The management of distribution channels has traditionally been the responsibility of:

Cited by (0)

Business-to-Business Channels

The first industrial distribution channel a business marketer can use is the direct-marketing channel, which is quite similar to channel 1, except the final consumer is replaced with a business customer. Most industrial goods such as raw materials, equipment, and component parts are sold through this business channel. There is no need for wholesalers or other intermediaries in this channel because the goods are sold in large quantities. In the case of small accessories, producers sell their products to wholesalers or industrial distributors, which in turn sell them to business customers. Brokers and sales agents are also common intermediaries in industrial marketing channels. Often small producers are represented by independent intermediaries called manufacturers’ representatives to market their products to large wholesalers or to final business customers [18,23].

Any one or more of the above-mentioned alternative distribution channels might be used by manufacturers to make their products available to final customers. An individual producer may choose different marketing channels with respect to its different types of products or customers.

5.9 Distribution and distribution channels

The distribution process shows how the product is transferred from the manufacturer to the consumer. The appropriate product placement and positioning have to be achieved, along with an adequate advertising strategy. The term ‘distribution’ also refers to the position of the apparel manufacturer. The selection of an internal or an external channel is determined mainly by the size of the company, the vertical integration and the competition pattern. If a company chooses internal distribution it assumes the responsibility for the product characteristics and the control of distribution. Such companies work with well-defined consumer segments and quick placement on the market.

Wholesalers have their merchandise stored centrally and distributed from there to the regional branches and shops. Others bring their goods to the non-food distribution center, from where the products go directly to the different points of sale. Strategies have to be adapted to the market development with the creation of new operational bases or distribution channels (see also Table 5.25 below).

The strategy can also be used for a desired market entry since it shortens the communication along the supply chain. Choosing external distribution requires a sound knowledge of the ever-changing market characteristics, specifically of the global competitors. Consequently also legal aspects become more important. High skills in communication and the development of (long-term) partnerships are required. In external distribution, intermediaries have the role of sales agents (contracting).

For a better understanding the external distribution process can also be analyzed from the viewpoint of the retailer and wholesaler as customers. The market differentiates between direct and indirect distribution (see Fig. 5.23). Distribution is shown here as export of manufactured apparel. The structure stands for all exports along the supply chain: raw material like fibers and crude oil, yarns, gray and finished fabrics as well as auxiliaries and accessories.

Direct distribution includes selling with and without contract as well as exporting by means of a retailer’s service account. Such sales offices may be either company owned, a foreign subsidiary or a foreign independent sales agent. The services provided include edition of markets, and monitoring of trends and styles. The offices often can provide access to specialty priced goods or new merchandise. Typically the associated fee is 7% of the sales price. Indirect sales require an export management company or an export trading company as an agent. The export management company may take a certain risk as it assumes the credit risks of the customer, whereas the US Export Trading Company is financed through the Export Trading Company Act’s loan guarantee program (see Fig. 5.24). Driving forces for the choice of a distribution or purchasing/sourcing strategy in a global market are prices and lead times of products (see sourcing strategies, Section 5.10).

Of growing importance for product placement are the Internet and mail order, often used in addition to other channels of direct distribution. Even if sales through these channels represent only about 4% of total sales, consumers may search for information before shopping in other channels. The Internet serves to reinforce a buying decision. Particularly the younger generation are taking advantage of this facilities.

Retail operates with various channels that provide specific characteristics and are preferred for these typical aspects by groups of consumers, as shown in Table 5.23. Over decades the channels have been defined exclusively. Only in recent years, as consumers have started shopping in different channels, have these classifications become less distinctive. Mixing characteristics of different channels might attract additional customers and increase sales.

Table 5.23. Retail distribution channels

Table 5.24 gives information about the attractiveness of the different retail channels in the USA, chosen for men’s and women’s apparel. The consumer preferences give valuable information for product positioning. Figure 5.25 shows the structure of the Swiss retail market channels in the late 1990s.

Table 5.24. Market share (%) of channels in the USA

The placement in distribution channels has to be chosen very carefully. Multiple strategies such as favored shop addresses, shop design and attractive Internet catalogs require high investment. Excellent services, especially at the point of sale, enable a personal relationship with customers. Promotion has to follow the trends: who and what is in or out, lifestyles of society, the role of women, patchwork identity of consumers, etc. Tables 5.25 and Example 5.5 give facts and details of US and European product placements and distribution channels.

Table 5.25. Distribution channels of US textile companies

Example 5.5 Sales outlet and direct distribution of a Swiss and a global designer.

AKRIS has eight boutiques worldwide in Paris, Boston, Monte Carlo, Düsseldorf, Frankfurt, Tokyo, Seoul and New York. In 1997 it opened the first shop-in-shop boutique, expanding to 35 shops worldwide. All stores are designed equally, with furniture by the architect Ferruccio Robbiani.

13.2.3.2 Vendor managed inventory

VMI is essentially a distribution channel operating system whereby the inventory at the retailer is monitored and managed by the manufacturer. It includes several tactical activities, such as determining appropriate order quantities, managing proper product mixes, and configuring appropriate safety-stock levels (Chopra & Meindl, 2001). The rationale is that by pushing the decision making responsibility further up the supply chain, the manufacturer or vendor will be in a better position to support the objectives of the entire integrated supply chain resulting in a sustainable competitive advantage. Refer to Fig. 13.4, in the conceptual evolution of inventory management, VMI stands in the middle.

In VMI, the retailer still owns the inventory and the manufacturer simply manages it. Consignment selling (CS) can be considered the next step (after VMI), where the manufacturer owns the inventory and the retailer charges a percentage for providing shelf space and customers. In practice, both VMI and CS are being practiced in the industry often under the more popular acronym VMI.

TAL Apparel is a Hong Kong-based apparel manufacturer having factories in HK, Thailand, Malaysia, Taiwan, China, Indonesia, Vietnam, Mexico, and the USA. TAL collects point-of-sale data of J. C. Penny’s shirts directly from its stores in the USA, then run the numbers through proprietary software to determine the number of different styles, colors, and sizes to make. One of its Asian factories manufactures the shirts and sends it directly to each J. C. Penny’s store (bypassing the warehouses and corporate decision makers. TAL’s New York office analyses the market, trends, sales data to design, and fine-tune the proprietary replenishment software. TAL provides similar services to Brooks Brothers and Lands’ End.

Wal-Mart believed to be the pioneer to start VMI with Procter and Gamble has supposedly moved to CS (Cid, Gordon, Kearns, Lennick, & Sattleberger, 2000). The enabling technology behind successful VMI is electronic data interchange that provides the manufacturer with essentially the same point of sales and inventory information.

2.2.3. Downstream Distribution Channels

The firms external downstream consists of all distribution channels and processes of the supply chain that help the product reach its ultimate destination – the end consumer. This involves inventory management, warehousing, distribution networks, retailers, after sales service, etc … Activities such as after sales and maintenance services ensures continual customer satisfaction and helps the firm not only manage its forward supply chain, but also its backward supply chain.

The length of internal supply chains and the subsequent delivery channels can vary quite enormously, depending upon the type of product. In the aerospace industry for example, the internal supply chain is fairly long, but its distribution channel is quite short – where the finished product is delivered directly to the customer. In the case of sweets or chocolates, the internal supply chain is short but the distribution channel is fairly long and extensive. From point of manufacture the packaged sweets/chocolates are sent to distributors, these are then sent to retailers (supermarkets, grocers, etc …) and only then does the consumer purchase them for consumption.

High SNR Regime

At high SNR (such that ρλk/nt ≫1 for all k), we may approximate C¯CDIT by

(5.65) C¯CDIT≈E∑k=1nlog2 ρntλk,= nlog2ρnt +E∑k=1nlog2(λk),

(5.66)=nlog2 ρnt+∑k=1nElog2χ2 (N-n+k)2,

(5.67)=nlog 2ρnt+1log2 ∑k=1n∑l=1N-n+k-11l-nγ,

(5.68)= nlog2ρnt+ 1log2∑k=1n∑l=1N-k1l-nγ.

Again, χ2(N-n+k)2 designates a χ2 variable with 2(N-n+k) degrees of freedom, and the last term in (5.66) can be upper-bounded by log 2N!(N-n)!.

The important result is that the ergodic capacity C¯CDIT at high SNR scales linearly with n (by contrast to the low SNR regime). The multiplexing gain gs is equal to n, similarly to the CSIT case. However, even though they scale similarly, C¯CDIT and C¯CSIT are not equal. This might seem surprising as the water-filling solution for uncorrelated channel approaches a uniform allocation at high SNR. Actually, the difference C¯CSIT-C¯CDIT is a constant equal to nlog2(nt/n) at high SNR [Gui05]:

if nr⩾nt, the constant is equal to zero, and C¯CDIT=C¯CSIT,

if nr<nt, the constant equals nr log2(nt/nr); intuitively, the gain offered by CSIT corresponds to the amount of energy saved by the transmitter not emitting in the subspace not seen by the receiver. If nr=1, we find again the classical beamforming (or array) gain of nt already discussed earlier.

Figure 5.3 compares the ergodic capacity of a various nr×nt i.i.d. Rayleigh fading channels for the CSIT and CDIT cases:

in the high SNR regime, the gain offered by CSIT vanishes for 2 ×2,4×2 and 4×4, in agreement with our earlier remarks, although a gap remains in the low SNR regime,

in the high SNR regime for the 2× 4 scheme, CSIT offers a gain equal to 2log2(4/2)=2 bps/Hz over CDIT, confirming the benefit of CSIT in asymmetric scenarios where nt>nr (which is the most common scenario in downlink configuration),

in the case of full CSIT, 2×4 and 4×4 schemes provide the same capacity at all SNR levels (naturally, this is not true in the CDIT case),

in all cases, the capacity scales asymptotically as n. Indeed, for every doubling of the SNR (3 dB increase), the capacity increases by n bps/Hz in the high SNR regime.

Finally, note that when nt=nr= 1, the Rayleigh fading channel capacity at high SNR reduces to

(5.69)C¯CDIT(N=n=1)≈log2(ρ)+ Elog2|h|2 ,=log2(ρ)-γlog2=log2(ρ)-0.83,

which is smaller (by ≈0.83 bps/Hz) than the capacity of a SISO AWGN channel (equal to log2(ρ) in the high SNR regime). Increasing the number of antennas at one side (i.e., increasing N, but keeping n=1) does not increase the multiplexing gain (which remains one), but might still increase the capacity for SIMO/MISO systems compared to (5.69):

for SIMO systems (nt=n=1,nr=N), the high SNR capacity becomes approximately log2(nrρ), i.e., there is an array gain of nr,

for MISO systems (nr=n=1, nt=N), the ergodic MISO capacity at high SNR reads as

(5.70)C¯CDIT ≈log2(ρ)+Elog2h2/nt,

(5.71)=log2(ρ)-γlog2+1log2∑l=1nt -11l,

Techniques in Image Distribution

The spectacular development of imaging sensors, communication, and internet infrastructure has also made the counterfeiting tools easily accessible and more affordable. In addition, the threats and vulnerabilities of the internet have also increased manyfold. All of these together pose a formidable challenge to the secure distribution and post-processing of images.

Requirements of Secure Digital Image Distribution

In light of the above-mentioned vulnerabilities of image distribution channels and processes, below are a few requirements of secure image distribution:

•

Data Integrity and Authentication: Digital image is easy to tamper with. Powerful publicly available image processing software packages such as Adobe Photoshop or PaintShop Pro make digital forgeries a reality. Integrity of the image content and authentication of the source and destination of the image distribution are, therefore, of utmost importance.

•

Digital Forensic Analysis: Consider a litigation involving a medical image that is used for diagnosis and treatment of a disease, as shown in the post-processing example in Figures 7 or 8 above. Integrity of the image again is vital for the proper forensic analysis.

•

Copyright Protection: Digital copyright protection has become an essential issue to keep the healthy growth of the internet and to protect the rights of on-line content providers. This is something that a textual notice alone cannot do, because it can so easily be forged. An adversary can steal an image, use an image processing program to replace the existing copyright notice with a different notice and then claim to own the copyright himself. This requirement has resulted in a generalized digital rights management initiative that involves copyright protection technologies, policies, and legal issues.

•

Transaction Tracking/Fingerprinting: In the continuous (re)distribution of digital images through internet and other digital media (CD, camera, etc.), it is important to track the distribution points to track possible illegal distribution of images.

Next, we will discuss how the digital watermarking technology can help enforce these security requirements.

Example of secure image distribution: digital watermarking

Digital watermarking is the process of embedding a digital code (watermark) into a digital content like image, audio, or video. The embedded information, sometimes called a watermark, is dependent on the security requirements mentioned above. For example, if it is copyright application, the embedded information could be the copyright notice. Figure 11 shows a block diagram of the digital watermarking process.

Figure 11. A simple watermarking process: embedder and detector.

Consider that we want to embed a message M in an image, fo, also known as the original image. For the sake of brevity, we drop the indices representing the pixel coordinates. The message is first encoded using source coding and optionally with the help of error correction and detection coding, represented in Figure 11 as e(M) such that

[26]Wm=eM

The encoded message Wm is then combined with the key-based reference pattern Wk, in addition to the scaling factor α, to result in Wa which is the signal that is actually added to the original image, given as

[27]Wa=αfoWM⊗,WK

Note that the scaling coefficient α may be a constant or image dependent or derived from a model of image perceptual quality. When the scaling factor α is dependent on the original image, we can have informed embedding which may result in more perceptually adaptive watermarking. The operation involving Wm and Wk is the modulation part, which is dependent on the actual watermarking algorithm used. As an example, in a typical spread-spectrum based watermarking, this operation is usually an exclusive NOR (XNOR) operation. The final operation is an additive or multiplicative embedding of Wa with the original image. We show a simple additive watermarking process as

[28]fw=fo+Wa

The result of this process is the watermarked image fw which may actually need some clipping and clamping operations to have the image pixel values in the desired range.

The detector, as shown in Figure 11, does the inverse operations of the embedder. If the original image is known, as in the case of nonblind watermarking, it is first subtracted from the watermarked image. Blind watermarking detector, on the other hand, uses some filtering with a goal to estimate and subtract the original image component. The result of filtering is then passed to the decoder, which collects the message bits with the help of the shared key, and finally gets back the decoded message M′.

Figure 12 shows the original image, watermarked image, and the difference image using a spread-spectrum based watermarking algorithm, respectively. The enhanced difference image represents a noise-like image, wherein the scaling coefficient α is clearly dependent on the original image. As evident from Figure 12b, the hidden message is imperceptible as well as the change in the original image. The imperceptibility is one of the primary criteria of digital watermarking.

Figure 12. (a) Original image, (b) watermarked image, and (c) the difference image.

There are other criteria such as unobtrusiveness, capacity, robustness, security, and detectability. Thus, digital watermarking is a multidimensional problem, which attempts to make a trade-off among a number of different criteria. The inserted watermark should not be obtrusive to the intended use of the original image. Because the watermarking process does not increase the size of the original image, it may be desirable to add as much information as possible. However, generally speaking, the more information one adds, the more severe will be the impact on the perceptual quality of the image. From the detection point of view, watermark should be robust enough to tolerate a range of image post-processing and degradation. In addition, the embedded watermark should also be secure enough so that it may not be removed from the watermarked image. Finally, detectability of the embedded watermark is an important criterion that places some constraints on the embedding algorithm.

In order to fully appreciate the use of watermarking for image distribution, it is instructive to look at different classes of watermark, which is delineated in Figure 13.

Figure 13. Watermark classifications.

For an example, depending on the robustness criterion of watermarking, they can be classified into three categories such as robust, fragile, and semi-fragile. Note that while the robustness of watermark is important, it is not always equally desirable in all applications. Let us take the authentication as an example. If you, as the owner of a digital image, embed some authentication information into the image using watermarking, you might want to see the authentication fail, when some one steals your image. This is an example of nonrobust watermark, where the fragility of the embedded watermark is required to detect any forging done on the digital image.

Now, consider how watermark provides the security requirements for secure image distribution. In particular, we look into the authentication of watermark. Figure 14 shows the use of watermark for image authentication, which has some similarities with the cryptographic authentication using digital signature. However, there are potential benefits to using watermarks in content authentication and verification. Unlike cryptography, watermarking offers both in-storage and in-transit authentication. Watermarking is also faster compared to encryption, which is very important in internet-based image distribution. By comparing a watermark against a known reference, it may be possible to infer not just that an alteration occurred but what, when and where changes have occurred. As shown in Figure 14, first we identify some features from the image and then we compute the digital signature, which is then embedded into the image. The detector takes a test image, computes the signature and extracts the embedded signature, and then compares the two. If they match, the image is authenticated. If they do not match, then the image may have gone through some forging/tampering effects and the algorithm may optionally detect the tamper area as well.

Figure 14. Image authentication using watermark.

One of the limitations of the watermarking approach in secure image distribution is that the technology is not yet widely deployed, and nor is the protocol satisfactorily standardized. Other major limitation is that the watermarking process cannot be made sufficiently robust to arbitrary types of different attacks. In reality, therefore, the hybrid combination of cryptography and watermarking is expected to improve the secure distribution of images across the unreliable internet.

15.2.3.2 OEM Recall Communication to Customers

Another important aspect of regulatory reporting is OEM recall communication to all direct distribution channels sometimes referred to as consignees (eg, distributors, contractors, customers). The OEM must provide information to identify the product and help reduce the risk of public health hazards in line with the level of health hazard classification assigned by the FDA evaluation (FDA, 2014c, pp. 21–25). A recall communication through various media (eg, detailed mailed instructions, verbally communicated instructions by telephone, or press release) generally reflects the firm recall strategy and report specific information by following regulations for the designated product, listing any OEM corrective action, and providing end user information on failure recognition and recommended customer action (FDA, 2014c; Brown, 2012). Checklists for reports of correction or removal of medical devices are referenced in 21 CFR § 806.10(a)(1–13) (FDA, 2014c, p. 48). The OEM must also conduct effectiveness checks to ensure that notification has occurred to all distribution channels and that proper recommended action has been taken by all customers (FDA, 2014c, pp. 21–25).

Additional OEM recall requirements fall under the QSR 21 CFR § 820.100 through 21 CFR § 820.250 guiding the establishment and maintenance of processes associated with CAPA plans required in OEM recall communication to all customers so that all distributed units receive proper through action that restores the product to acceptable levels of conformity or proper disposal (FDA, 2014c, pp. 27–28).

OEMs must maintain contact with device user facility representatives, generally from the purchasing department or representative of central dispatch, who have the responsibility to ensure that recall notifications systems are monitored, received, and documented in master log and transaction files. Service bureaus, such as the National Recall Alert Center, federal agencies, and industry sources are some of the places where notifications on product recalls can be located (Table 15.3).

Table 15.3. Examples of Service Bureaus and Agency Resources for Product Recall Notifications

Resources for product recall and device information include regulatory organizations such as the Joint Commission, National Fire Protection Agency, Centers for Medicare and Medicaid Services, and state organizations. Additional resources include professional organizations such as the American College of Radiology and the American Society for Healthcare Engineering.

12.3.3 Channel loss, quantum repeaters, and quantum memory

Early in Section 12.3.1 we mentioned that the quantum states transmitted in the QKD channels cannot be amplified for the same reason it cannot be cloned. This makes the quantum channel loss a very important factor in practical QKD implementations. This loss not only limits the key generation rate but also affects the key security, because we have to assume that Eve is able to collect all of the lost photons.

Performance of various CV and DV QKD protocols with respect to the loss is reviewed in Ref. [80]. Since different QKD protocols depend on a large number of different parameters, a direct comparison is difficult. However, a general pattern can be seen from the two examples in Fig. 12.3. Here the channel efficiency, characterized by the ratio of the key generation rate to the pulse rate, is plotted as a function of the channel loss for various protocols: continuous variables with Gaussian modulation (CV), perfect single-photon source (1-ph), weak coherent pulses with and without decoy states (decoy and WCP, respectively), entanglement-based (EB), and coherent one way (COW). In both Fig. 12.3(A) and (B) the mean intensity for DV protocols and variances for CV protocols are assumed to be optimized, and Bob’s receiver is assumed to have a unity transmission. The error-correction codes are implemented as described in Ref. [80] (we will talk more about it in Section 12.3.4). Other relevant parameters are listed in Table 12.1.

Table 12.1. Parameters for Fig. 12.3A and B.

P&amp;M, Prepare-and-measure; ES entanglement-sharing protocols. The electronic noise variance is relative to the shot noise.

The most important message of Fig. 12.3 is that the losses degrade performance of all QKD techniques. Up to a certain point, the key rate reduction is approximately a polynomial function of loss, but when the critical loss value is reached, the rate drops catastrophically and the QKD becomes impractical. This happens when the raw key fraction required for the error correction and privacy amplification approaches the entire raw key length. The critical loss value depends on the system’s parameters. Notice that as we improved these parameters from the set (a) to the set (b) in Table 12.1, the critical value of the loss increased significantly. For an ideal system the catastrophic key rate drop does not occur at all, even though the key-to-pulse rate ratio still declines as a function of loss. Upper limits of this ratio derived for such an ideal system under different sets of assumptions and known as the TGW [90] and PLOB [91] bounds:

(12.47)RTGW=log 2(1+η1−n)≈2.89η,RPLOB=log2(11−η) ≈1.44η.

Here η is the channel transmission, and the approximation is made for η <1.

Unfortunately, the hard limit imposed by the hardware imperfections on the QKD range is prohibitive for many important applications requiring long-range communications. For example, a 40 dB loss corresponds to some 200 km of a telecom fiber. Such an experiment using a DPS QKD protocol was performed in 2007 [92], demonstrating 12.1 bps secure key rate produced out of 10 GHz raw key pulse rate, that is, approximately 1.2×10−9 channel efficiency.

To get around the loss problem, one may introduce a number of communication nodes along the communication channel. With one such extra node (Charlie) the channel topology will be: A–C–B. If Alice shares a private key with Charlie by using a QKD protocol, and Charlie likewise shares another private key with Bob, then Alice and Bob also can share a key. To do so, Charlie can use Alice’s key to encrypt Bob’s and send the result to Alice via a public channel. Alice now is able to recover Bob’s key; this key is also known to Charlie, who therefore has to be trusted. We will review some practical realizations of this approach later, but note that for a large number of intermediate nodes, A–C1–C2–…–B, the requirement that all C nodes (or relays) must be trusted becomes practically equivalent to building a fully secure communication channel, which is assumed impossible under the QKD paradigm.

Quantum mechanics allows Charlie to help Alice and Bob to establish a secure key without learning it in the process. This approach to the long-range QKD is known as the quantum repeater. To implement a basic quantum repeater, Charlie, instead of establishing keys with Alice and with Bob, performs a Bell measurement on the two photons sent to him by these two parties, just like he does it in the MDI protocol. Communicating its result to Alice or Bob, Charlie facilitates the entanglement swapping as described in Section 12.2.4.2. Now Alice and Bob share a two-photon entangled state unknown to Charlie and can generate a private and secure key.

Entanglement swapping between three nodes is easily generalized to an arbitrary long chain of nodes and can be made noise resilient by implementing error correction and entanglement distillation or purification steps. In theory, this should not only eliminate the hard limit on the channel loss but also allow for exceeding the TGW bound (12.47). However, a practical realization of such a protocol requires a synchronization of the states’ measurement and preparation, and this requires a capability to store quantum states for sufficient intervals of time while preserving their coherence, that is, a quantum memory.

A quantum memory is based on strong coupling between photons and matter qubits. The latter may be implemented as individual ions or atoms, atomic ensembles, quantum dots, color centers, and other quantum systems. Here we discuss three examples of different approaches to quantum memory based on atomic ensembles and rare-earth ions. The same approaches can be adapted to other matter qubit implementations.

In our first example [93], quantum state of a weak light pulse was recorded in two cesium vapor cells with the opposite spin polarizations, and preserved for up to 4 ms. The state was then recovered with a better fidelity than could be achieved by a classical memory. The underlying principle of this quantum memory realization was a QND measurement (see Section 12.2.2) on the light pulse and the atomic ensemble, followed by an electro-optical feedback system. To utilize this system for a practical quantum repeater in an entanglement-sharing QKD protocol, one needs to implement the same type of measurement with an entangled photon in place of a weak laser pulse. However, most of the available entangled photons sources have optical bandwidth far exceeding that of a typical atomic transition, and cannot be efficiently used for this purpose. A progress in this direction was achieved recently with building an SPDC source based on a very high-finesse optical resonator, which can match the important cesium and rubidium transitions in optical frequency and bandwidth [94,95].

The second example of a quantum memory realization is based on electromagnetically induced transparency. In this process, the signal pulse group velocity is controlled by a classical optical pump field. When the pump field is gradually turned off, the signal pulse slows down and eventually stops, being transformed into a dark-state polariton wave. By turning the pump back on, the process is coherently reversed and the initial pulse is retrieved. This approach has been used in a number of experiments involving cold atomic ensembles and demonstration of storing DV and CV quantum states. For CV states, a squeezed vacuum is most typically used. The typical storage efficiency in this case is about 10%–15%, but the storage and retrieval can be accomplished practically without adding noise [96].

Our third example is based on a photon echo technique. The advantage of this approach is the increased optical bandwidth, which is allowed by the inhomogeneously broadened absorption in the active medium. The broadening causes the signal dephasing, which is, however, compensated due to the phase-reversing nature of the photon echo. Furthermore, the photon echo allows for control over the release of the stored quantum state. This approach has been experimentally realized using solid host crystal, such as lithium niobate or yttrium orthosilicate, doped with rare-earth ions, such as Pr3+ or Er3+; see for example, Refs. [97,98].

8.2 End-of-life product management options

When it leaves the manufacturer's gate, a textile product enters the retail and distribution channels, via which it then reaches the customer. From this point, the use phase begins. After a certain period, the customer decides to dispose of the product and this is the end-of-life phase. There are many destinations or fates of a product at the end of its life. These include

reuse for primary and secondary purposes,

recycling (open- and closed-loop types),

landfilling,

incineration.

From the perspective of life cycle assessment (LCA) or environmental impact assessment, two terms can be applied at this point: one is ‘impact’ and the other is ‘benefit’ or ‘credit’. These terms are self-explanatory; clearly the end-of-life option that brings credits or benefits is favoured over one that causes any impacts. Each of the options for disposal has its own pros and cons, and these will be discussed in the sections that follow, with a focus on textile products.

Reuse is the first and best option in terms of gaining only environmental benefits or credits with nil impacts. Recycling is the next best option, and it brings both benefits and certain impacts, followed in the hierarchy by incineration. Incineration can be performed with or without energy recovery. Clearly, incineration with energy recovery is preferred (generating impacts as well as bringing benefits). The final and most undesirable option is disposal in a landfill, which creates environmental impacts only. The perceived degree of these impacts and benefits is quite subjective and cannot be generalized, as it is influenced by multiple factors. This aspect is not covered in the current discussion.

To summarize, as defined by the 2008 European Union Waste Framework Directive (European Council, 2008), waste prevention is the best choice of waste management, followed by the other options in the hierarchy of reuse–recycle–energy recovery–landfill.1

Owing to the limitations of space in landfills and also to promote the concepts of resource conservation and to minimize the environmental impacts of producing a new product, waste management options and concepts are being refined. A very famous concept called 3R's – Reduce, Reuse and Recycle – has been refined to 5R's as Refuse, Reduce, Reuse, Repurpose and Recycle in the same order of hierarchy, as keeping the same order is also crucial. Similarly, the new hierarchy of waste management has been known as the ARRE strategy: Avoid, Reduce, Recycle and Eliminate.

2.4 Distribution Channels

In a digital-based smart energy era, the expectation is that the main distribution channel will be online, and the energy retailing prize will hinge on innovative digital platforms (refer chapter by Cooper in this volume) in order to secure the energy automation, own generation, and energy efficiency customer space. Already, many companies are shifting their positioning to cluster energy management offerings around a central energy efficiency and energy saving proposition, and using new channels, such as social media to engage with customers, such as E.ON.11

A risk for traditional energy companies is that their distribution channel to end customers becomes disintermediated in ways that are not dissimilar to what has happened to incumbent publishers and booksellers with the advent of Amazon. Not only is the channel to market for incumbents dominated by the new platform, but the actual demand for product is eroded as the platform acts as an aggregator for self-publishing and second-hand sales. And, of course, the offering is now much wider than just books, with the combination of a trusted brand, and sheer presence, providing a marketplace joining consumers to a wide range of product providers. There is some evidence that this is happening in New Zealand, where it is possible for households to be directly exposed to the wholesale market (eg, via the retailer, Flick Electric). The next stop is for individual households to trade energy with each other across the AMI technology, which some companies are experimenting with already (such as Reposit Power in Australia)12 and the ability to do this via online communities is being actively researched (Bourazeri et al., 2012).13

Smart grids, microgrids, local generation, and local storage all create opportunities to engage customers in new ways—for example, see the discussion of mircogrids (chapter by Marnay in this volume). Increasingly, we are seeing interest in the power sector from companies in the online, digital, and data management world, who are looking at media and entertainment, home automation, energy saving, and data aggregation opportunities. In a grid-connected but distributed power system, there are roles for intermediaries that can match supply and demand, rather than meet demand itself.

A key consideration for incumbent power utilities is if their brands are perceived as being part of the past that is being broken away from, rather than the future for customers. An energy saving or demand management proposition may be perceived as more credible, coming from a new entrant rather than an incumbent, so use of the brand needs to be carefully considered. At the same time, old brands could become valuable again as a cacophony of new brands emerges and causes confusion.

Another important challenge for companies arises from the need to be an expert in managing data in a smart home, smart city, and smart company environment. As well as data from smart devices and the grid, additional layers of information about demographics, behavior, customer characteristics, and other factors will often be required to exploit the data opportunity best. Many power utility companies already use sophisticated data analytics for customer segmentation purposes, and this can be built on and supplemented by enhanced analytics, big data from social media, and learning from other industries.

However, as consumers become more conscious of the value and potential data security issues associated with AMI technologies, there may be pressure to have more control of their own data, and who has access to it. Active consent for data collection and analysis may need to be sought, and consumers may be able to access and sell their data on to new entrants offering new services reducing the natural data analysis advantages of current utilities.