Search

The article presents information related to several issues of the journal "IEEE Software," to be published this year. The January-February issue will focus on aspect oriented programming related to software development. The March-April issue will focus on innovative methods and techniques emerging from research to support the practice of software architecture. It will also presents insights for research to improve the state of practice in the architecting of software-intensive systems. The May-June issue will discuss practices and experiences in producing quality requirements that lead to successful systems, products and services. The July-August issue will discuss the importance of testing in software development. The September-October issue will focus on case studies from successful and failed global projects as well as experiences with methods, processes and tools developed to overcome these challenges. The November-December issue will discuss the software development and maintenance challenges faced by small and medium-size software companies and will presents practical solutions and best practices to such companies.

The article discusses some of the works of Victor R. Basili, a professor at University of Maryland, related to empirical software engineering. Basili has mostly contributed in the field of measuring software development processes and products. The Goal Question Metric (GQM) approach developed by Basili made software measurement achievable. The GQM approach is based on the assumption that an organization must first specify goals for itself and its projects to measure its products and processes usefully. This approach has helped many organizations such as Schlumberger Ltd., and others in both guiding improvement initiatives and developing metrics for managing software projects. Basili has also developed QIP, which uses the principles of continuous improvement approaches such as Total Quality Management to software development. QIP is a cyclical process for planning organizational improvement on the basis of various factors including understanding the current baseline, setting quantifiable goals for improvement, choosing development processes to meet those goals and others.

The article discusses aspect-oriented programming (AOP) technologies aimed to improve software system's modularity. AOP aims to improve system modularity by modularizing crosscutting concerns which includes error handling or transaction code, interacting features and reliability and security. Using an AOP language, a developer can define an update signalling aspect, a module that expresses code that crosscuts the structure and functionality of a system. When needed aspects can be composed into the system. This modularization allows independent application of the UpdateSignalling aspect or implementing different updating strategies. Many AOP languages support selecting join points and modifying behavior at those join points from the aspect. Using this approach, the developer of code that will later be advised by the aspect can remain oblivious of crosscutting concerns. Aspect-oriented mechanisms were first developed for source code. The difficulties associated with nonmodularized crosscutting structure are not limited to implementation but also occur during requirements and design activities.

The article presents information about portable C/C++ code for portable XML data. XML makes use of context encapsulation to separate content from presentation and to support a hierarchical structure among data from various sources. XML data is reusable, easily derivable, and reconfigurable. However, XML-based applications need a programming technology to perform tasks such as parsing, manipulating and validating the data. Software application developers commonly use toolkits based on C/C++, such as Gnome's XML C parser and toolkit, Oracle's XML Developer's Kit for C/C++, and Microsoft's XML Parser. XML application developers often use a two or three-tiered architecture to facilitate application development. They implement the application logic using XSL stylesheets and the lower-level components using a scripting language such as Python, Tcl, PHP, or Perl. Ch is an embeddable C/C++ interpreter for cross-platform scripting, shell programming, numerical computing, network computing, and embedded scripting. The Ch XML for these toolkits is designed to integrate portable XML data with portable C/C++ scripting code.

The article presents information on portability of a project's non-source code asset. The portability of a project's non-source code assets means more than allowing those assets to move freely between different organizations and developers. Portability also means that a marketplace for tools can evolve without the artificial restrictions of the vendor lock-ins imposed by incompatible data formats and associated switching costs. Such an environment would let different tools compete on their actual technical merits, without the artificially cozy protection of their installed base's captive audience. One should interchange format and let the tools compete of stability, usability, performance, and data manipulation capabilities. Databases, compilers, and Web browsers have all flourished under a regime of standardized interfaces. Nowadays, a software system's program source code makes up only a small part of its assets stored as bits-taking up a larger percentage are specifications, design diagrams, application server deployment scripts, build rules, version history, documentation, and regression tests.

The article presents information on the author's opinion on the research papers on magnetic recording discussed at the 16th Annual Magnetic Recording Conference held at Stanford University, Stanford, California from August 15-17, 2005.

Practical level performance for ∼200 Gb/in² has been verified by AlOx barrier tunneling magnetoresistive (TMR) heads, which resistance area product (RA) is more than 3 ohm.μm2, in perpendicular recording mode. In addition, improved AlOx barrier magnetic tunnel junctions (MTJs) formed on plated bottom shield with smoothed surface achieved TMR ratio of 25% and 16% with RA of 1.9 and 1.0 ohm.μm2, respectively, indicating over 200 Gb/in² is also possible by the AlOx barrier TMR heads with lower RA. Furthermore, TMR heads with crystalline MgO barrier were fabricated. The MgO barrier MTJs formed on plated bottom shield with smoothed surface achieved TMR ratio of 88% with RA of 2.0 ohmμm2, which is 3.5 times higher than that of AlOx barrier MTJ5 under similar RA. Dynamic electrical test was also performed for TMR heads with the MgO barrier. As a result, good readback waveform with huge output was obtained. This is the first confirmation of readback waveform generated from TMR heads with crystalline MgO barrier. Our results indicate that the future of TMR heads technology is promising beyond 200 Gb/in² application. [ABSTRACT FROM AUTHOR]

Tunneling magnetoresistive (TMR) reading heads at an areal density of 80-100 Gbit/in² in a longitudinal magnetic recording mode have for the first time been commercialized for both laptop and desktop Seagate hard disk drive products. The first generation TMR products utilized a bottom TMR stack and an abutted hard bias design. These TMR heads have demonstrated three times the amplitude of comparable giant magnetoresistive (GMR) devices, resulting in a 0.6 decade bit error rate gain over GMR. This has enabled high component and drive yields. Due to the improved thermal dissipation of current-perpendicular-to-plane geometry, TMR runs cooler and has better lifetime performance, and has demonstrated the similar electrical static discharge robustness as GMR. TMR has demonstrated equivalent or better process and wafer yields compared to GMR. The TMR heads is proven to be a mature and capable reader technology. Using the same TMR head design in conjunction with perpendicular recording, 274 Gbitfin2 has been demonstrated. Advanced design can reach 311 Gbit/in² [ABSTRACT FROM AUTHOR]

Inserting a 4 A-Mg metal layer between the amorphous CoFeB bottom electrode layer and the MgO barrier layer was found to be effective in realizing huge magnetoresistance effect in low-resistance CoFeB/MgO/CoFeB magnetic tunnel junctions (MTJs). As a result, magnetoresistance (MR) ratio as high as 138% at resistance-area product (RA) of about 2.4 1Ω /μm² was obtained. This value is about seven times that of state-of-the-art MTJs for magnetic sensor application. X-ray diffraction analysis clarified that crystal orientation of the poly-crystalline MgO(001) barrier layer was improved by the Mg layer. It is suggested that the higher crystalline orientation of the MgO(001) barrier layer could have enhanced the coherent tunneling of Δ1 electrons, resulting in an increase of MR ratio at the low RA (thin Mgo thickness) region. The annealing temperature and free layer materials have also been optimized to satisfy the requirements for practical read head application. Although this optimization resulted in a reduction in the MR ratio to about 45%-53%, this value is still more than twice the highest MR ratio of conventional MTJs. The currently developed fabrication process will accelerate the development of highly sensitive read heads for ultrahigh-density hard-disk drives. [ABSTRACT FROM AUTHOR]

In magnetic data storage the noise and signal-to-noise ratio (SNR) of the read heads is becoming more and more of a limiting factor given the smaller readout devices needed to follow the ever higher areal densities. For higher data rates this problem is aggravated by the needed wider signal bandwidths. This study presents the magnetic simulation of noise, SNR, and bandwidth for the most common types of head: tunneling magnetoresistive heads, CIP-GMR heads, and CPP-GMR heads. Such simulations allow the comparison of these different types of head. The sensitivity factors of the SNR for the most important head parameters are determined as well. The SNR of the total read path is also determined, including the head-amplifier interconnect and the electronics. [ABSTRACT FROM AUTHOR]

This paper considers magnetization fluctuations (mag-noise) in current-perpendicular-to-plane (CPP) read sensors, due to either thermal fluctuations or spin-torque induced instability. The former are relevant to both tunneling (TMR) and giant-magnetoresistance (GMR) devices, the latter a potentially more serious limitation most relevant to all-metal GMR sensors. Discussion of thermal mag-noise is here treated by micromagnetic simulation. Discussion of spin-torque instability includes analytical modeling combined with experimental measurements. [ABSTRACT FROM AUTHOR]

A new type of dual spin valve (DSV) structure with enhanced giant magnetoresistance (GMR) ratio is proposed, and the performance is characterized by the prototype read heads. The bottom part of DSV is kept as a synthetic pinned structure and only the top part adopts a monolayer pinned structure. The top monolaycr pin DSV (TM-DSV) showed three percentage points higher GMR ratio and 20% higher ΔR value. After the mechanical lapping process, the monolayer pinned layer can be stabilized by not only sense current but also large stress induced anisotropy due to Villari effect. After the quantitative analysis, the induced stress is measured to be about 4.2 x 109 N/m2. The bias point was tuned by adjustment of Cu spacer layer thickness using the RKKY interaction between pinned layer and free layer. There is no reliability concern related to the monolayer pinned structure. In a perpendicular magnetic recording system, TM.DSV technology could successfully demonstrate the output of 35.5 mV/μm and∼ 170 Gbits/in2 density feasibility. [ABSTRACT FROM AUTHOR]

We investigated the relationship between head design, pole-tip magnetization, head field, and media magnetization in longitudinal recording. At first we estimate the effective switching field of current mobile media with isotropic anisotropy, and introduce longitudinal corresponding field for the head field assessment. The smaller perpendicular component or large perpendicular field gradient of the head field at the upper pole improves the transition quality, but the small perpendicular field degrades the overwrite performance. So the ideal head field is large longitudinal and perpendicular field for good overwrite and fast attenuation with the large perpendicular field gradient for the transition quality. Some head designs make it possible to control the ratio of the perpendicular component to the longitudinal head field, by changing the pole-tip magnetization direction. As a project to control it, we discuss the effect of the bottom and upper poles configuration on the transition, overwrite, and adjacent track erasure in longitudinal recording. [ABSTRACT FROM AUTHOR]

Pulse deposition of 2.4 T Co37Fe63 alloy in photoresist features with 40 nm critical dimension and high aspect ratio is presented. The design of the pulse deposition parameters is described in terms of the transport limitations through the diffusion layer, electrochemical interface stability with respect to Fe(OH)3 precipitation, and the optimum conditions for additive (Saccharin) adsorption. The alloy grain size and crystal structure in the nanoconfined electrode geometry is compared versus the thin film and relevant implications for magnetic recording are discussed. The 2.4 T Co38Fe61Pd1 alloy is introduced as a possible way to improve the corrosion properties of 2.4 T Co37Fe63 alloy. [ABSTRACT FROM AUTHOR]

Excellent overwrite (OVW) performance has to be guaranteed for two cases, namely, the case where a low frequency (LF) background data pattern is overwritten by a high frequency (HF) data pattern and the reverse case, e.g., the reverse overwrite (ROVW) situation where a high linear density background pattern is overwritten by a low linear density data pattern. In this work, the spectra of both OVW and ROVW processes have been investigated experimentally to identify the intrinsic characteristics of residual noise in the OVW and ROVW processes in perpendicular recording. Specifically the spectra of both OVW and ROVW processes have been quantitatively analyzed to quantify the characteristics of relevant amplitude modulation or amplitude cancellation process. [ABSTRACT FROM AUTHOR]

A novel spin-stand method for measuring the perpendicular write pole "footprint" is proposed. This method allows the detailed study of the spatio-temporal pole footprint dynamics and the two-dimensional media saturation and noise. Experimental footprint measurements are presented and the problems caused by the magnetic nonuniformity of the poles are addressed. The perpendicular overwrite process is studied by time-domain waveform analysis. The spatial distribution of the overwrite signal is determined by the geometry and the magnetic uniformity of the write pole footprint. The dependence of recording parameters on write current overshoot in perpendic- ular recording is also analyzed. [ABSTRACT FROM AUTHOR]

We demonstrate a technique to measure the field of a perpendicular head on a soft underlayer. We built a current-perpendicular-to- the-plane giant magnetoresistive sensor deposited on a soft underlayer and placed it on the write-read contact tester. Scanning a perpendicular head over the device with sub-nm resolution allowed us to map the field spatially and by changing the writer current, obtain saturation curves of both the main pole and the return pole. [ABSTRACT FROM AUTHOR]

This paper reports on progress in perpendicular drive integration. In the first part, the performance of a series of perpendicular media with different coercivity (5-6 kOe), nucleation field, and soft-underlayer thickness (120/80 nm) is compared using two different head designs. Good overwrite and bit error rate performance is achieved for all head/media combinations. Higher coercivity leads to lower linear density but higher track density, resulting in approximately equal areal density capability. The thinner soft undelayer causes wider erase bands and therefore a loss of areal density. Wide-area track erasure measurements were performed on all media. The media with lower coercivity do not show any measurable amplitude loss, due to their larger absolute value of the nucleation field and the smaller write current needed to write on them. Two types of single-head, 7200 RPM desktop perpendicular drives (40 and 80 GB capacity) were built using 95 mm perpendicular media. They both achieve good bit error rate and off-track capability in all zones. The thermal decay for the 40 GB drive was measured and compared to a similar longitudinal drive. It was found that the thermal decay rate is much smaller in the case of the perpendicular drive in the range -6 °C to +80 °C. [ABSTRACT FROM AUTHOR]

We present dibit extraction with a nonlinear representation of the readback signal in terms of a finite Volterra series. We show that the extracted dibit response is the linear combination of the cyclically shifted Volterra kernels. Then we explore several implications of this to dibit extraction and system characterization such as the indistinguishability of physical mechanisms from output observation, bandwidth expansion, and requirements on the length of the pseudorandom binary sequences that arise when viewing the system in the wider context of the finite Volterra series. [ABSTRACT FROM AUTHOR]

In magnetic recording systems, it has been reported that sufficient signal-to-noise ratio (SNR) was obtained with a narrow track width, using a spin-valve head and metal-evaporated tape. In order to apply this to a helical-scan tape system, techniques for writing narrow tracks and reading them exactly were required. However, the degree of mechanical accuracy demanded was too high to realize. With this in mind, we have developed multichannel heads which can write and read certain tracks with a single scan, as well as a nontracking technique removing the need for such precise mechanical accuracy. A prototype tape drive employing these techniques was developed, and a track density exceeding 15 ktpi was achieved. [ABSTRACT FROM AUTHOR]

The term "data recovery" herein refers to accessing logically and/or physically damaged storage media, for which no functioning backup exists. The state-of-the-art physical techniques for recovering data from failed hardware can all be described as "part replacement." To achieve high data density and high manufacturing yields, modern drives are "hyper-tuned" in the factory so that their data layout, zone frequencies, and various channel settings are optimized for each head, surface, and zone. This greatly complicates part replacement because a transplanted headstack, for example, no longer matches the servo, preamp, and read channel parameters that were optimized for the original headstack. Methods and challenges are discussed for replacing, or "refreshing," firmware and system area information and for replacing all of the drive's electronics. The data recovery industry's point of view, limitations of current techniques, and some probable future directions in data recovery are also presented. It is predicted that data recovery will be more important in the future as drives are exposed to more extreme mobile environments. Drive manufacturers may be able to differentiate themselves from their competition by designing for recoverability. [ABSTRACT FROM AUTHOR]

In this paper, we consider the problem of modeling the timing error process in magnetic recording systems. We propose a discrete- valued Markov model for the timing error process, and design two methods (data-aided and nondata-aided), based on the Baum-Welch algorithm, to extract the model parameters from the readback waveforms. The channel model we consider is an intersymbol interference (ISI) channel with additive Gaussian noise. The continuous-time readback signal at the output of the channel is sampled at baud-rate. Simulation results show that the estimated parameters are close to the actual values and the convergence is attained in a few iterations of the Baum-Welch algorithm. We also demonstrate the usefulness of the accurate model extraction by comparing a fine-tuned Markov timing recovery loop to the standard Mueller and Muller detector with a tuned second-order loop filter. [ABSTRACT FROM AUTHOR]

We have three main goals in this paper. First, we will show the equivalences and relationships among some of the well-known classes of structured low density parity check (LDPC) codes. Second, we consider application of these codes to magnetic recording channel. We will make performance comparisons not only with bit-error-rate (BER) metric but also with sector failure rate (SFR) as the comparison dimension. Such a comparison is more meaningful from a system analysis perspective. Finally, we consider the case of these structured LDPC codes concatenated with conventional Reed-Solomon (RS) codes. We believe LDPC codes will not completely replace RS codes in the future storage devices. From this perspective, for LDPC codes to be contenders for future architectures, they will need to work in harmony with the outer RS codes. This implies that we have to take into account the redundancy introduced by the overall coding scheme, that is, RS code in concatenation with LDPC code, to reach meaningful comparisons from a system design perspective. [ABSTRACT FROM AUTHOR]

High-rate low-density parity-check (LDPC) codes are the focus of intense research in magnetic recording because, when decoded by the iterative sum-product algorithm, they show decoding performance close to the Shannon capacity. However, cycles, especially short cycles, are harmful to LDPC codes. The paper describes the partition-and-shift LDPC (PS-LDPC) codes, a new class of regular, structured LDPC codes that can be designed with large girth and arbitrary large minimum distance. Large girth leads to more efficient iterative decoding and codes with better error-floor properties than random LDPC codes. PS-LDPC codes can be designed for any desired column weight and with flexible code rates. The paper details the girth and distance properties of the codes and their systematic construction and presents analytical and simulation performance results that show that, in the high signal-to-noise ratio region, PS-LDPC codes out- perform random codes, alleviating the error floor phenomenon. [ABSTRACT FROM AUTHOR]

A parity code and a distance enhancing constrained code are often concatenated with a Reed-Solomon code to form a coding system for magnetic recording. The tensor-product parity coding scheme helps to improve efficiency of the parity code while retaining the same level of performance. In this paper, we present two methods for combining a tensor-product parity code with a distance-enhancing constrained code. The first method incorporates a constrained code with unconstrained positions. The second method uses a new technique, which we call word-set partitioning, to achieve a higher code rate relative to the first method. We simulate the performance of several coding systems based upon the two combination methods on a perpendicular recording channel, and we compare their symbol error rates and sector error rates with those of a system that uses only a Reed-Solomon code. [ABSTRACT FROM AUTHOR]

The maximum transition run (MTR) and running digital sum (RDS) constraints have a wide variety of applications in magnetic recording systems. The property of an MTR constraint is to eliminate long runs of transitions so that certain common error events can be eliminated. The property of an RDS constraint is desirable for achieving high-density recording for recording channels exhibiting low frequency components, such as perpendicular recording channels. It is very common to implement more than one constraint in practice to achieve the desired effect in the recording channel. In this paper, the labeled graph representations of RDS and MTR [or the time-varying MTR (TMTR)] constraints are combined to obtain new classes of constraints: RDS-MTR and RDS-TMTR constraints. The capacity of a combined constraint is slightly lower than that of its constituent constraints. The spectra of RDS-MTR and RDS-TMTR constraints show the suppression of low frequency contents compared to the conventional MTR and TMTR constraints. A method for designing block codes is proposed to implement the properties of RDS, MTR (or TMTR), the run-length-limited (RLL) (0, k), and twin constraints. The cardinalities of block codes are maximized over various situations of codeword boundaries. The code design method- ology is demonstrated via a rate 100/108 RDS-TMTR block code based on a rate 20/21 RDS-TMTR mother code. [ABSTRACT FROM AUTHOR]

Iterative codes such as low density parity check (LDPC) codes and turbo product codes (TPC) are of interest for data storage appli- çations. To apply these codes to digital recording systems, two possible schemes are considered in this paper. In the first scheme, a single LDPC code of column weight j > 2 is used to replace the Reed-Solomon (RS) code of the conventional channel. In the second scheme, an iterative code is used as the inner code and is concatenated with an outer RS code. Single parity check TPC code is considered for this scheme. We developed a high-throughput field programmable gate array (FPGA) platform to evaluate the error floor performance of LDPC codes and the error statistics of TPC codes in partial response (PR) channel with turbo equalization. High rate codes (rate 8/9 for LDPC code and rate 0.935 for TPC code) are evaluated for magnetic recoding application. The bit error rate (BER) performance of LDPC code down to 1010can be reached within 2 h using the FPGA platform. The TPC code error statistics are evaluated using about 1011 bits at different signal-to-noise (SNR) levels. For practical implementation complexity and power consumption, 2 channel iterations and 2 TPC decoder iterations are employed. The results show that the gain by applying TPC code under 2 channel iterations and 2 TPC decoder iterations is marginal. [ABSTRACT FROM AUTHOR]

We have studied the behavior of failure in lifetime acceleration test of TMR prototype with performance meeting 100 Gb/in2 recording. The failure mode is identified to be same as extrinsic breakdown of the AlOx barrier, while failure due to magnetic change is not observed. The extrinsic breakdown is further found to be the increase in pinhole conduction in the barrier, involving thermal energy from ambient and read current. Results also suggest the existence of a critical voltage below which pinhole is not enlarged. Furthermore, we will discuss our hypothesis in the formation of pinholes and the role of charge traps in the barrier. Finally, we will discuss a method for screening out heads with relatively short lifetime, which is feasible for mass production. The screening allows a guarantee of lifetime well enough for drive application. [ABSTRACT FROM AUTHOR]

Through an electrostatic discharge (ESD) damage case study of giant magnetoresistive (GMR) head in shunting process of head gimbal assemblies (HGA), this paper unveils that GMR heads would be damaged by a combination of field-induced voltage on HGA and surrounding trigger event. This investigation further demonstrates that the ESD protection technique, which derives from metal-to-metal contact to GMR terminals and thus focuses on eliminating metal-to-metal contact to GMR terminals, is not sufficient to prevent ESD. The electromagnetic interference caused by nearby trigger event needs to be noted also in the ESD control. [ABSTRACT FROM AUTHOR]

To achieve the area! density goal in hard disk drives of 1 Tb/in2 the minimum physical spacing or flying height (FH) between the read/write transducer and disk must be reduced to about 2 nm. At such low spacing new nanoscale forces act between the slider and disk, such as intermolecular and electrostatic forces, which must be taken into consideration in the air bearing design. These forces increase the level of flying height modulations (FHMs), which in turn creates dynamic instability and intermittent contact in the flying head slider similar to what has been observed in experiments. Here, we present three possible approaches to minimize such forces and/or reduce FHM by FH control, including a micro-trailing pad slider, a thermal flying height control slider, and a piezoelectric flying height control slider for hard disk drives. [ABSTRACT FROM AUTHOR]

In a hard disk drive with a dual-stage actuator the standard rotary actuation of the voice coil motor (VCM) is combined with an additional micro- or milli-actuation to accomplish high-bandwidth and highly accurate track following. For extremely high density recording, the servo control algorithm designed for a specific dual-stage actuator needs to maintain high performance track following in lieu of external disturbances and uncertainties in the dynamic response induced by product manufacturing tolerances. Essential in the robust design of a servo control algorithm is to bound these uncertainties and design a robust track following servos accordingly. In this paper, we present an experiment based methodology to characterize and bound the uncertainties in the dynamic response of a dual-stage actuator in the form of an uncertainty model. Subsequently, the uncertainty model is used for the analysis and synthesis of a robust dual-stage servo controller that is able to optimize position error variance in the presence of uncertainty information. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel shockproof head gimbals assembly (HGA) structure incorporated with "a balanced-type head suspension" for 0.85-in hard disk drives (HDD5), and investigated its performance on external shock resistance during read/write operations. Experimental drop impact test results confirm that the 0.85-in drive equipped with the new suspension structure is durable over 2000 G (1.0 ms duration) of external shock, ensuring no head slider and/or disk surface damage. According to computer simulation and experimental results, potential shock resistance of more than 2000 G may also be possible with the head actuator itself, which points to its application in cellular phones. [ABSTRACT FROM AUTHOR]

The use of scattering parameters (S-parameters) to characterize high frequency interconnects has widely been used in the signal integrity field and is now being accepted into the disk drive arena as the preferred modeling tool for write and read path interconnects. The motivating factor for the shift from traditional lumped circuit models to S-parameter models has been the increase in data rates which exceed 1 GB/s. S-parameter models are convenient because they can be used directly in several high end spice circuit simulators as "black box" models. This paper will describe recent measurement techniques which make vector (VNA) network analyzer measurements possible for various interconnect structures up to 8.5 GHz. This paper will verify the results using time domain refiectometry (TDR) measurements of the same interconnect structures. Specific examples on signal losses, impedance, and coupling will be shown. Advances in achieving a broad range of differential impedances while maintaining high bandwidth will be demonstrated with idealized test structures. [ABSTRACT FROM AUTHOR]

The article briefly discusses research papers that are contained within the section, titled "Algorithm Design and Scheduling Techniques (Realistic Platform Models) for Heterogeneous Clusters," of this issue. The purpose of these papers is to explore the impact of platform heterogeneity on the design and evaluation of new algorithms, such as, distributed algorithms and scheduling algorithms. One paper by Ligang He and his colleagues discusses two new workload allocation strategies for large-scale heterogeneous platforms. The paper "On the Distribution of Sequential Jobs in Random Brokering for Heterogeneous Computational Grids," by Vandy Berten and colleagues, focuses on resource brokering for scheduling sequential jobs onto a grid platform. In their article, "Multiple Job Scheduling in a Connection-Limited Data Parallel System," authors Alessandro Amoroso and Keith Marzullo describe a new method for scheduling jobs in a distributed system.