Insufficient Memory Research Articles

Deployment of Embedded Web Services in Real-Time Systems

Service-oriented architectures (SOA) become more and more important in networked embedded systems. The main advantages of Service-oriented architectures are a higher abstraction level and interoperability of devices. In this area, Web services have become an important standard for communication between devices. However, this upcoming technology is only available on devices with sufficient resources. Therefore, embedded devices are often excluded from the deployment of Web services due to a lack of computing power, insufficient memory space and limited bandwidth. Furthermore, embedded devices often require real-time capabilities for communication and process control. This paper presents a new table driven approach to handle real-time capable Web services communication, on embedded hardware through the Devices Profile for Web Services.

Compiling for memory emergency

There has been a continued growth in the sales of mobile and embedded devices, in spite of the economic recession in many parts of the world. Many of these devices operate under tight memory bounds. Past research dealt with this problem and proposed both hardware and software solutions oriented toward reducing memory space requirements of embedded and mobile applications. One of the common characteristics of most of these prior efforts is that they assume the memory space available to an embedded application is fixed for the entire execution. Unfortunately, this is not a valid assumption in many execution environments since a typical embedded/mobile platform can have multiple applications executing concurrently and sharing a common memory space. As a result, the amount of memory available to a particular application can vary during execution. A particularly interesting scenario is what we call "memory emergency", where the size of the memory available to an application suddenly drops. If the application is not written to cope with this emergency scenario, the result would normally be a premature termination due to insufficient memory. In this paper, we propose compiler-based solutions to this memory emergency problem. The proposed compiler support modifies a given application code assuming a memory emergency model and reduces memory space demand (when necessary) by recomputing data values, thereby performing a tradeoff between memory space reduction and performance overhead. Our goal is to be able to work with the reduced memory space but minimize the performance overhead it brings. We evaluate the proposed approaches using twelve array-based embedded benchmarks. Our experimental analysis shows that the proposed approaches are very successful in responding to many memory emergency scenarios.

Dynamic Reconfiguration of IP Domain Middleware Stacks to Support Multicast Multimedia Distribution in a Heterogeneous Environment

Seamless connectivity to multiple wireless networks independently of a fixed point is becoming increasingly important for mobile devices however wireless networks differ in bandwidth, size and access costs each requiring protocol functions to enable devices to communicate efficiently. In addition, due to the divergence of users and applications, traditional stacks are frequently enriched with additional functionality such as transport protocol functionality, synchronisation and presentation coding which can lead to a performance bottleneck due to the insufficient processing power and memory of portable devices. We argue that an extensible middleware is needed to cover small resource constrained devices to full-fledged desktop computers thus we investigate dynamic micro-protocols which enable devices to adopt specific protocol stacks at runtime in an attempt to optimise the stack to the functionality that is actually required by the application thus eliminating additional overhead functionality provided by generic stacks. A side effect of this is that it allows devices such as PDAs to offer protocol functions, which would not normally be available due to their memory constraints. Memory constrained devices are catered for through the deployment of a client-proxy overlay network where proxies offload processing. The problem of the ‘common denominator bandwidth’ is overcome through multicast media groups where clients subscribe to different quality of services in accordance with resource availability and move between groups according to bandwidth availability over time. Our end result is a Java middleware for multimedia streaming to heterogeneous mobile clients, utilising dynamic configuration of protocols with respect to application requirements and available network resources. Performance is increased through application specific tailored protocols and reducing protocol complexity allows stacks to fit inside the limited memory space of current mobile devices. We evaluate the dynamic reconfigurability of the middleware and present some new results from a series of applications utilising runtime adaptation.

Assessment of agitation in elderly patients with dementia: correlations between informant rating and direct observation.

Assessment of behavior problems in elderly persons with dementia is important for understanding and managing those behaviors. The most common method for assessing agitation is the use of informant ratings; however, these ratings may be affected by staff bias, inaccurate or insufficient memory, or stress. An alternative method is direct observation, which is more objective, but very costly and necessitates time sampling, thereby limiting the period covered by the assessment. To date, little research attention has been given to the degree to which these two methods converge. In the present study, 175 elderly persons with dementia who manifested problem behaviors were recruited from 11 nursing home facilities in Maryland. The average age for the participants was 87 years; 78% were female. Two methods were employed for assessing agitation: the Agitated Behaviors Mapping Instrument (ABMI), which is based upon direct observations, and the Cohen-Mansfield Agitation Inventory (CMAI), which is a frequency rating scale completed by a formal caregiver. The ABMI and CMAI contain some identical items for tapping behavior problems. Data analysis revealed significant Pearson correlations between identical items on the two assessment instruments, as well as significant correlations of summary measures based on these different instruments, demonstrating a strong convergence between informant ratings and direct observations. Informant ratings can achieve moderate agreement with direct observation when valid instruments and informants are used.

Improving effective bandwidth through compiler enhancement of global cache reuse

The performance of modern machines is increasingly limited by insufficient memory bandwidth. One way to alleviate this bandwidth limitation for a given program is to minimize the aggregate data volume the program transfers from memory. In this article we present compiler strategies for accomplishing this minimization. Following a discussion of the underlying causes of bandwidth limitations, we present a two-step strategy to exploit global cache reuse—the temporal reuse across the whole program and the spatial reuse across the entire data set used in that program. In the first step, we fuse computation on the same data using a technique called reuse-based loop fusion to integrate loops with different control structures. We prove that optimal fusion for bandwidth is NP-hard and we explore the limitations of computation fusion using perfect program information. In the second step, we group data used by the same computation through the technique of affinity-based data regrouping, which intermixes the storage assignments of program data elements at different granularities. We show that the method is compile-time optimal and can be used on array and structure data. We prove that two extensions—partial and dynamic data regrouping—are NP-hard problems. Finally, we describe our compiler implementation and experiments demonstrating that the new global strategy, on average, reduces memory traffic by over 40% and improves execution speed by over 60% on two high-end workstations.

An End to End Adaptable Architecture for Streaming Media over IP Networks

Wireless networks differ in bandwidth, size and access costs each requiring a set of protocol functions to enable devices to communicate efficiently. Portable multimedia devices such as PDA's and laptops will also vary greatly however all these devices will require optimal multimedia delivery. A traditional method is for sources to limit their transmission rates to accommodate lower bandwidth links, even though high-bandwidth connectivity is available to many participants. This method similar to others does not provide optimum throughput to heterogeneous clients due to its quest for a common denominator bandwidth. In addition, due to the divergence of users and applications, traditional protocol stacks are frequently enriched with additional functionality such as transport protocol functionality, synchronization and presentation coding which can lead to a performance bottleneck due to the insufficient processing power and memory of portable devices. Micro-protocols attempt to eradicate this bottleneck by optimising the protocol stack to the functionality that is actually required by the application. A side effect of this is that it allows a device such as a PDA to offer protocol functions, which would not normally be available due to its memory constraints achievable by downloading necessary micro-protocols for new environments and discarding previous micro-protocols. Multicast media groups overcome the heterogeneous client problem where clients subscribe to different quality of services in accordance with resource availability and move between groups according to bandwidth availability. Chameleon is 100% Java middleware for multimedia streaming to heterogeneous mobile clients, which allows the dynamic configuration of protocols with respect to application requirements and available network resources. We evaluate the dynamic reconfigurability of the middleware in order to demonstrate runtime adaptation. We especially concentrate on the primary quality transformation technique (PQT) of the middleware which enables clients to subscribe to media groups in accordance with available resources and network capacity.

Communication effects on memory and judgment

AbstractThe present research explored the mechanisms of judgmental and memory correction for communication influence. Participants described a target person to an addressee who either liked or disliked the target. Participants' descriptions were more positive when the addressee liked the target than when the addressee disliked the target. Mediation analyses showed that the effect of the addressee's attitude on judgment and memory was mediated by its effect on participants' descriptions. Participants' evaluative judgments of the target were influenced by the addressee's attitude only when the attitude was presented subtly. When the presentation of the attitude was blatant, participants used the attitude as a situational inducement of their communication behavior and corrected their judgments. Both participants' recall and false recognition of personality traits were evaluatively consistent with the addressee's attitude regardless of its presentation. However, mediational analyses suggested that participants in the blatant presentation condition engaged in an insufficient memory correction by discounting the evaluative implications of their descriptions. Copyright © 2002 John Wiley & Sons, Ltd.

Solving Some Large Scale Semidefinite Programs via the Conjugate Residual Method

Most current implementations of interior-point methods for semidefinite programming use a direct method to solve the Schur complement equation (SCE) $M \Delta y=h$ in computing the search direction. When the number of constraints is large, the problem of having insufficient memory to store M can be avoided if an iterative method is used instead. Numerical experiments have shown that the conjugate residual (CR) method typically takes a huge number of steps to generate a high accuracy solution. On the other hand, it is difficult to incorporate traditional preconditioners into the SCE, except for block diagonal preconditioners. We decompose the SCE into a 2 × 2 block system by decomposing $\Delta y$ (similarly for h) into two orthogonal components with one lying in a certain subspace that is determined from the structure of M. Numerical experiments on semidefinite programming problems arising from the Lovász $\theta$-function of graphs and MAXCUT problems show that high accuracy solutions can be obtained with a moderate number of CR steps using the proposed equation.

Design of a multiple leaky buckets shaper

Asynchronous Transfer Mode (ATM) has been considered as the future trend for the information transfer in Broadband Integrated Services Digital Network (B-ISDN). ATM networks provide a guaranteed Quality of Service (QoS) for a variety of applications. One of the important QoS attributes is Cell Delay Variation (CDV). CDV measures how much delay variation will appear in cell transmission times. Call Admission Control (CAC) and Usage Parameter Control (UPC) are two main functions for achieving traffic control in ATM networks. The UPC scheme monitors and controls traffic in order to determine whether the established connections conform to the predefined parameters or not. For a real source operating at predefined parameters, some delay variation will occur in cell transmission times. CDV Tolerance (CDVT) controls the amount of acceptable deviation. Two classes of UPC methods have been proposed in the past to deal with CDV: one is to design an UPC method considering the worst traffic at User-Network Interface (UNI); the other is using a shaper for each connection of each link to accommodate CDV. In this paper, we propose a new shaper architecture to implement the UPC function that will eliminate CDV introduced between Terminal Equipment (TE) and the edge node of public networks. Our new shaper can follow multiple Generic Cell Rate Algorithm (GCRA) rules and consider all the traffic parameters. Furthermore, we use a simple method to reduce the memory requirement in the shaper. Insufficient memory might lead to incorrect cell departure in previous shapers. Simulation results show the performance is quite good and as expected.

Abstract interpretation using typed decision graphs

This article presents a way of implementing abstract interpretations that can be very efficient. The improvement lies in the use of a symbolic representation of boolean functions called typed decision graphs (TDGs), a refinement of binary decision diagrams. A general procedure for using this representation in abstract interpretation is given; we examine in particular the possibility of encoding higher-order functions into TDGs. Moreover, this representation is used to design a widening operator based on the size of the objects represented, so that abstract interpretations will not fail due to insufficient memory. This approach is illustrated on strictness analysis of higher-order functions, showing a great increase in efficiency.

Adding Temporary Memory to ZCS

In a recent article, Wilson (1994) described a "zeroth-level" classifier system (ZCS). ZCS employs a reinforcement learning technique comparable to Q-learning (Watkins, 1989). This article presents results from the first reconstruction of ZCS. Having replicated Wilson's results, we extend ZCS in a manner suggested by Wilson: The original formulation of ZCS has no memory mechanisms, but Wilson (1994b) suggested how internal "temporary memory" registers could be added. We show results from adding one-bit and two-bit memory registers to ZCS. Our results demonstrate that ZCS can exploit memory facilities efficiently in non-Markov environments. We also show that the memoryless ZCS can converge on near-optimal stochastic solutions in non-Markov environments. We then present results from trials using ZCS in Markov environments that require increasingly long chains of actions before reward is received. Our results indicate that inaccurate overgeneral classifiers can interact with the classifier-generation mechanisms to cause catastrophic breakdowns in overall system performance. Basing classifier fitness on accuracy may alleviate this problem. We conclude that the memory mechanism in its current form is unlikely to scale well for situations requiring large amounts of temporary memory. Nevertheless, the ability to find stochastic solutions when there is insufficient memory might offset this problem somewhat.

SUPRAPAT: A value added patent database — A personal view and vision

A short look back to the beginnings of patent documentation in industry shows that the austerity of data available in the old patent databases of the 1960s and 1970s (because of insufficient memory and disk space) still has effects on the multitude of present databases. Millions of patent records added each year may lead to frustration of many users when confronted with rising costs and search times. The resulting non-use of some bases is dangerous for patent departments and leads to diminishing revenues for database producers. This gives rise to the proposal of one and only one comprehensive logical (not necessarily physical) patent database containing extended data. Many of them are available in one or the other of the existing databases, but not all together in one file. EPO, Derwent and CA are regarded as centers of competence for bibliographic data collecting and correcting; abstracting; indexing and deep analysis, respectively.

Photonic Bandgaps: What is the Best Numerical Representation of Periodic Structures?

Abstract The numerical analysis of photonic bandgaps in periodic dielectric structures leads to an infinite-dimensional eigenvalue problem which must be truncated to be solved. The truncation alters the numerical representation of the dielectric structure. By changing the formulation of the problem, the representation of the periodic structure and the convergence of the solutions can be improved. Our calculations have shown that convergence can be reached for one- and two-dimensional structures but insufficient computer memory has prevented us from reaching convergence for three-dimensional structures. High-order super-Gaussians can be used to estimate photonic bandgaps accurately and to overcome the problems caused by the poor numerical representation of step functions. However, results obtained with super-Gaussians in one- and two-dimensional structures converge to the same values as those obtained with step functions; hence we conclude that the expansion of the dielectric function with step functions y...

Graph-based techniques to speed up floorplan area optimization

A well known approach for the floorplan area optimization problem is to first determine a list of all non-redundant implementations of the entire floorplan and then select an optimal floorplan from the list [1–7]. For large floorplans, this approach may fail due to insufficient memory space available to store the implementations of sub-floorplans generated during the computation. An effective method to reduce memory usage is as follows: During the computation, whenever the set of non-redundant implementations of a sub-floorplan exceeds a certain predefined size, we only retain a subset of the implementations that can best approximate the original set. In this paper, we present two algorithms to optimally select implementations for rectangular and L-shaped sub-floorplans. Our algorithms can be applied to existing floorplan area optimization algorithms such as [3–7] to improve their performances. The common key idea of our two algorithms is to reduce the problem of optimally selecting a subset of implementations to a constrained shortest path problem, which we can solve optimally in polynomial time. We have incorporated the two algorithms into [7] and obtained very encouraging experimental results.

The enable construct for exception handling in Fortran 90

It is apparent that a good mechanism for handling computing exceptions is essential for the construction of robust and maintainable Fortran 90 code. Ideally such code would execute rapidly when all is well and not fail in situations such as when data is near the arithmetic limits or there is insufficient memory for an automatic array. A language feature is needed that permits transfer of control when necessary from straightforward fast code to slower code that uses an alternative algorithm.

Solving Large and Sparse Linear Equations in Analog Circuit Simulation on a Cluster of Workstations

For developing analog circuits at Siemens AG, the method of Harmonic Balance was integrated into the circuit simulator TITAN. The equations of Harmonic Balance are solved by Newton's method. The corresponding Jacobian is a very large sparse block matrix. Hitherto this kind of problem was computed either on a single workstation or on a vector computer. In this paper, we present an algorithm that makes it possible to simulate analog circuits on a cluster of workstations. Since the bulk of data is stored in a distributed way, even very large problems which often cannot be processed on vector computers due to insufficient memory capacity may now be solved.

Diagnosing the Computer User Addicted, Infected or Technophiliac

This paper explores how computer culture is altering the subjectivity of computer users and culture more broadly. It foregrounds the way language and metaphors of the human body and social body alter with computers and how this affects people's experience of themselves. The mechanical self of plumbing and tubes and sparking on all four cylinders has given way to the hardwired self, with serious identity crises about its insufficient memory banks. The first half of the paper makes use of the metaphor of diagnosis to think about various subjective relations to computers. Addiction, infection and technophilia are elaborated as exciting as well as distributing in their subjective and wider political meanings. The second half of the paper locates computer users historically, using a concept of technology as embodying and perpetuating social relations - including cultural, aesthetic, sexual, economic and political relations. In particular, issues of control, memory and difference are developed and explored.

A fast multilayer general area router for MCM designs

An efficient multi layer general area router is developed as an alternative to the 3-D maze router for solving the multi-layer multichip module (MCM) routing problem. The router, named SLICE, is independent of net ordering, requires much shorter computation time, and uses fewer vias. A key step is to compute a maximum noncrossing bipartite matching, which is solved optimally in O(n log n) time where n is the number of possible connections. The router was tested on a number of examples, including two MCM designs from MCC. The total wirelength used by SLICE is only a few percent away from the optimal on average. Compared with a 3-D maze router, SLICE is six times faster and uses 29% fewer vias. Another feature of SLICE is that it works on only a 'thin slice' of a two-layer routing grid at a time, so it can produce solutions for large MCM routing examples where 3-D maze routers fail due to insufficient memory.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A differential compiler for computer animation

A program for the real-time display of computer animation on a bit-mapped raster display is presented. The differential compiler performs temporal domain image data compression using frame replenishment coding on successive frames of animation stored in memory as bitmaps and saves only the differences. A small run-time interpreter then retrieves and displays the differences in real-time to create the animated effect. This results in a significant reduction in storage requirements, and allows animation on general purpose computers which would otherwise be too slow or have insufficient memory. Frame creation is both device and method independent. An animation environment supports interactive editing capabilities, reconstructing any arbitrary desired frame for later modification. Frames can be added, modified, or deleted, and the animated sequence can be viewed at any point during the session. The compiler is automatically called as needed; its operation is transparent to the user. The compiler is described in detail, both in terms of data compression and the requirements of interactive animation editing.