There is no exact definition for the term ‘virtual clusters’ in the literature; however, the definition of Virtual Breading Environments (VBEs) is very close to what it is referred to as virtual cluster in this paper. In addition, a number of authors including Galeano & Molina (2005) have used these two terms interchangeably. Camarinha-Matos & Afsarmanesh (2005) define VBEs as “an association or pool of organizations and their related supporting institutions that have both the potential and will to cooperate with each other through the establishment of a “base” long-term cooperation agreement and interoperable infrastructure.” One of the main goals of a virtual cluster is to assimilate all of the benefits of a geographical cluster; this includes the formation of business collaborations. However, apart from this purpose, a virtual cluster should be able to satisfy the needs of its members as a GC does in terms of easy access to suppliers, employees, and customers. Innovation should be achieved through alliances between firms and knowledge flow within the cluster. The following are some examples of VBEs that are currently operational.
28 Read more
Starting from the historical perspective of virtual clusters in Grid computing, we see the new architectural model for high performance virtual clusters illustrated in Figure 1.3 that is implied by the efforts described in this dissertation. Here, we leverage commodity hardware, as well as some advanced HPC hardware withint he same system. Providing advanced hardware in this scenario has a twofold effect. First, we look to provide classic distributed memory HPC applications in an virtualized environment that still utilize the same hardware that such applications have grown accustomed to. Second, it allows for novel big data analytics and cloud platform services to move towards hardware that may drastically speed up their computational efforts. Specifically, we allow for such platform services to leverage the high-speed, low-latency interconnects that HPC systems have relied upon. While such cloud systems may not be able to take full advantage of the latency improvements, it is estimated the the increased bandwidth may have a significant impact on the overall runtime of these systems, reducing the overall time-to-solution. While this notion could incorporate a wide array of differing technologies, we focus here on GPU-based accelerators and InfiniBand interconnects in conjunction with x86 CPUs.
230 Read more
At SDSC, we have created a novel framework and infras- tructure by providing virtual HPC clusters to projects us- ing the NSF sponsored Comet supercomputer. Managing virtual clusters on Comet is similar to managing a bare- metal cluster in terms of processes and tools that are em- ployed. This is beneficial because such processes and tools are familiar to cluster administrators. Unlike platforms like AWS, Comet’s virtualization capability supports installing VMs from ISOs (i.e., a CD-ROM or DVD image) or via an isolated management VLAN (PXE). At the same time, we’re helping projects take advantage of VMs by providing an enhanced client tool for interaction with our management system called Cloudmesh client. Cloudmesh client can also be used to manage virtual machines on OpenStack, AWS, and Azure.
Interactive cloud offerings are expanding, providing virtual com- puting laboratories, remote desktop environments and online col- laboration tools. For example, North Carolina State University’s Virtual Computing Laboratory (VCL)  is a production cloud system hosting virtual desktops with a variety of applications for more than 13,000 students at NCSU and other nearby schools. These new platforms bring individual users easy access to popular applications/tools with low management overhead. Such systems also yield significant residual, or unused, resources, due to overpro- visioning and the bursty, unpredictable nature of interactive work- loads. Traditional techniques such as virtual machine (VM) pack- ing are unlikely to be performed aggressively in this environment, due to users’ bursty resource consumption patterns combined with response time requirements. Conservative workload consolidation, on the other hand, will likely leave significant amounts of resid- ual resources idle, as we show in §3.1. By aggressively harnessing such resources, cloud providers will benefit from higher cloud uti- lization as well as considerable energy savings, as the incremental
13 Read more
For the command-line and programmable interface, we support long running instances after execution of the user specified application. This advanced resource control is designed for applications requiring multiple-rounds of calculation, and supports HBase as persistent storage. Furthermore, the amount of running instances of existing virtual clusters can be adjusted by adding/removing VM nodes with the programmable interface cgl.salsa.salsadpi.plugin.AddANodeToExistingIaaSCluster. The user or program developer needs to provide the cluster name (normally master node hostname), master node hostname and virtual network private IP address, a shorter salsaDPI configuration file (excluding the application object), and amount of add/remove instance(s). Figure 8 shows a salsaDPI configuration file for adding HBase regionserver nodes to an existing cluster which installs HBase and Ganglia when starting the instance.
18 Read more
The presented approach for virtual clusters creation from Linux containers was found to be ef- ﬁcient and its performance was proven comparable to that of an ordinary physical cluster: not only the use of containers does not incur processor virtualization overheads, but also the network virtual- ization overheads can be totally removed if the host’s network name space is used and the network bandwidth is saved by automatically transferring only those ﬁles that are needed through network- mounted ﬁle system rather than the whole images. From the point of view of a system administrator, storing each HPC application in its own container makes versioning and dependencies control easily manageable and their conﬁguration does not interfere with the conﬁguration of the host machines and other containers.
12 Read more
This paper presents virtual clusters based proportional fairness and resource allocation scheme for Localized Single Carrier Frequency Division Multiple Access (L-SC-FDMA). L-SC-FDMA has been selected as the uplink transmission scheme in 3GPP Long Term Evolution (LTE) due to its low Peak to Average Power Ra- tio (PAPR) over OFDMA in general and high rate-sum capacity over Interleaved SC-FDMA in particular. Virtual cluster-based proportional fairness (VCPF) scheduler exploits the link adaptation information avail- able at MAC layer to form virtual clusters. The distributed proportional fairness scheduler ensures a mini- mum throughput for all users in the coverage area by assigning contiguous RBs, proportional to the throughput and the number of users in a particular cluster or group. Simulations have been performed using practical scenario of uniformly distributed users in Rayleigh faded coverage area and design formulas have been devised for network planning to get the best possible fairness with promising level of quality of service (QoS).
Our experiment results indicate that Hadoop systems provi- sioned by bare metal servers with powerful storage options can be better options to build a high performance virtual clusters for processing various workloads with a cost consideration. The result of our experiments delivers a thorough analysis of production environments with extensive research on storage options, i.e. block storage and local NVMe. Our results for Hadoop workloads on Amazon, Google Oracle, and Microsoft expose underlying hardware requirements e.g. IOPS, along with service limitations e.g. throughput allowance per instance. High storage performance made a significant impact on HDFS based jobs with a large number of virtual cores. JBOD- style (’just a bunch of disks’) non-RAID storage attachment shows 6 times better results with additional volume counts and CPUs per server than large numbers of low-end servers according to our result in Fig 1. Data intensive workloads, for that reason, may gain better scalability and efficiency on high capacity servers and bare metal servers are suitable for seeking performance improvements and cost savings.
Visual Cryptography(VC), an emerging technology for secret sharing which allows visual information (pictures, text, etc.) to be encrypted in such a way that the decryption can be performed by the human vision system(HVS).Originally it was proposed by Naor and Shamir in 1994 for black and white images. Later this technique is extended for gray level images as well as for color images. This paper compares and analyze the performance of various VCS on various parameters such as pixel expansion, contrast, shares generated etc. The compared algorithms came into aura by rectifying limitations of one another.
The advantage of the Virtual Move Monte Carlo cluster algorithm over sim- ilar cluster algorithms is that the cluster is selected dynamically according to the pairwise energy gradients. This allows us to guarantee a physically more realistic topology of the moving clusters, and to approximate the kinetic evolution of the system by Monte Carlo simulation in situations where collective modes of motion cannot be neglected. The derivation and implementation of the algorithm involves novel and non-trivial concepts in the design of Monte Carlo schemes, which we be- lieve were not sufficiently described in the previous papers. The first part of this thesis thus provides a concise description of the simulation method, along with the formal proof of its validity. It is then shown that the maximum displacement size δ significantly affects the cluster size distribution and the efficiency of the algorithm. The moving clusters tend to span the entire system for large δ, and the cluster motion is negligible for small δ. The choice of the optimum δ may depend on the properties of the system and the aims of the simulation. In an equilibrium fluid, the moving cluster size distribution may follow a universal exponential behaviour. In the simulation of phase separation, the moving cluster size distribution develops over time towards larger clusters, and intermediate sized rearrangements become rare. It is proposed that a slow lowering of δ at those stages is likely to enhance the efficiency of the simulation. In the simulation of stabilized isolated aggregates, the cluster generation and acceptance probabilities also evolve in time for different sizes of the moving clusters. It is suggested that a slow increase of δ during the course of the simulation may enhance the sampling in the multiscale modelling of hierarchical self-assembly. Interestingly, simulations of kinetically slowed down phase separation led to the same product, independently of the choice of δ, and in spite of entirely different trajectories leading to the final kinetically arrested states. This suggests that the products of VMMC simulation may have universal significance, and may be physical.
175 Read more
Dankzij de kennis die is opgedaan tijdens de Enermet periode had de VCS Group al een vrij goed inzicht in alles wat zich bij de energiebedrijven afspeelde. Door die kennis te koppelen met ICT dienstverlening kon de VCS Group inspringen op de nieuwe mogelijkheden binnen de energiesector. De VCS Group heeft dan ook bij veel energiebedrijven een ondersteunende rol gespeeld om van de ‘monopolistische’ structuur (één groot energiebedrijf dat alle taken uitvoert) naar de huidige vrije markt structuur (taken opgesplitst tussen meerdere bedrijven) van het te komen, al dan niet gekoppeld met een implementatie van “Generis” (een ICT-oplossing voor administratieve bedrijfsprocessen door Process Vision ontwikkeld).
92 Read more
4.4. Computational software service HeartFlows Blood flow through aortic valve can be described by numerical models based on viscous incompressible Navier-Stokes equations . The developed software service HeartFlows based on Ansys Fluent platform is applied for numerical analysis of low flow, low pressure gradient aortic stenosis, which is a highly challenging condition in terms of diagnosis and therapeutic management . Computational flow analysis is a representative example of Ansys Fluent usage in various fields of engineering . Advanced software and large computational times have been associated with the numerical analysis of complex flows , . Thus, cloud infrastructure and software services are perceived as a promising avenue for future advances in this multidisciplinary area of medical engineering. Large computations can be carried out on clusters of virtual machines acquired on-demand and managed by the Eucalyptus Cluster Controller (Fig. 2) while the interactive exploration of computational results can be performed by using the developed software services on a single virtual machine saving computational resources. Moreover, semi-automatic medical image segmentation, which requires user interaction, can also be performed on virtual machines employing interactive graphics services to obtain patient-specific geometry.
Individuals inside the virtual societies are involved in all kinds of social and economic activities. The theory of structural holes already tells us that the agent occupies special position in social network will have great inﬂuence on his economic performance . Here, we empirically investigate the underlying connections between network structure and economic output for individuals in virtual societies. Our data are game logs from a popular MMORPG in China (see data description in Section ). For each society, we con- struct a dependence network of individuals, which only keeps the multiscale backbone of original friendship network (see construction of dependence network in Section ). Fig- ure illustrates the dependence network from a virtual society and some basic statistics about the topological structure. One can see that there are lots of in-stars in dependence network, which is in accordance with the ﬁndings  that the in-degree distribution is much fatter than out-degree distribution. Such phenomena can be explained by that in virtual societies the individuals with high levels may play a relatively important role in friend lists of low level individuals, which is supported by the strong preference of con- necting high level individuals for low level individuals .
14 Read more
TBX5-dependent activation of an Scn5a enhancer. We hypothesized that TBX5 directly drives a molecular network required for fast VCS conduction via direct regulation of Gja5 and Scn5a, and focused our efforts on Scn5a, as no direct regulators of Scn5a in the VCS have been identified; Gja5 is a known TBX5 target in the embryo (11). We bioinformatically interrogated the Scn5a locus to identify potential TBX5-responsive enhancers using the overlap of 4 independent data sets: (a) evolutionary conservation (21); (b) ChIP-seq studies identifying TBX5 binding sites in the atrial cardiomyocyte HL-1 cell line (22); (c) p300 ChIP-seq peaks to mark active enhancers, both in vitro and in vivo (22, 23); and (d) bioinformatic predictions of cardiac enhancers (24). Within the genomic region including the Scn5a locus and complete upstream and downstream intergenic regions, a single region approximately 15 kb downstream of Scn5a demonstrated overlap in all 4 data sets (Figure 6A). We hypothesized that this genomic region represents a TBX5-dependent Scn5a enhancer.
11 Read more
China’s VC market was established in the mid-1980s with foreign VCs initially dominating the market. Facilitated by the relaxed regulations on domestic institutional investors investing in VC/PE funds and creation of additional stock market exit routes (SME and Venture Boards), domestic VCs has grown rapidly and now dominated China’s VC market (Lu et al., 2013). In the US where formal institutions are mature, VCs can employ traditional VC mechanisms and reduce agency costs by carefully screening, evaluating and providing assistance to their portfolio firms (Amit et al., 1998). In China, however, intellectual property and investor protection, government intervention, IPO markets for exits are still matters of tremendous concerns to VC investors (Jeng and Wells, 2000; Lu et al., 2013). Under such circumstances, institutional theorists highlight the importance of social network ties in entrepreneurial activities (e.g., Xin and Pearce, 1996; Peng and Luo, 2000). Compared to other types of social network ties (e.g., inter-firm and interpersonal ties), the value of PTs should be greater in China, since firms continue to depend on the government for critical information and resources despite the progress in its marketization process (Shi et al., 2014).
185 Read more
A given yield VC can be set apart as inaccessible on the off chance that it gets demands from any information VCs, as one of these solicitations will result in a given in . The consequences of yield side mediation are then disseminated again to sources of info VCs. Since a specified information VC just subjects solicitations to solitary yield port, it is sheltered to consolidate the approaching stipend data from all directions of router utilizing a P port as input OR gate; the result is a bit vector that shows the allowed yield VC—assuming any—at the chose goal port. The yield first use variety is depicted in Fig. 4. Here, every data VC sends requesting to all cheerful yield VCs at the picked objective port, where intercession is executed as in the data first case. Since decision among yield VCs isn't performed until after yield side affirmation, we can speak to VCs that are difficult to reach just by covering their yield specialists' permit signals. At the data side, we can join the honors from the assorted yield ports using a P-input OR as in the data first case. In any case, as a given data VC may get stipends from different yield VCs at its objective port, an additional intercession sort out is relied upon to pick a champ among them. Since yield VCs that were distributed must be separate as blocked off, the last data side permit signals must be multiplied back to the yield side; thusly, while yield first bit reduces the delay for creating the data side give signals, it assembles the deferral give, out for enrolling the give signals for individual yield VCs. A wave front-based VC allocator, as appeared in Fig. 4, involves an acknowledged P×V - input wave front allocator, with additional method of reasoning for making the P×V – wide requesting vector for every data VC as in the particular yield first case
A visual cryptography scheme (VCS) is a kind of secret image sharing scheme which allows the encoding of a secret image into shares. The drawback of visual cryptography scheme (VCS) is that even a layman to cryptography is able to decode the secret image without having any cryptographic knowledge and computational tools/devices. An extended visual cryptography scheme (EVCS) is a type of visual cryptography scheme (VCS) which consists of meaningful shares (not like in the traditional VCS which consists of meaningful as well as meaningless shares). In this paper, we propose the construction of EVCS which is recognized by embedding random shares into meaningful covering shares, and hence we call it the Embedded Extended Visual Cryptography Scheme (EEVCS). Experimental results in recent years show that the proposed Embedded EVCS has reasonably good visual quality compared to many of the well-known EVCSs.
Abstract—It is challenging to execute an application in a heterogeneous cloud cluster, which consists of multiple types of virtual machines with different performance capabilities and prices. This paper aims to mitigate this challenge by proposing a scheduling mechanism to optimise the execution of Bag- of-Task jobs on a heterogeneous cloud cluster. The proposed scheduler considers two approaches to select suitable cloud resources for executing a user application while satisfying pre- defined Service Level Objectives (SLOs) both in terms of exe- cution deadline and minimising monetary cost. Additionally, a mechanism for dynamic re-assignment of jobs during execution is presented to resolve potential violation of SLOs.
A. Performance: This is one among the difficulties of CC, which should be viewed as and when characterization of VM assignment systems is done. From client’s perspective execution is absolute first part of cloud benefit. Cloud specialist co-ops essential thought process is to accomplish 100%performance; however, it is difficult to achieve 100%. Since performance is impact by system, simultaneous ward forms, VM relocation and so on. S. Shirinbab et al.  presented a unique methodology of the over- allotment where the physical CPU centers are treated instead as vCPU centers. It gives us more Virtual Machine arrangement contrasted with firmly time-sharing Virtual Machines. Henceforth, the execution of progressively parallel assignment is done to improve the execution. MB Nagpure et al.  shared a dynamic resource distribution dependent on dissymmetry. The dissymmetry calculation estimates irregularity in condition, additionally it anticipates the load of future by using calculation of Fast Up and Slow Down (FUSD). The above framework if utilized properly can be used to present the challenges of security in SDN.
17 Read more
virtual e-business clusters and increasing the limit num- ber of start-ups. Second, for virtual e-business clusters with a certain scale, there is a limit to the conduction of entrepreneurial behavior among enterprises, which is in line with what has been observed in practice. Not all en- terprises can always adopt an innovative development strategy in their development. There are always some enterprises Will decline or gradually withdraw from the market. Therefore, the growth and development of vir- tual e-business cluster has a certain threshold. Thirdly, more attention should be paid to the passive impact mode in the cooperative relations between enterprises when promoting the conduction of entrepreneurial be- havior of enterprises in virtual e-business cluster. That is to say, while encouraging and guiding the start-up enter- prises to find suitable e-business related enterprises for