Which service frequently require direct contact between the customer and the service provider

Service provider

A service provider can offer to customers high-level services that operate over ATM connections provided by a separate network provider. The service provider is therefore a customer of the network provider in addition to being a supplier of services to the end-customer, and therefore has many of the requirements listed above. A more specific requirement of the service provider is to be able to charge at the service level in a way that relates logically to the ATM-level charges imposed by the network provider. Service-level charging is discussed more fully in Chapter 6.

Some service providers might also provide and charge for content. Usually the service provider would like to offer an integrated charge to the end-customer, who would not then be able to distinguish between connection charge and content charge.

2.4.2 Service Provider

Service providers are the entities that provide services to others. These services could be applications, infrastructure, or data services. As “the cloud” grows in popularity, many people have become aware of the three main cloud services models. They are IaaS (Infrastructure as a Service), PaaS (Platform as a Service), and SaaS (Software as a Service). The applications and services provided by a service provider are called the relying party. This is because they rely on the IdP for authentication and identity information.

4.5.1.2 Facilitating administration and control

Service providers must respect legal rules concerning privacy and data protection [EUR 95] or face sanctions and damage to their brand image. Certain tools have been developed to assist in ensuring conformity. These solutions follow three main approaches:

an ergonomic user interface for the definition of personal data processing rules in accordance with the applicable regulations, with sufficient documentation to provide users with appropriate information, or for step-by-step learning of the correct application of legal rules for service providers;

the creation of minimalist and maximalist regulatory ontologies, where the relevant authorities publish a definition of a first (strict) and second (permissive) category of data protection policies (see section 4.5.1.1). The first definition corresponds to a service provider requiring a set of personal data simply to provide a service, using default parameters. The second definition corresponds to a service provider collecting data not only for service purposes but also with the aim of profiting from the obtained data;

a comparison engine for personal data management policies (preferences and practices), allowing entities to identify whether or not a policy corresponds to the expectations of the person concerned.

Note that the sole objective of the user interface is to assist in the definition of practices and preferences. It may usefully exploit regulatory ontologies (defined above) to ensure that service providers correspond to legal regulations and to allow users to define their preferences. It does not guarantee conformity to all legal obligations for service providers, particularly in terms of personal data confidentiality and security aspects. However, the interface enables independent monitoring bodies to automatically check the conformity of personal data processing practices [EUR 95], either on-site or remotely. This may be useful for the CNIL in implementing their new investigative powers, enabling remote detection of infringement of the French Data Protection Act over the Internet33.

From a technical perspective, the use of regulatory ontologies in a user interface requires the capacity to compare rules concerning the processing of personal data. This is not currently possible with respect to the existing standards (P3P). P3P defines a set of values for the PURPOSE, RETENTION and RECIPIENT criteria, but these values are not ordered on a scale, from “most permissive” to “strictest”, for example. Moreover, these criteria are not intercomparable for any given rule. Finally, privacy policies generally correspond to a set of distinct rules with the same weightings.

Dari-Bekara’s doctoral thesis [DAR 12b] offers several responses to these issues, including a scale of P3P values and a preference definition interface based on this scale (see Figure 4.4). Note that this scale may differ from one service type to another.

7.3 Efforts to Reduce Operational Costs

Service providers today face formidable challenges. As customers add new services, the volume of data traffic traversing the service provider’s network continues to grow exponentially. This is driving the cost of networks up with demand. At the same time, revenue per user is remaining flat or increasing at a much smaller rate thus impacting profitability. This situation is depicted in Fig. 7.14 below.

This is putting pressure on the service providers to reduce operating costs. Service providers are using various means to achieve this goal. One of them is to out-source help-desk function. Fig. 7.15 shows the schematic of the out-sourced help-desk function. This requires integration of the IVR and call center system with the OSS/BSS systems so that remotely located call centers can validate customers, enter trouble tickets, resolve issues, and close trouble tickets.

Another technique that service providers are using is to encourage self-provisioning by customers. As shown in Table 7.2, the cost savings from self-provisioning is considerable. Self-service moves contacts from high-cost call center channels to lower cost IVR and web channels.

Table 7.2. Rationale for Self-care Portal of Order Entry

Focus on self-service via IVR and Web allows for the highest and best use of capital and operating resources. Additional channels can be used as alternatives to call centers or self-care portals but have not proved as effective as the IVR and self-service Web portals in delivering savings. For example, the cost for e-mail and Web chat interactions is close to that for live telephone interactions. Fig. 7.16 shows a typical implementation of self-service portal.

Analytics is another area that service providers are using to improve efficiency and quality of decision making to make business competitive and reducing costs. It is shown in Fig. 7.17. It involves extracting data from various OSS/BSS systems, staging the data, loading the transformed data in data marts and enterprise data warehouses, and then integrating these with analytic applications to create dashboards and score cards. The service providers are also using historical data for data mining to determine patterns using artificial intelligence applications and using predictive analytics for predicting future trends.

There are two emerging areas that offer huge potential for further reducing operating and capital costs and also for implementing on-demand services. These include NFV and SDN. These are currently getting lot of attention because NFV and SDN will allow operators to use cheaper generic hardware and allow customers to not only order the services but also configure them as well in near real-time. We will cover these topics and their impact on OSS/BSS in the next section.

Migration

The service provider object offloading process follows a three-step approach: First, the target PF JAR file is transferred to VM and started. Then, a proxy object is created to intercept and capture service request to a remote target service. Finally, the PF containing target service provider object is stopped.

The first step prepares for the service request sent by the proxy object that is created by Java dynamic proxy technique. The proxy object registers to POEM framework with a higher-ranking number than the target service provider object. The POEM framework treats the proxy object as the service provider according to its ranking order. The request to the service interface is sent to the proxy object. Up to now, the service composition works fine and the last step stops original PF.

The migration happens according to the migration decision module command. POEM constructs the migration decision module as a plug-in framework. A user can develop the migration decision strategy plug-ins and install the strategy bundle into POEM, which not only provides the flexibility for the user customized migration strategy but also scales the POEM intelligence.

SLAs

Many service providers are not at the point where they can offer truly substantive SLAs. Some providers don’t offer SLAs at all. Others offer SLAs, but the service guarantees they make are not suitable for many organizations. Your organization may need 24/7 availability for a particular service or application, but there might not be a provider that can offer that. One thing to remember is that if your organization cannot provide a certain level of availability because of a technical limitation, a service provider may face the same technical limitation for the given service or application.

Cloud Platform-as-a-Service

PaaS providers usually deliver a bundling of software and infrastructure in the form of a programmable container and provide a cloud for an end user to host their own developed applications or services. PaaS is similar to SaaS, but with PaaS, the service is the entire application environment—typically, PaaS includes the computing platform as well as the development and solution stack.

Google's Google App Engine is an excellent example of a PaaS architecture. So is Salesforce.com's Force.com platform. In both cases, the end user receives an environment from the provider (also called a container) that is ready to host a particular application or service that the end user requires. The end user does not need to worry about lower-level services such as the infrastructure; these are provided for them within the service.

Service provider strategy

Service providers mirror the modernization deployment options outlined at the beginning of this chapter. These include tool-centric, service-supported, in-sourced, and outsourced approaches to deploying modernization projects. Many times management creates a strategy based on a preexisting relationship with a service provider and this may be a bad strategy. For example, management may feel comfortable with a certain firm that always in-sources modernization work — a scenario under which the service provider completely owns the project, brings in their own tools, and does not provide skills transfer to the client. This may not be appropriate if there are a number of projects or areas that could benefit from assessment, refactoring, or transformation-related activities on an ongoing basis.

This situation is more common than one would think, but it can be avoided by stepping back and understanding what is to be accomplished. For example, if an organization has large volumes of COBOL code on a mainframe and no immediate need to move that code off of the mainframe, the focus is more likely to be on an assessment and refactoring strategy if a business case is in place to drive such a plan. The business case in this situation, for example, might require a plan to clean up, streamline, and consolidate a series of redundant applications and return those systems to the existing platform. This “consolidation scenario” would drive the modernization plan.

Under this scenario, management should craft a strategy that couples the in-sourced project option with the service-supported option. A specialized or high-priority project could be in-sourced, where an outside firm owns the project, while other projects can be jump-started using the service-supported deployment option. This second option should incorporate a rollout program that includes technology and skills transfer to enterprise employees. Over time, the tool-centric approach could be rolled out to various projects across the enterprise.

Another example involves an enterprise where systems were deployed under a highly decentralized management structure for decades. This resulted in numerous systems running on a wide variety of unsupported platforms, using unsupported technologies. In this case, the outsourced option could be selectively pursued to transform those environments into standard languages, running on a standardized platform. As this program evolves, a service-supported approach could be deployed to further consolidate and transform these migrated environments into common data and application architectures.

Executives should not jump into an approach randomly or based on personal relationships with a given firm. Rather, services strategies should follow a consistent philosophy that makes good sense for the business.

13.9.1 Service Provider NightOwl Requirements

Service Provider NightOwl has the following services supported by the network infrastructure:

Broadband users accessing data in the Internet

Enterprise VPN customer with managed CPE and Gold, Silver, Bronze, and Best-Effort traffic classes

Enterprise VPN customers with unmanaged CPE with only Best-Effort traffic class

NightOwl has a modest network infrastructure as follows:

The network spans four regions: central, north, south, and east

Twelve core POP sites with 24 routers partially meshed with high-speed POS interfaces, three core sites in each region

Thirty-two PE routers for enterprise customer connectivity, both managed and unmanaged, eight PEs in each region

Forty-four edge routers aggregating broadband users, 20 routers in central, eight in each of the remaining three regions

Four Internet Gateway routers (IGW) in a single central location for domestic and international peering

4.2.4 Class of Service-Related Attributes (Queuing and Scheduling Related to QoS)

Service providers may offer on their CENs different classes of service (CoS) to subscribers defined by various CoS identifiers. Based on a white-paper titled Metro Ethernet Services—A Technical Overview by Ralph Santitoro, published by the MEF in 2003, the CoS identifiers include the following:

Physical port—this is the simplest form of CoS that applies to the physical port of the UNI connection. All traffic that enters and exits the port receives the same CoS. It is also the least flexible. If the subscriber has different services that require different CoS then they would be required to purchase separate physical ports for each service with different CoS and that would be very expensive.

CE—VLAN ID–based CoS (IEEE 802.1Q Priority Code Point)—this is a very practical way of assigning CoS if the subscriber has different services on the physical port where a service is defined by a VLAN ID. We have discussed this in Chapter 3, Section 3.1, and the VLAN tag format is shown in Fig. 3.7. The 802.1p p-bit values in the IEEE 802.1Q VLAN tag (Fig. 3.7) allows the carrier to assign up to eight different levels of priorities to the customer traffic. Generally, a service provider assigns certain performance parameters such as frame delay, frame delay variation, frame loss, and availability to each CoS level. This bundling of CoS and performance parameters together with bandwidth profile defines QoS. Ethernet switches use this CoS field to specify some basic forwarding priorities, for example, frames with priority number 7 get forwarded ahead of frames with priority number 6, and so on. This is one method that can be used to differentiate between VoIP traffic and regular traffic or between high-priority and best-effort traffic. In all practicality, service providers are unlikely to exceed three levels of priority for the sake of manageability.

IP ToS/DiffServ—the CoS may also use additional behaviors assigned by IP ToS or DiffServ applied to frames. The IP ToS field is a 3-bit field inside the IP packet that is in Ethernet frame’s payload and used to provide eight different classes of service known as IP precedence. This field is similar to the PCP field if used for basic forwarding priorities; however, it is located inside the IP header rather than the Ethernet 802.1Q tag. DiffServ has defined a more sophisticated CoS scheme than the simple forwarding priority scheme defined by IP ToS. DiffServ allows for 64 different CoS values called DiffServ code points (DSCPs). Although DiffServ gives much more flexibility to configure CoS parameters, service providers are still constrained with the issue of manageability. Beyond that, the overhead of maintaining these services and the SLAs associated with them becomes cost prohibitive. Unlike CE, VLAN CoS, IP ToS, and DiffServ require network switch to inspect IP packet header in payload of the Ethernet frame to determine ToS or DSCP value. Generally Ethernet switches support this capability. Because both IP ToS and DiffServ are applied to IP packets and not Ethernet frames, additional details are out of scope for this book.

In MEF 10.3 published in 2013, the CoS identifiers are classified as

EVC Based—each ingress service frame mapped to a given EVC has a single class of service identifier. The class of service identifier can be determined from inspection of the content of the ingress service frame. When class of service identifier is based on EVC, all ingress data service frames mapped to the EVC must map to the same class of service at the given UNI associated with that EVC.

PCP based—when the class of service identifier is based on the priority code point (PCP) field, the CE-VLAN CoS must determine the class of service name. PCP was described in Chapter 3 and is the field that refers to CE-VLAN CoS in the customer VLAN tag in a tagged service frame.

Internet Protocol based—when the class of service identifier is based on Internet Protocol, the class of service identifier is determined from the DSCP for a data service frame carrying an IPv4 or an IPv6 packet.