SSDs. In physical access pattern depends on management pool. studied impact...