Managing Diverse Infrastructures

Chapter 7: Managing Diverse Infrastructures

• Before the emergence of the commercial Internet in the 1990s, companies relied on proprietary technologies installed/managed inside each firm. This approach was expensive and unsatisfactory, because:
• To reach business partners/customers, every company had to develop its own communication infrastructure, resulting in massive duplication in infrastructure investment and confusing multiple technologies.
• The technologies did not interoperate well. Complex software programs were necessary just to serve as a bridge between incompatible systems.
• Reliance on proprietary technologies meant that companies were locked into specific vendor technologies and were at the mercy of these vendors.
• Companies that installed hardware/software from many vendors suffered performance and reliability difficulties.
• IT managers were blamed by business managers for delivering expensive systems that performed poorly or, worse, never worked at all.
• The accessible public Internet based on open standards has changed the way companies build IT capabilities, and companies now gain leverage from their connections to public infrastructure:
• Companies can share a communication infrastructure common to all business partners and customers. Customers and business partners can interact via common interfaces (usually Web browsers) in seamless interaction that dramatically reduces complexity and confusion.
• The open Transmission Control Protocol/Internet Protocol (TCP/IP) standard enables communication technologies to interoperate well. Software that bridges systems is simple, standardized, inexpensive, and sometimes even free.
•  Companies are much less locked into specific vendor technologies, thus creating more competition among vendors, lower prices, and better-performing technology.
• Companies can combine technologies from numerous vendors and expect them to interconnect seamlessly.
• Reliable, secure connections to public networks provide new options for delivering IT services:
• Services historically provided by IT departments can be acquired from service providers.
• Firms can obtain smaller increments of service from outside vendors, with shorter lead times and contract duration.
• Incremental service delivery (outsourcing limited amounts of service) is easier to manage, because it represents a series of small steps toward improvement, rather tha a few large “all-or-nothing” steps.
• Incremental service delivery facilitates experimentation and learning.
• Incremental service delivery also means that IT managers now must manage relationships and share information with service partners.
• Increasingly, cell phones and PDAs are the tools used to interact with IT systems to conduct business.

 

• New Service Models
• With the advent of reliable, high-capacity networks local software execution and data storage are no longer the only alternative or the best alternative from a business standpoint.
• Software is now designed to operate in geographically distant facilities that belong to service providers who deliver similar services to many customers.
• Service delivery scenarios:
• The end-user’s company owns little of the infrastructure involved in service delivery and pays a monthly fee for a service bundle.
• Components of the infrastructure that delivers IT services may be outsourced—a company could rent space for its data center in a vendor-owned hosting facility or employ a specialized outside firm to monitor its intrusion detection systems and guard against sophisticated new security threats.
• Benefits of incremental outsourcing:
• Managing the shortage of specialized IT workers—by reducing the need for internal IT staff. Particularly helpful to small and medium-sized businesses.
• Reduced time to market—helps in developing new capabilities quickly.
• The shift to 24X operations—vendors can achieve economies of scale that justify large investments in redundant infrastructure needed for real-time operations. Vendors can invest in levels of availability/security that individual firms cannot provide.
• Favorable cash flow profiles—Firms pay a monthly fee to acquire services equivalent to those provided by internal systems in the past. With limited upfront purchases, payback flows in more quickly. Compare this to traditional IT investments, which required large up-front cash outlays that only yielded deferred and often uncertain benefits, due to high project failure rates.
• Cost reduction in IT service chains—centralized service delivery reduces support costs. Upgrades to new versions of software are done centrally, eliminating the need for support personnel to upgrade individual client computers. Since software is never physically distributed, risk (and costs) of software piracy are reduced. Vendors realize savings from economies of scale in staffing which they can pass to customers in the form of reduced prices.
• Making applications globally accessible—services are available at any computer with a Web browser for any user with the authority to access the service. The cost of moving a worker from one location to another is eliminated.
• Web Services--examples
• Human resources benefits administration—vendors can supply benefits expertise
• Employee retirement investment plan management—online vendor websites can provide guidance.
• Salesforce.com—example of an application service provider, from which companies can rent software functionality for a monthly fee and access the functionality via a Web browser.
• Impacts of Web services model:
• Allows for highly dynamic provision of IT services.
• Rather than a long-term relationship with specific service providers, firms using web services negotiate/procure services in real time from an ever-changing market composed of companies offering those services.
• These technologies are being used inside firms to improve efficiency in using computing assets, through on demand, utility computing, and grid computing models.

1. On Demand, Utility, and Grid Computing Models

• Definitions:
• Financial models that make IT services easier and less risky to procure and manage, as well as contracting models based on management of service levels.
• Restructuring and reengineering of existing applications to make them easier to manage and use.
• Enhancements to infrastructure to improve interoperability and efficiency in use of computing assets.
• Financial models that underlie on demand, utility or grid models can be…
•  simple—when an IT vendor makes equipment available to a customer via favorable leasing arrangements, rather than traditional purchases.
• sophisticated—customers can contract for a variable amount of computer power or storage capacity or both, to be provided by a vendor for a fee that varies in proportion to the amount of power or capacity used. Customer saves money by only paying for surge capacity when it is needed.
• Applications often have to be restructured:
• Centralized and communized in a central data center and accessed from anywhere.
• Large-scale reengineering of business processes and rewriting or replacing applications with off-the-shelf packages.
• Replacement of exiting applications with ASPSs (application service providers) or Web services.
•  At the infrastructure level, companies can make access to IT services like electricity or telephone. For example, server capacity or disk storage capability can be managed as a fluid resource and reallocated to handle surges or to create efficiencies from sharing equipment.
• To support this scenario, middleware must address issues such as:
• Provisioning—providing access to new services or additional capacity in an automated manner.
• Resource virtualization—server or storage capacity can be accessed and referenced independent of its physical characteristics and location.
• Change management-permits centralized changes to infrastructure, to reduce the cost of making changes and to exert additional levels of control over processes critical to maintaining high availability.
• Performance monitoring and analytics—allowing constant evaluation of the performance of computing infrastructure and suggest improvement to performance or return on investment.

 

B. Managing Risk through Incremental Outsourcing

• Which services should be outsourced? See decision-making steps in Figure 7.3 (p. 339)
• IT services that are unique to a company and provide it with significant advantages over competitors tend not to be outsourced, unless a company cannot develop a vital capability internally.
• For commodity-like services like email and word processing software, which do not provide a competitive advantage, reliability and lost cost can be obtained from vendors.
• Incremental outsourcing, as compared to outsourcing of large segments of a company’s IT, does not create a lot of risk, and mistakes cost less.
• Incremental outsourcing offers new, attractive choices to managers seeking to improve IT infrastructure. In the past, they faced two choices: 1) do nothing and risk slipping behind competitors, or 2) wholesale replacement of major components of computing infrastructure, with the risk of huge cost overruns and potential business disruptions due to implementation failures.
• Now, with incremental outsourcing, managers can capture significant economic benefits without putting the entire future of the firm at stake.

2. An Incremental Outsourcing Example: Hosting

Outsource hosting of a company’s systems involves deciding where they should be located physically. Managers can decide how much of company IT will be turned over to the vendor.

• The Hosting Service Provider Industry—
• Hosting companies own and manage the facilities that house computers that provide over-the-Net services.
• Benefits of outsourcing hosting—downtime is reduced by 87% and costs by 80-90%.
• Service Levels in Hosting—
• Colocation hosting—customers rent floor space, connectivity, and power. Customer owns/is responsible for design, maintenance, and operation of all the servers and equipment. This model supports a wide range of architectural possibilities and offers high availability.
• Shared hosting—servers are owned and operated by the hosting provider and customers purchase space on servers. Multiple customers share a single physical server.
• Dedicated hosting—servers are owned and operated by the hosting provider, but customers do not share servers. Servers are dedicated to individual customers. Complete managed services packages are offered to run the customer’s systems at the required level of security and availability.

C. Managing Relationships with Service Providers
Choosing reliable service providers and managing strong vendor relationships are critical skills for an IT manager.
1. Selecting Service Providers

• An RFP (request for proposal), submitted to a set of apparently qualified vendors, includes:

• Descriptive information
• Financial information
• Proposed plan for meeting service requirements
• Mitigation of critical risks
• Service guarantees
• Pricing

Information from industry analysts, other companies, and visits to service provider sites are also used in deciding on a service provider.
Rule: An outsourcing deal that is too one-sided, too favorable to one party at the expense of the other, usually ends up as a bad deal for both sides.

• Relationship Management
• Problem-tracking and customer relationship systems must be able to exchange problem-tracking and customer account info.
• Procedures and technical interfaces between partner systems must be properly designed and maintained.
• Service-level agreement must align incentives in relationships with service providers.
• Describe the specific conditions by which the service provider is held liable for a service interruption and the penalties it will incur as a result.
• SLAs should have “teeth” so that service providers feel pain when failures occur—but, determining what penalties should be is difficult, and the most successful relationships emphasize shared objective and helping all the partners earn a reasonable return.
• SLAs also need to contain provisions about a customer’s rights to control its own data. 

D. Managing Legacies

• Company operations are often constrained by the ways legacy systems process information.
• Business managers often do not even know how their company performs certain vital business functions because the details are buried in how legacy systems operate.
• Difficulties that arise from legacy systems:
• Technology problems—incompatibilities
• Residual process complexity—systems address problems that no longer exist. Ex: batch processing is still used, but because of today’s computing power/bandwidth, it is not necessary.
• Local adaptation—legacy systems were built for very focused business purposes and do not facilitate today’s global uniform parts management.
• Nonstandard data definitions—differences in data definitions are built into a company’s IT infrastructure in many specific places and are difficult to change, because making definitions common requires an expensive and difficult to achieve companywide consensus.
• There is no prescriptive approach to dealing with legacy issues, and tactics for solving these problems must fit individual companies and their specific situations.
• Table 7.5 lists questions managers should consider regarding legacy systems that will have to interact with a new system or service. Basically:
• Can legacy systems in any modified or enhanced form, perform consistently with real-time infrastructure objectives?
• If no, replacement may be the only option.
• Or, enterprise application integration can be used, in which interfaces are used to facilitate interaction between new and old infrastructure elements. Question: are they sustainable, and do they represent reasonable cost/functionality tradeoffs?
• Enterprise Application Integration practitioners recommend:
  -non-invasive interfaces that minimize changes to the internal operations of legacy systems.

-attention to organizational aspects of legacy systems, like the rigidity of workers’ attachment to how a system works. The degree to which systems force cultural and process changes is a key management decision.

E. Managing IT Infrastructure Assets

• Since the emergence of PCs, client servers, the Web, portable devices, and distributed network infrastructure, a company investment in IT is diffuse and scattered—hard to keep track of and manage.
• The variety of asset configurations makes it hard to answer questions like:
• How are IT investments deployed across business lines or units?
• How are IT assets being used?
• How can we adjust their deployment to create more value?
• Total Cost Ownership Analysis (TCO) analyzes IT services in terms of the costs and benefits associated with service delivery to each client device. Arriving at this number requires a detailed study to determine the total monthly costs associated with the delivery of each service available on that client, including costs shared with other clients and costs not necessarily accounted for as line items in budgets or accounting systems. Once monthly costs are computed, they must be allocated on a per client basis.
• Usage information on a per client, per month basis can also be helpful when compared with costs of service delivery on a per client, per month basis.

 
Questions for Discussion:

• Discuss the reasons why relying on proprietary technology install/managed inside each firm, as was done before the emergence of the commercial Internet in the 1990s, was expensive and unsatisfactory.

 

• Discuss how the accessible public Internet, with its open standards, has changed the way companies build IT capabilities and given them leverage.

• Discuss the concept of incremental service delivery and the several benefits its offers.

 

• Discuss the ways applications often have to be restructured when incremental service delivery is used.

• What are the issues middleware must address when companies decide to make access to IT services like access to traditional utilities like electricity and telephone?

 

• In the past, what were the two choices (discussed in the text) that managers had in seeking to improve IT infrastructure? How has incremental outsourcing impacted on this decision? What factors do companies take into consideration in deciding which services should be outsourced?

• Define and discuss the three service levels of outsource hosting.

 

• What information should an RFP contain? In addition, what other information do managers use in choosing an outsource vendor?

• What are the major difficulties that arise from legacy systems? What suggestions do Enterprise Application Integration practitioners give about changing legacy systems?

 

• What is the purpose of Total Cost Ownership Analysis, and how is it done?