Page Allocation Research Articles

Editorial for April, 2011 Issue JPVT

Editorial for April, 2011 Issue JPVT

Editorial: Phys. Status Solidi RRL 1/2011

Happy 50th anniversary pss, but your youngest family member pss – Rapid Research Letters is turning only four, and the word about its fast track handling service is still being spread throughout the solid state physics community. 18 months after the first Impact Factor release and with submission and publication numbers rising steeply (see Fig. 1), we have arrived at a point where we can put the announcements and advertisements on solid statistical grounds in more detail. Manuscript submissions to pss (RRL) and published Letters since journal launch. pss (RRL) was designed as a double peer-review, rapid publication medium for high-quality content with need for express communication to the community. Submissions enjoy priority editorial handling, and the time to the first editorial decision is kept as brief as possible by the commitment of referees who are reporting on the three-page manuscripts on short notice. The acceptance rate of 25% in 2009 and 2010 is partly due to the fact that we refer good, but less urgent novel results to the regular editorial handling of the sister journals pss (a) and (b) prior to peer review. Along with maintaining a large international referee panel, these choices are a measure to reduce the pressure of urgent review requests on the schedule of busy experts. Purely editorial and/or board decisions are communicated to the authors within a week, but we will concentrate on the statistics for published papers in the following. So let us take all regular submissions that were accepted as RRLs in 2009 and 2010 and look at how concept and organization are faring in every-day reality (Focus Issues excluded, since special projects may differ in handling). Figure 2 shows the time that elapsed between submission and acceptance of a Letter in calendar days, including peer review, editorial decisions, and author revisions if required. We are proud to say that half of the final decisions are reached within a month. In spite of the pronounced rise in submission and publication numbers (Fig. 1) this median is fairly constant over the four years, ranging between 24 and 29 days. Extracting the cases where acceptances did not require any revision we still arrive at a median of only 16 calendar days, in line with our two-week target for outstanding submissions. We note however, that this applied to only 7% of all acceptances in 2009/10, and observe the trend with mixed emotions. After all, the tendency of our reviewers to be more careful and critical about Rapid Research Letter manuscripts reflects the rise in quality expectations and esteem, and seemed particularly pronounced ever since the Impact Factors of 2.15 (June 2009) and 2.56 (current) were released. Taking the time from receipt of the final version of a manuscript (revised, or no revision required) to acceptance as displayed in Fig. 3, the spectacular-looking median of only 2 calendar days reflects the handling at the editorial office rather than the time required for peer review. Number of accepted manuscripts versus days to decision, individual count (red) and accumulated count (green). Number of accepted manuscripts versus days elapsed since receipt of the final manuscript version (revised versions or initially submitted version when no revision was required), individual count (red) and accumulated count (green). Only slightly longer and even more spectacular for all the processing steps included are the times from acceptance to online publication shown in Fig. 4. Number of online-published Letters versus days from acceptance, individual count (red) and accumulated count (green). Once in the hands of the our Technical Editor Heike Höpcke and the electronic publishing department, Letters are copy-edited, typeset, proof-read by the authors, corrected, and published online in final form with issue and page allocation as well as additional material (graphical table of content entry, supporting information) within 5 calendar days median. All data considered, we gladly observe that pss (RRL)'s fast track publication service is indeed working as projected, achieving an excellent combination of speed and quality. And obviously, our capacity for high-quality content with need for express release to the solid state physics community is not nearly reached. Starting with a [email protected] on 'Competing energy scales in high-temperature superconductors: Ultrafast pump–probe studies' by Elbert E. M. Chia et al. in the current issue, pss (RRL) is now also launching a new article type. In line with the journal's dedication to most recent novel results, a [email protected] will summarize progress in particularly active fields over the last three years. Looking forward to great papers in 2011.

Business Models

Business Models

Generalized ERSS tree model: Revisiting working sets

Accurately characterizing the resource usage of an application at various levels in the memory hierarchy has been a long-standing research problem. Existing characterization studies are either motivated by specific allocation problems (e.g., memory page allocation) or they characterize a specific memory resource (e.g., L2 cache). The studies thus far have also implicitly assumed that there is no contention for the resource under consideration. The inevitable future of virtualization driven consolidation necessitates the sharing of physical resources at all levels of the memory hierarchy by multiple virtual machines (VMs). Given the lack of resource isolation mechanisms at several levels of the memory hierarchy within current commodity systems, provisioning resources for a virtualized application not only requires a precise characterization of its resource usage but must also account for the impact of resource contention due to other co-located applications during its lifetime. In this paper, we present a unifying Generalized ERSS Tree Model that characterizes the resource usage at all levels of the memory hierarchy during the entire lifetime of an application. Our model characterizes capacity requirements, the rate of use, and the impact of resource contention, at each level of memory. We present a methodology to build the model and demonstrate how it can be used for the accurate provisioning of the memory hierarchy in a consolidated environment. Empirical results suggest that the Generalized ERSS Tree Model is effective at characterizing applications with a wide variety of resource usage behaviors.

A New Era Begins

The journal Stroke celebrates its 40th year and there is much to be proud of based on prior accomplishments. I begin my 5-year term as Editor-in-Chief inheriting a flourishing, successful and highly respected journal that has thrived under the leadership of Dr Vladimir Hachinski, who has served as Editor-in-Chief for the past 10 years. During Dr. Hachinski’s term the number of manuscripts submitted yearly has doubled to just over 2800 and the impact factor has risen by a point to 6.5. Additionally, publishing ahead of print began, as did a completely electronic manuscript submission system. Translation of Stroke into 7 languages increased the submission and publication of manuscripts, and the number of subscriptions, from these regions. An author-mentoring program for non-native English speakers was initiated along with a waiver of page charges for authors from many less developed countries. These accomplishments, along with many others too numerous to enumerate, have marked the stewardship provided by Dr Hachinski and will provide a sustainable legacy for the field of cerebrovascular disorders. Dr Hachinski can be described as an internationalist with insightful ideas who has strived to be inclusive and innovative. I submit that he has excelled by providing all of his visionary initiatives and I offer him my thanks, and those of our entire community, …

Runtime monitoring on multicores via OASES

Runtime monitoring support serves as a foundation for the important tasks of providing security, performing debugging, and improving performance of applications. Often runtime monitoring requires the maintenance of information associated with each of the application's original memory location, which is held in corresponding shadow memory locations. Unfortunately, existing robust shadow memory implementations are inefficient. In this paper, we present OASES: OS and Architectural Support for Efficient Shadow memory implementation for multicores that is also robust. A combination of operating system support (in the form of coupled allocation of memory pages used by the application and associated shadow memory pages) and architectural support (in the form of ISA support and exposed cache events) is proposed. Our page allocation policy enables fast translation of original addresses into corresponding shadow memory addresses; thus allowing implicit addressing of shadow memory. By exposing the cache events to the software, we ensure in software that the shadow memory instructions execute atomically with their corresponding original memory instructions. Our experiments show that the overheads of runtime monitoring tasks are significantly reduced in comparison to previous software implementations.

Design and implementation of a uniformity-improving page allocation scheme for flash-based storage systems

Flash memory is being actively employed in a variety of embedded systems such as digital cameras, MP3 players, cell phones, solid state disks (SSDs), and digital media broadcasting (DMB) devices. This paper considers performance issues in file systems that employ Flash memory as a storage medium. Firstly, it explores the characteristics of Flash memory and identifies the cost of block cleaning as the key performance bottleneck for Flash memory analogous to the seek time in disk storage. Then, it defines three performance parameters, namely, utilization, invalidity, and uniformity and derives a formula for block cleaning cost based on these parameters. It is shown that, of these parameters, uniformity exerts the strongest influence on the cost of cleaning and that uniformity is a file system controllable parameter. Finally, we design a uniformity-aware page allocation scheme and analyze how enhanced uniformity affects the block cleaning cost with various workloads. Real implementation experiments conducted on an embedded system show that the scheme proposed here typically reduces the cleaning time by 20 to 30% compared to the traditional sequential allocation scheme that is used in YAFFS.

Where to now for our journal?

As this will be my last editorial for the Australian Occupational Therapy Journal (AOTJ), I decided to review the aims and objectives of the journal. It is pleasing to note that they are being fulfilled. We do produce a high-quality peer-reviewed journal, which promotes research and communication and provides a forum for discussion and debate. The manuscripts published are relevant to occupational therapy theory, practice and education, and we are able to provide analysis and commentary of other research articles in the form of Critically Appraised Papers. We have a large number of voluntary reviewers who donate their time to review submitted manuscripts – in fact we had 140 individual reviewers over the last 18 months, some providing reviews for four different manuscripts. It is also encouraging to note that online downloads have increased from 2005 by 49%, to 135 939 in 2006! This demonstrates the increasing use and significance of the journal as a professional, clinical, educational and research resource. Over the 2-year period 2004–2006, the submission rate of all manuscripts increased by more than 100% including an increase of 30% in international manuscripts over the period 2005–2006. It is also interesting to note that the cost per member for the AOTJ has reduced from $A20.00 in 2005, to $A16.00 per annum in 2006. The members of the AOTJ Editorial Board feel this is a small percentage of the National Levy to be paying for such a quality publication. The AOTJ is the public face of occupational therapy and OT AUSTRALIA nationally and internationally. The increased submission rate is related to a number of factors. There are more occupational therapy courses with more postgraduate and honours students who are undertaking research; there is a demand from the universities that staff publish; senior clinicians are required to undertake research and publish, and there is demand for evidence-based practice. In summary occupational therapists are becoming more research active, and our journal is highly regarded by profession. However, it is not all good news. The page allocation during this time has fluctuated, from 380 pages in 2005, 352 pages in 2006, to 304 pages in 2007. This of course results in a backlog of manuscripts waiting to be published, notwithstanding the rejection of manuscripts which do not reach the required standard. The Editorial Board of the AOTJ recently made some very hard decisions to develop strategies to overcome the backlog of manuscripts. The Board decided to discontinue the ‘In Practice’ section of the journal—any new manuscripts for this section will be returned to the authors, suggesting that they consider submitting to OT Connections, or to another journal. Thesis Abstracts will no longer be published, Book Reviews and Critically Appraised Papers will be reduced to two per issue. In addition, the Board decided to reduce the word limit for all research and feature manuscripts to 4000 (from 5000), reduce the number of figures and tables to four per manuscript and reduce the number of references to 25 (from 60) per manuscript. These strategies should result in the equivalent of an extra 100+ pages of available space per volume, which will enable more manuscripts to be published. At the current submission rate, it is proposed that from 2009 AOTJ be published as six 80-page issues per year. Online Early is to be retained, which offers the opportunity for authors to cite their publications prior to their publication in the hard copy of the journal. In addition to these strategies the Council of OT AUSTRALIA has been asked to consider the importance of the AOTJ to the profession and allocate more funding to publish a special supplement in 2007 and six large issues in 2008 to overcome the balance of the backlog of manuscripts. Good times should lie ahead for our Journal, we have an innovative Editorial Board; the journal has been a leader internationally with Critically Appraised Papers and Online Early and was in fact the first occupational therapy journal to be published online. It is important that this fine reputation is maintained and that the journal is respected and supported by the professional body under whose banner it exists.

플래시 메모리 파일 시스템을 위한 순수도 기반 페이지 할당 기법에 대한 연구

플래시 메모리 파일 시스템을 위한 순수도 기반 페이지 할당 기법에 대한 연구

Data space mapping for efficient I/O in large multi-dimensional databases

In this paper, we propose data space mapping techniques for storage and retrieval in multi-dimensional databases on multi-disk architectures. We identify the important factors for an efficient multi-disk searching of multi-dimensional data and develop secondary storage organization and retrieval techniques that directly address these factors. We especially focus on high dimensional data, where none of the current approaches are effective. In contrast to the current declustering techniques, storage techniques in this paper consider both inter- and intra-disk organization of the data. The data space is first partitioned into buckets, then the buckets are declustered to multiple disks while they are clustered in each disk. The queries are executed through bucket identification techniques that locate the pages. One of the partitioning techniques we discuss is especially practical for high dimensional data, and our disk and page allocation techniques are optimal with respect to number of I/O accesses and seek times. We provide experimental results that support our claims on two real high dimensional datasets.

Moving address translation closer to memory in distributed shared-memory multiprocessors

To support a global virtual memory space, an architecture must translate virtual addresses dynamically. In current processors, the translation is done in a TLB (translation lookaside buffer), before or in parallel with the first-level cache access. As processor technology improves at a rapid pace and the working sets of new applications grow insatiably, the latency and bandwidth demands on the TLB are difficult to meet, especially in multiprocessor systems, which run larger applications and are plagued by the TLB consistency problem. We describe and compare five options for virtual address translation in the context of distributed shared memory (DSM) multiprocessors, including CC-NUMAs (cache-coherent non-uniform memory access architectures) and COMAs (cache only memory access architectures). In CC-NUMAs, moving the TLB to shared memory is a bad idea because page placement, migration, and replication are all constrained by the virtual page address, which greatly affects processor node access locality. In the context of COMAs, the allocation of pages to processor nodes is not as critical because memory blocks can dynamically migrate and replicate freely among nodes. As the address translation is done deeper in the memory hierarchy, the frequency of translations drops because of the filtering effect. We also observe that the TLB is very effective when it is merged with the shared-memory, because of the sharing and prefetching effects and because there is no need to maintain TLB consistency. Even if the effectiveness of the TLB merged with the shared memory is very high, we also show that the TLB can be removed in a system with address translation done in memory because the frequency of translations is very low.

Exciting developments and a new image

We are delighted to have a new look journal from 2005. However, the new cover is not the only change; we have more pages and a new layout. I would like to thank the Editorial Board of AOTJ, Katrina Lok (Production Editor) and also designers from Blackwell Publishing who have worked together to develop and produce these new designs and changes. When you look at the new layout inside the journal, you will see many differences. We were aiming to increase the number of words per page so that we can publish more articles per issue and volume. This has been achieved without making the font too small to read. We have also reduced the size of the headings and the top and bottom margins. The position of the abstract and author details has changed, as well as the style of font for the headings and subheadings; all these innovations have saved space. The cover design continues to use the official OT AUSTRALIA colours; we aim to maintain the clarity and accuracy of these colours in the production process. We wanted to have a simple design, which would promote the profession and also stand out on library and occupational therapy department displays. We have the OT AUSTRALIA logo featured as a watermark on the front cover, with the full logo on the back cover. The contents list of the journal issue is maintained on the back cover for easy access. Finally, we have more pages. I wish to thank the Council of OT AUSTRALIA for supporting the Journal's application for more pages per volume. The number of pages has increased from 56 to 92 per issue from 224 to 368 pages per year or volume. This can also be interpreted as an increase of approximately 14 additional manuscripts published per year, which will enable the backlog of manuscripts to be cleared more rapidly. In addition, we will maintain the OnlineEarly facility whereby manuscripts that have been accepted and are ready for publication are published as soon as possible online, prior to allocation in the print version of the journal. It is important to understand that these articles can be cited and used, as they are in fact officially published with a DOI reference number. More details about OnlineEarly are available on the website http://www.blackwellpublishing.com/aot. When visiting the website, go to the page that lists the issues, click the OnlineEarly logo and then read the article ‘What is OnlineEarly?’ All articles published online are complete and searchable articles in HTML, with active links, and there is a PDF with journal-style formatted pages for download if that is preferred. In the HTML online version, you can click on highlighted references that will take you through the links to the reference details and then either to Crossref, Medline or the ISI Abstract. In some cases, you can link right through to the full text reference. I have written in detail about the online facility because it is important for members to try it out and use it! If you have forgotten your pin number, there is a section on the website where you can email Blackwell Publishing so that you can be issued with a new number. While in Canada during late 2004, I received many compliments about our Journal and the many recent innovations — particularly the Critically Appraised Papers section and the online facility, including OnlineEarly. I thank the Journal Board and OT AUSTRALIA for their support of these innovations — we are truly becoming a more international publication. The extra page allocation has enabled the publication of a wide variety of articles in this issue; we hope that the breadth of topics will mean that most readers will find something of interest in their specialty. In addition to the regular departments we have also been able to include a review article on occupational therapy off-road driver assessment tools and a very interesting paediatric case study. We anticipate that we will overcome the backlog of manuscripts during the year and therefore encourage you to submit your research and other articles so we can maintain the variety and high standard of our journal.

Editorial

Editorial

Editorial

Editorial

Faster Web page allocation with neural networks

To maintain quality of service, some heavily trafficked Web sites use multiple servers, which share information through a shared file system or data space. The Andrews file system (AFS) and distributed file system (DFS), for example, can facilitate this sharing. In other sites, each server might have its own independent file system. Although scheduling algorithms for traditional distributed systems do not address the special needs of Web server clusters well, a significant evolution in the computational approach to artificial intelligence and cognitive engineering shows promise for Web request scheduling. Not only is this transformation - from discrete symbolic reasoning to massively parallel and connectionist neural modeling - of compelling scientific interest, but also of considerable practical value. Our novel application of connectionist neural modeling to map Web page requests to Web server caches maximizes hit ratio while load balancing among caches. In particular, we have developed a new learning algorithm for fast Web page allocation on a server using the self-organizing properties of the neural network (NN).

The publishing cycle

As I prepare this editorial during the last week of January, I am aware that it will not be published and therefore read until June 2002. It is remarkable to think that when this issue is distributed some of you will be fortunate enough to be on your way to Stockholm for the 13th World Congress of Occupational Therapists. I plan to take the opportunity to attend a meeting of international occupational therapy journal editors while there. This meeting will provide an opportunity for the editors to discuss the common issues relevant to maintaining a high quality professional journal. One of the major problems we are encountering here in Australia is the time taken for manuscripts to be published. This results in frustration for authors and also the Editorial Board, and means that, at times, published research has become less current. Part of the problem is due to the fact that we have such enthusiastic and prolific authors and, because of high costs of publishing, a limited page allocation per issue. This year, due to increased membership, we have gained an extra eight pages per issue, equivalent to one feature article. Because of the demand for space, we are currently investigating the feasibility of publishing more issues per year. It is important for readers to know why there is such a long lead-time when publishing a quarterly journal. As the journal is a peer-reviewed publication, this contributes to the length of time from submission to acceptance for publication. The process is as follows. Following submission, as Editor I select two referees for anonymous peer review of the article, these referees then have six weeks allocated to review the manuscript and return their comments to the editorial office. We have found over the last 12 months that the referees have been very efficient, with few reminders being required. If minor or major changes are recommended, the manuscript is then returned to the author with the referees’ reports. Manuscripts which are not of the required standard or relevance are also returned to the author with a covering letter. It is at the revision stage that delays appear to be occurring, with authors taking a long time to return their amended manuscripts. If a major rewrite is required, then the manuscript is reviewed again by the referees, and the process is repeated until the paper is accepted for publication. There are many stages involved in the editorial process once a manuscript has been accepted for publication, this is the Production Process. For example, for this current June issue, we began planning in December 2001. Articles were selected based on a variety of criteria, and sent to the Production Editor by mid January, any queries were then emailed to authors by early February with replies required within 10 days. The authors then received proofs by email in late March, with corrections being required by early April. The Synergy online publication date for the June issue was 23 May, and the June issue was dispatched to subscribers at the end of May. As can be seen, the production process takes time and a great deal of hard work by the editorial staff and the authors. There are many reasons for the length of time taken for manuscripts to reach the publication stage. However, with all the due processes in place and carefully adhered to, it ensures that we have a high quality and well-respected journal. As occupational therapists have now become more involved in research, the dissemination of their findings becomes essential: to provide evidence, for the sharing of information and to generate further areas of research. Through presentation of research at professional meetings and conferences, and publication in professional journals we will increase the occupational therapy research profile and add to the evidence base of our profession. In addition to research findings, theoretical and feature articles, we have other sections in our Journal, such as book and media reviews, ‘In Practice’ and ‘Viewpoint’, which, we hope, provide a wide range of topics to stimulate the interest of occupational therapists in their many areas of practice.

Power aware page allocation

One of the major challenges of post-PC computing is the need to reduce energy consumption, thereby extending the lifetime of the batteries that power these mobile devices. Memory is a particularly important target for efforts to improve energy efficiency. Memory technology is becoming available that offers power management features such as the ability to put individual chips in any one of several different power modes. In this paper we explore the interaction of page placement with static and dynamic hardware policies to exploit these emerging hardware features. In particular, we consider page allocation policies that can be employed by an informed operating system to complement the hardware power management strategies. We perform experiments using two complementary simulation environments: a trace-driven simulator with workload traces that are representative of mobile computing and an execution-driven simulator with a detailed processor/memory model and a more memory-intensive set of benchmarks (SPEC2000). Our results make a compelling case for a cooperative hardware/software approach for exploiting power-aware memory, with down to as little as 45% of the Energy• Delay for the best static policy and 1% to 20% of the Energy• Delay for a traditional full-power memory.

Power aware page allocation

One of the major challenges of post-PC computing is the need to reduce energy consumption, thereby extending the lifetime of the batteries that power these mobile devices. Memory is a particularly important target for efforts to improve energy efficiency. Memory technology is becoming available that offers power management features such as the ability to put individual chips in any one of several different power modes. In this paper we explore the interaction of page placement with static and dynamic hardware policies to exploit these emerging hardware features. In particular, we consider page allocation policies that can be employed by an informed operating system to complement the hardware power management strategies. We perform experiments using two complementary simulation environments: a trace-driven simulator with workload traces that are representative of mobile computing and an execution-driven simulator with a detailed processor/memory model and a more memory-intensive set of benchmarks (SPEC2000). Our results make a compelling case for a cooperative hardware/software approach for exploiting power-aware memory, with down to as little as 45% of the Energy• Delay for the best static policy and 1% to 20% of the Energy• Delay for a traditional full-power memory.

Sharing objects in a distributed, single address space environment

With reference to an object type defining the two basic operations, read and write, we present solutions to the object sharing problem, classified according to the migration and/or replication of the shared objects. We refer to a memory management system supporting a single address space view of storage in a distributed environment. Our system defines a small, powerful set of primitives that allow processes to explicitly control the allocation and deletion of the virtual pages in the physical storage as well as the page movements across the memory hierarchy. By using real programs, we demonstrate that these primitives make it possible to integrate an object sharing algorithm within the implementation of the given object at little programming effort. The discussion takes a number of salient issues into consideration, including network costs and a conceptual framework for actual implementation of the memory management system.

Power aware page allocation

One of the major challenges of post-PC computing is the need to reduce energy consumption, thereby extending the lifetime of the batteries that power these mobile devies. Memory is a particularly important target for efforts to improve energy efficiency. Memory technology is becoming available that offers power management features such as the ability to put individual chips in any one of several different power modes. In this paper we explore the interaction of page placement with static and dynamic hardware policies to exploit these emerging hardware features. In particular, we consider page allocation policies that can be employed by an informed operating system to complement the hardware power management strategies. We perform experiments using two complementary simulation environments: a trace-driven simulator with workload traces that are representative of mobile computing and an execution-driven simulator with a detailed processor/memory model and a more memory-intensive set of benchmarks (SPEC2000). Our results make a compelling case for a cooperative hardware/software approach for exploiting power-aware memory, with down to as little as 45% of the Energy Delay for the best static policy and 1% to 20% of the Energy Delay for a traditional full-power memory.