Abstract—Value-addedservices (VASs) are an integral part of todays Content Delivery Networks (CDNs). They can be implemented as a chain of Virtual Network Functions (VNFs). These chains need to be placed in an efﬁcient way in CDNs in order to optimize quality of service (QoS) for end-users(EUs) while minimizing cost for providers. We formulate the problem as an Integer Linear Program (ILP) and propose a Pattern-based Cost-efﬁcient Proactive VNF placement and chaining (PCPV) algorithm. The objective is to ﬁnd the optimal number, location and chaining of VNFs in such a manner that the cost is minimized while QoS is met. Apart from cost minimization, the support for large-scale CDNs with a large number of physical machines (servers) and EUs is an important feature of the proposed algorithm. Through simulations, the algorithm behavior for small-scale to large scale CDN networks is analyzed.
Ehsan Ahvar is with Université Rennes, INRIA, CNRS, IRISA, France
advertisement overlays/tickers and content adaptation (e.g., transcoding, compression). Website/application acceleration relies on a combination of optimization techniques and allows end-users (mostly enterprises) to access a website/application with improved response time. CDN analytics services provide awareness at various levels including network, device and content, thereby enabling content providers to make more informed and business-friendly decisions. Increasingly, mul- timedia content providers are finding it useful to outsource their media-related services such as digital rights management and content adaptation to the CDN provider domain. This is because providing such services requires the content provider to store customized content for every end-user. The over- lay/ticker service is perceived as a fast monetization strategy by content providers with, for instance local news/weather updates inserted on the top of the video as static or scrolling tickers. Similarly, short video advertisements are linearly overlaid on the video delivered to end-users. Traditional CDNs face numerous obstacles towards efficiently provisioning value- addedservices. First, the value-addedservices are deployed on a dedicated hardware in the CDN infrastructure. As a result, it is both time-consuming and expensive to deploy and manage the service, resulting in more time to market and cost- inefficiency. Second, the explosive growth in end-users and amount of content  delivered to them raise the need to scale the deployed services as needed. Third, end-users’ growing interest in customized content fueled by a continual innovation in video formats over the years requires rapid provisioning of novel video-based value-addedservices. Network Functions Virtualization (NFV)   is an emerging paradigm that can be used to make CDNs meet the above-mentioned require- ments. NFV is a novel telecommunication service provisioning approach in which the network function is decoupled from the physical devices on which they run and are implemented as virtualized software, termed as virtual network functions (VNFs) which run on top of a virtualized infrastructure and chained together to provide a required service. They can be implemented on any computational node (e.g., CDN surrogate server, switches, data center) that meets their resource demand. The computational nodes must provide Network Functions Virtualization Infrastructure (NFVI) functions to support the execution environment of VNFs. These nodes are referred to as NFVI nodes. It should be noted that NFV has been tradi- tionally used to virtualize middle-boxes (e.g., firewall, Network
2.4. Conclusion 20
Therefore, these kinds of applications should be treated near the geographical location of their origin to achieve lower latency and higher throughput.
The importance of Cloud brokering for telecommunication services is highlighted by Carella G. et al. [ CMCS12 ]. In this approach, the Cloud broker enhances his placement mechanisms based on: real-time data on network performance, QoS requirements and Cloud providers’ prices. The goal is to provide to telecommunication service operators a minimum QoS to satisfy customer’s requirements by monitoring the deployed services. This approach is evaluated in a testbed composed of a Cloud broker and an IP Multime- dia Subsystem (IMS) deployment. The cost-effective placement of Web 2.0 applications with high-availability and fault-tolerance requirements across multiple Cloud providers is proposed by Frincu et al. [ FC11 ]. In this approach, authors consider applications con- sisting of several components and connectors (C/Cs). C/Cs are reallocated by making a snapshot, stopping the execution of each C/C, moving the snapshot to a new VM and starting the C/C from the snapshot. A Cloud broker architecture with the intelligence to react to changes in business processes by changing the Cloud configuration across multiple Cloud providers is described by Grivas et al. [ GKW10 ].
2. Evolution of modern logistics introduces new concepts (such as JIT, QR, ECR, and value-addedservices), which require the adapted transport system to support themselves. The characteristics of intermodal transport can meet these needs, fundamentally soldering intermodal logistics. The advantages (door–to-door, transport cost, environmental impacts, reliability and safety, value-addedservices, advanced information services, etc) of intermodal transport offer the enterprises the opportunities to build up an efficient, economic, and reliable logistical system to respond to the challenges of globalization and internationalization.
Our main result may appear quite surprising: Common value information can be handled without costs, as if the information was not private. The optimal concession contract specifies auctions rules biased in favor of the incumbent, so as to preserve incentives to invest; the analysis is still preliminary, but we are able to show that the optimal contract can be implemented by introducing a golden parachute, to be given to the incumbent whenever it is replaced by an outsider. Our results are thus quite similar to those in Laffont-Tirole (1993, Chapter 8), who consider a similar model but only under private values. These results are however not trivial as the existence of a common value destroys the results that are applicable in the case of private values for standard auctions, as we stressed.
and hence the need for hindcast comparisons. The analysis concentrates on the CC signal obtained from the seasonal- averaged temperature and precipitation fields and shows that the fine-scale variability of the CC signal is generally small compared to its large-scale component, suggesting that little AV can be expected for the time-averaged fields. For the temperature variable, the largest potential for fine-scale addedvalue appears in coastal regions mainly related with differential warming in land and oceanic surfaces. Fine- scale features can account for nearly 60 % of the total CC signal in some coastal regions although for most regions the fine scale contributions to the total CC signal are of around *5 %. For the precipitation variable, fine scales contribute to a change of generally less than 15 % of the seasonal- averaged precipitation in present climate with a continental North American average of *5 % in both summer and winter seasons. In the case of precipitation, uncertainty due to sampling issues may further dilute the information pres- ent in the downscaled fine scales. These results suggest that users of RCM simulations for climate change studies in a delta method framework have little high-resolution infor- mation to gain from RCMs at least if they limit themselves to the study of first-order statistical moments. Other possible benefits arising from the use of RCMs—such as in the large scale of the downscaled fields– were not explored in this research.
Of course, the quotation phase is based on a knowledge database that contains the major variables characterizing the enterprise processes with an Excel interface. This is a dynamic tool through the consequences immediate visualization of one of these variables modification (value of the hour rate). These functionalities allow mastering the economical characteristics of the manufacturing processes. Due to the complete integration with technical indicators of the processes, quotation step is possible in less than ten minutes and, when considering the past productions and when comparing the effective results with the previsions, the precision is less than 5%. Then the industrial and commercial plan indicates the contribution of the production to the enterprise result thanks to the comparison with the dynamic stabilization point.
Figure 11. Relative decrease in d eﬀ uncertainty resulting from the addition of six FIR channels to four MIR channels, for Arctic winter and tropical conditions.
So far, the obtained results are theoretical because the instrument which they refer to does not exist yet. To give it a more realistic ﬂavor, the same method was applied to the real instruments described in section 2.2. For this, the addedvalue of the future TICFIRE satellite was quantiﬁed with regard to MODIS (see section 2.2). Figure 12 is similar to Figure 11 except that the reference corresponds to the 10 bands of MODIS, and the full instrument corresponds to MODIS channels complemented by the nine channels of TICFIRE. The patterns for the eﬀective particle diameter are very similar to those of Figure 11. The values are slightly lower, though, because the 10 channels of MODIS contain more information than the four MIR channels of the synthetic radiometer. The largest diﬀerence is the more limited improvement for cloud particles beyond 100 μm in the tropical case. The reductions in uncertainty for optical thickness and cloud top altitude are also shown for the Arctic case. FIR observations help constrain all cloud parameters for a large variety of cloud characteristics. They also reduce the correlated errors of the retrieved parameters (Table 4). The reductions in uncertainty for cloud top altitude and optical thickness in the tropical case are not shown since they are not as signiﬁcant as in the Arctic case. These results overall highlight that adding TICFIRE observations to existing MODIS obser- vations would greatly increase the performance of cloud properties retrieval algorithms in the polar regions and also for thin cirrus and convective ice clouds in the tropics [Hong et al., 2007].
5.3 Beyond Stockless and onto One4All
The CFO of one hospital reflected that the primary driver for change was the need to figure out ways to drive down cost and become more flexible to perform capital investments. Even though an appropriate inventory monitoring system was not in place he said that visibly “a lot of waste was lying around”. Furthermore, there was the added pressure of addressing the issue of patient safety which for one of the hospitals was costing as much as $9,000 on average to accommodate the adverse effects caused in patients. Worth noting is that the calculated cost did not include any losses of potential lawsuits, but rather the required extra patient days, extra drugs, extra tests and extra labour to rectify a mistake. Such was not acceptable at either one of the researched hospitals as both were found to be in financial distress and at a time which coincided with the decision made to engage JAFKO Enterprises in its One4all proposition.
(Email : email@example.com, firstname.lastname@example.org )
This plant is designed to meet the challenges of process development dedicated to the use of residues to produce high valueadded products. It is developed with flexibility in mind as to the choice of substrates (conventional residues, industrial wastewater and synthetic media), microbial strains (known, genetically modified or new) and according to appropriate fabrication protocols. Its strategic orientations lie within the scope of the training of highly trained researchers and new initiatives of industrial research for sustainable development. This new facility is equipped with high tech instruments divided into two distinct fermentation lines. All equipments are food grade or better which enables us to work with many different types of substrates and for various purposes. Research carried out by our team includes biofertilizers, bioinsecticides, bioplastics, biodiesel and enzymes. Several pretreatments, formulation and purification steps can be carried out at the plant itself (Continuous centrifuge, spray dryer, freeze dryer, mixing tanks, ultrafiltration system, reverse osmosis, chromatography system, flow cytometer and hydrolyser). It also offers support from highly trained researchers for optimization of the process. The scale up can reach up to 1 500 L which can be centrifuged on site by a pharmaceutical grade continuous centrifuge. This article describes in details the equipment and services that are available at this pilot plant.
Little is known about sponge genome organization. According to the most complete study of sponge karyotypes published to date, work that focused on freshwater demo- sponges only, sponge genomes are thought to be organized on about 23 micro-chromosomes.  And in terms of gene content, sponge genomes have between 17 000 and 41 000 genes. [30,40–43] The highest estimates are high even compared to the estimated gene number of some bilaterians.  This gene number may be overestimated by the accidental inclusion of symbiont or transposable element-coding sequences in draft genomes, and further analysis is needed on a larger set of sponges for a better estimate. Nevertheless, according to recent data on A. queenslandica, sponge genomes may be compact, with gene density at about 9 genes per 50 kb window, and therefore more similar to choano ﬂagellates than to bilaterians.  This may partly explain the relatively high number of genes compared to total genome size. If con ﬁrmed, sponges would illustrate the “G-value paradox” that is, a disconnect between gene number and body complexity.  It is nevertheless premature to extrapolate from these data to all sponges.
Over the WAC area (Fig. 11 a), NARR occurrences range from about 0 to 12 storms per month, with a sta- tistical median and mode of around 4 storms per month. The regular regime is materialized by the compact shape of the distribution, implying that relatively small disper- sion occurs around the mean occurrence value (~5 storms per month). ERAI distribution is shifted toward the right- hand side, suggesting more presence of months with higher storm occurrence values, as suggested by the larger median (5 storms in ERAI instead of 4 in NARR). This distribu- tion shift also suggests that ERAI positive occurrence bias over the coastal area (Fig. 3 ) may be related to a larger than observed contribution of very active regimes, with an occurrence range of about 6–9 storms per month during the winter season. It is shown that ERAI-driven CRCM5 sig- nificantly reduces the shift of the pattern, leading to better agreement in terms of median and modes with respect to NARR. The intensity KDE shows that, in this WAC regular regime, storm intensity has a median value of ~7 CVU and is pretty well reproduced by ERAI and CRCM5 compared with the NARR distribution. It is worth adding that, even for extreme events (12–14 CVU), the RCM still behaves very well. On the other hand, MPI-ESM-LR and CanESM2 models show fair skill in reproducing the observed shape of the occurrence distribution in a manner very compara- ble to ERAI. Interestingly, it is shown that, as with ERAI, the regionally downscaled simulations can still bring some improvement, especially for a weaker occurrence regime. Finally, serious discrepancies arise in the EC intensity distribution of the CanESM2, in which a weaker storm regime dominates the distribution. The correction brought by CRCM5 is very meaningful even though there is an overestimation of higher intensity storms in return. This too-intense storm category, captured by both GCM-driven CRCM5 runs, could be linked with the SST warm biases inherited from the GCM simulations.
In this study, an experimental setting reproducing conditions of solid-state fermentation was used for high addedvalue metabolites production and was compared with a classical submerged culture. The experimental setting contains a metal structured packing allowing microbial growth as a biofilm. SSF-like conditions lead to technologic progress such as absence of foam formation, persistence of the microbial system and improvement of downstream processing operations. Moreover, dynamic of metabolites excretion is improved in SSF-like conditions but are not yet optimized. Consequently, it is required to define physicochemical and biological parameters affecting the biofilm distribution inside the structured packing and the excretion of the metabolites. X-ray tomography proved to be a relevant tool to characterize parameters influencing microbial growth inside the support. The next step of this work will consider hydrodynamics aspects such as liquid and gaz flow rate and physiologic aspects such as amount and origin of the carbon source.
Equity in the value chain and upgrading potential for smallholder farmers
Both teak planters and traders had a positive return from their activities. Profitability of activities is an important criterion to expect those agents to continue to perform their functions in the value chain. Two distinct conclusions can be draw from the analysis of costs and benefits sharing among value chain stakeholders, by considering separately the evenness in benefits sharing and equity in the value chain. First, the sharing of benefits is drastically uneven to the detriment of teak planters. This suggests a poor competitiveness of farmers with respect to value appropriation in the value chain. Second, given that traders were the main contributors to value addition and bear the major part of costs in the value chain (Figure 3), the results did not provide sufficient evidence of inequity to the disadvantage of teak planters, hence the rejection of the second hypothesis. The latter conclusion tallies with other studies reporting on the lack of evidence of an exploitive role of middlemen (11).
Source: Author’s own elaboration
The turnover generated by the wholesale trade of coal is the product of the number of bags of charcoal by the sale price of a bag of coal in Bujumbura. The unit price is BIF 40,000/bag (25, 9 US$). The valueadded is positive. The marketing of charcoal by wholesale merchants is a wealth-creating activity. Net operating income (RNE) is positive and represents 62, 9% (the majority) of the valueadded. The wholesale marketing of charcoal generates jobs and workers are the second largest beneficiary of the addedvalue. The marketing of charcoal by wholesale traders is somehow a wealth-creating activity.