An Enterprise In Implementing Private Clouds

Published Date: 02 Nov 2017

Abstract: In today’s world where companies are growing at a rapid pace, adopting Information Technology practices has become the norm to simplify the processes and operations around the business. As a result, increasing headcount also implies increased IT infrastructure need like buying new computers, purchasing software, software licences, tools, etc. thereby pushing the company’s IT spending upwards. In the recent past, one of the most cited alternatives for reducing IT cost is building systems that allow applications to be loaded and maintained centrally with high availability over a network. The technology that enables such a centrally hosted program with minimal maintenance overhead and low infrastructure costs is called Cloud Computing and is often times quoted as a universal solution for cost reduction in IT spending.

Cloud computing is defined in technical terms as "a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (for example, networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service-provider interaction". (1)The motivation for today’s enterprises to opt for a cloud computing solution lies in the fact that most of the IT resources are a huge one time investment and the frequency of use is not the same throughout the year. Some applications/systems will be used continuously where as many are used very rarely. A typical example is the use of employee tax calculation systems which is done once in a year, but the company will have to buy the software and licence the same for the entire year. But the maintenance of these systems must be done every year, irrespective of the use. This demands a solution that is elastic, pay per use and has cost effective or zero maintenance. Cloud computing is the solution that quenches all these needs.

The advantages provided by Cloud computing include on-demand selling, billing per use on hour or minute basis, management by a cloud service provider, who is either internal or third party vendor, among others, and the client or the user needs to have a simple personal computer connected to a network. These factors have made Cloud computing to be embraced by large and small enterprises equally. These capabilities also result in faster Time-to-market for the enterprise IT services.

Since the area of cloud computing is recent, there are many open challenges which are a cause of concern in leveraging the benefits that cloud computing provides. Security, reliability and performance, vendor lock in, maturity of solutions are the important ones. (2) Though most of these are faced by public clouds, private clouds also do not provide a convincing method to address them. The economics of establishing a cloud data centre by an enterprise is not only determined by the cost of the hardware/software required at the data centre, but also on infrastructural costs like power, facilities and maintenance, network, cooling and real estate costs which play a vital role in the total ownership cost of the cloud computing system. (3) Choosing the best of the available resources to address these costs can push down the total cost of ownership significantly.

In this paper, we elaborate on the costing methods for cloud computing and provide a detailed economic formulation for optimal cost incurred based on the factors like proximity of the data centre to the enterprise, the climatic conditions of the city where the data centre resides, the power costs in different geographies and the alternate sources of power that are available at the nearest destination. We also provide details of the cost workout by illustrating a fictitious example of a midsized enterprise that houses a private cloud solution. Thus, we hope to burst the myth that hosting a private cloud is not the best option as compared to public cloud for an enterprise. We emphasise that considering the right or optimal values of these factors are vital and once chosen and the cost worked out, it makes cloud computing a better alternative to traditional ways of enterprise computing.

Keywords: Cloud Computing, Optimal Costing Model, Private Clouds, Cost Estimation, Enterprise Information Technology.

Introduction:

This section gives a brief introduction to the cloud system architecture, especially the private cloud, and the costing model:

Introduction to the cloud architecture: As defined in the abstract, cloud computing is the new IT model to enable elastic, on-demand provisioning of computing resources which are centrally managed. Figure 1 describes the bird’s eye view of the cloud computing system in contrast with the traditional computing systems shown in Figure 2.

E:\Supreeth\College\Research\Cloud\Infra Provisioning Cloud.png

Figure 1 – Cloud Computing Infrastructure Provisioning (2)

E:\Supreeth\College\Research\Cloud\Infra Provisioning Traditional.png

Figure 2 – Traditional Infrastructure Provisioning (2)

Broadly, there are three types of cloud computing solutions found in today’s enterprise – Public Clouds, Private Clouds and Hybrid Clouds. Public Clouds are hosted by a third party service providers like Google, Microsoft, SalesForce.com, Amazon Web Services, etc. Private clouds are hosted and maintained internally by an enterprise. Hybrid clouds include some part of both public and private clouds. There are three kinds of services offered by each of these types of cloud as below (4)

Infrastructure as a Service (IaaS) – This is the most basic form where it’s a server or servers out there in the cloud, or a bunch of storage capacity or bandwidth. IaaS customers, which are often tech companies, typically have a lot of IT expertise; they want access to computing power but don’t want to be responsible for installing or maintaining it.

Platform as a Service (PaaS) – This is a cloud-based platform that companies can use to develop their custom applications or write software that integrates with existing applications. PaaS environments come equipped with software development technologies like Java, .NET, Python, and Ruby on Rails and allow customers to start writing code quickly. Once the code is ready, the vendor hosts it and makes it widely available. PaaS is currently the smallest segment of the cloud computing market and is often used by established companies looking to outsource a piece of their infrastructure.

Software as a Service (SaaS) – It is the largest and most mature part of the cloud. It’s an application or suite of applications that resides in the cloud instead of on a user’s hard drive or in a data centre. One of the earliest SaaS successes was Salesforce.com’s customer relationship management software, which provided an alternative to on-premise CRM systems. More recently, productivity and collaboration software—spreadsheets, word processing programs, and so on—has moved into the cloud with Google Apps, Microsoft Office 365, and other similar offerings.

Figure 3 shows the above cloud service models working together along with the major commercial players in each of the category.

E:\Supreeth\College\Research\Cloud\Cloud Service Models.png

Figure 3: Cloud Computing Service Models (5)

Introduction to Costing of Cloud: Costing is the process of estimating costs for items in a program. The result of costing is a budget for the program. The development and maintenance of a cloud involves a lot of cost which the company has to bear for the installation of the cloud setup. The different costs that form the cost units include infrastructural costs like power, facilities and maintenance, network, cooling, real estate, green taxes, etc.; Platform costs like procurement cost, development cost, operational cost, maintenance cost of data centres; One time software costs like purchasing cost of software, software licences fee, etc.

Cost Type

Description

Cost Factors

Strategic decision, selection of Cloud Computing Services and Cloud Types

Strategic decision on sourcing a Cloud Computing Service: as-is analysis of the IT infrastructure and business applications, analysis of performance indicators, application of decision tools; choice of Cloud Computing Service type (IaaS, PaaS, SaaS or combinations); choice of Cloud type (Public, Private or Hybrid Cloud); definition of service requirements (as e. g. hardware configuration for IaaS, programming language support for PaaS and functionalities for SaaS).

expenditure of time, consulting services, information for decision-making

Evaluation and Selection of Service Provider

Search process for providers offering the desired service based on the previously defined requirements.

Service evaluation and analysis: evaluation of the functionalities of Cloud Computing Services; Identification of the best alternative.

Evaluation of the provider and SLA analysis: determining the provider’s reputation, analysis of the SLAs (quality of service) and of the security requirements (e. g. data recovery)

expenditure of time, consulting services, information for decision-making

Service Charge

Pricing schemes vary depending on the service type and the provider. The service charge can be calculated on the basis of the pricing schemes.

IaaS: computing power, storage capacity, inbound data transfer, outbound data transfer, provider internal data transfer, number of queries, domain, Secure Socket Layer (SSL) certificate, licence, basic service charge.

PaaS: user-dependent basic charges, storage capacity, inbound data transfer, outbound data transfer, provider internal data transfer, extra user data storage capacity, extra user document storage capacity, queries to the Application Programming Interface, sent emails, database, secured logins, connections with other providers’ applications.

SaaS: access to the service System, user.

Implementation, Configuration,

Integration and Migration

Implementation and configuration of the service, including, for example, access authorizations (creating groups and users including their specific rights).

Integration into or merging with other systems and business processes. This includes the option of merging two Clouds into a hybrid Cloud.

Migration of the system (porting of data)

Expenditure of time, porting process.

Support

Phone, email and ticket support and/or support via chat (instant messaging)

Expenditure of time, support costs, problem solving.

Initial and permanent training

Internal (by own employees) or external training (by third-party providers): User training and administrative training

preparation time of internal employees, participating time of internal employees, instruction material, external consulting services

Maintenance and Modification

Modifying the service to guarantee operability

Testing the service operability; configuration of settings; tariff changes.

Monitoring and Reporting: Performance and Cost management

Service Level Management: testing whether the provider fulfils contractual obligations (aspects of service quality, as e. g. availability)

Expenditure of time.

System Failure

Lost working time

Contract penalty for non-delivery of services

Loss of reputation

loss per period

Back sourcing or Discarding

Porting of data from the Cloud

Reestablishment

expenditure of time, porting process

Table 1 – Costs, description and the factors affecting cloud implementation lifecycle (1)

Objectives of the Study:

The purpose of this study is to

1. To understand the concept of Cloud Computing.

2. To build an efficient evaluation formula for calculating the cost of cloud.

3. To propose an optimal costing model for setting-up the cloud.

Theoretical Background:

There have been many works done in the past to establish a relationship between these above mentioned costs and their significance in setting-up and maintaining the cloud. To provide theoretical evidences to support our hypothesis, we applied the method of a concept-centric systematic literature review. As a first step we defined the review scope on keywords like Cloud Computing, Optimal Costing Model, Private Clouds, Cost Estimation, and Enterprise Information Technology. Next we applied these keywords to scientific databases like IEEE Xplore, EBSCO and Springer to receive scientific, peer-reviewed papers. To amplify the number of papers, we used forward (review of reference lists) and backward search (author-centric review).

The opportunities that a cloud can find in a business enterprise are enormous since the businesses are moving towards the next evolutionary step of elastic IT. Cloud computing provides tremendous potential for optimizing IT services to businesses around the globe and across myriad domains. However, it is also important for a business to seek next-generation management solutions that can aid in the proper adoption and expanded use of the cloud computing model. (5) Cloud allows technology to be accessed as a utility, creating a lean and mean virtual enterprise positioned wherever customers need to be served in the marketplace. This ability to "plug in" to IT services opens up transformational opportunities. (6) With these opportunities, Cloud computing changes the economics from a capital investment to a pay-as-you-go model and it’s becoming more ingrained in businesses today. Cloud computing solutions also offer a large number of business benefits that enables the business in reducing expenses, better business support, automation, better financial management, reaching the markets faster, mobile business, continuity in operations and also easy to install the technology. (7) Making use of cloud computing effectively and efficiently in a business can not only increase profits for a company by allowing fewer employees to work remotely, but can also increase the productivity of a company. Employees no longer need to wait for its members to gather to work on a single project, rather they can connect to the cloud via the Internet to work from wherever and whenever, while, still remaining up to date with their project partners thus improving the value of the business. (8) With all these advantages the enterprises have the option to choose between the available types of clouds whether it is private, public or hybrid. Public clouds have its own benefits as they provide faster time to market, on demand elastic infrastructure and also pay as you go. But they possess a lot of drawbacks with regard to security, reliability and performance, vendor lock-in, inability to leverage existing investments, complex corporate governance and auditing, lack of maturity of the solutions etc. Enterprise private clouds have an upper hand over the public clouds as they provide automation, management and monitoring, virtualization, etc. Thus enterprise private clouds provide a good opportunity to get started with cloud computing and reap the associated benefits of agility, cost savings and on-demand services while meeting the stringent enterprise security, performance and reliability requirements. (2) Thus, enterprises which have started moving towards the cloud will have to consider the following guidelines for effective implementation. These guidelines include identifying restrictions and grey areas, running experiments with SaaS, developing projects on the cloud and speaking to company’s core software vendors to understand their plans for the cloud. (4)

Optimal Costing model:

Before we actually throw light on the costing model we analyze the various costs involved in setting-up and implementing the cloud. As mentioned earlier, the various cost involved from the initial decision to the last step of discarding the cloud include Strategic Decision, Selection of Cloud Computing Services and Cloud Types, Evaluation and Selection of service provider, Service charge of IaaS, PaaS and SaaS, Implementation, Configuration, Integration and Migration, Support, Initial and Permanent training, Maintenance and Modification, System Failure and Back sourcing or Discarding.

Apart from these costs, an important cost involved that is the amortization cost. The Cloud cost also involves amortization because of different purchase prices and duration over which purchased resources are planned to be used, cost of different metrics items that cannot be put together directly or be compared fairly, etc. In such a scenario, it is extremely significant and necessary to analyze how the costs of server, power and cooling and other metrics that contributes to the monthly renting. For this purpose, the monthly depreciation cost (or amortization cost) is calculated for each metrics item from the initial purchase expense based on the duration over which the investment is amortized (which is related to its expected lifetime) and the assumed interest rate. Especially for the items contributing to operational expense like power, cooling and maintenance, amortization costs are easy to calculate based on their monthly duration pays. Typically, real-estate is comparatively depreciated over periods of 10 years and other items like server and facilities are depreciated over three years with 5% money cost. By amortization, we can get a cost amortizable rate parameter, Arp, which we can apply to both onetime purchases (server cost, facility cost and etc.) and operational expenses (power cost and coolig cost, etc.) (9)

Cost of Strategic Decision (CostSD):

The costs of strategic decisions and the selection of suitable Cloud Computing Services are made dependent on the expenditure of time (EoT) described in Equation [1] that is necessary for decision-making (expressed in monetary terms), expenses for information on which the decision may be based (EoI), as e. g. scientific literature or market studies, as well as costs of external consulting services (EoC) including choosing the service provider. The total CostSD is defined as in equation [2].

----- [1]

---- [2]

Service Cost of selected Cloud Service (CostCS):

This cost is associated with the kind of service that is desired (IaaS, PaaS or SaaS). Different pricing schemes are available for each of these services based on the type of vendor. (1)

Cost of Implementation and Facilities (CostIF):

This cost takes into account the cost incurred for real estate (CostRE), power (Costpower), facilities (Costfacilities) and cooling (Costcooling) that are required to set up the cloud hardware which are defined in equations [3] through [6].

---- [3]

----- [4]

---- [5]

---- [6]

Hence CostIF is defined as in equation [7]

---- [7]

Cost of Support (CostSupport):

The cost type "Support" depends on the costs of support services via telephone, email, ticket systems or chat. It is assumed that the customer has access to the internet for reasons other than technical support. Therefore this type of costs depends on the expenditure of time required for interactions with support personnel, as well as on occurred costs. However, some providers charge users on the basis of the time needed for problem solving and support. Generally this cost is calculated based on the number of employees involved and their compensation.

Cost of Initial and Permanent training (CostIPT):

The total costs of the cost type "initial and permanent training" can be subdivided into internal training (staff members as coaches) and external training (third party coaches from outside the company). The costs of an internal training depend on the amount of preparation time invested by one or more employees, the amount of time invested by participating employees and the costs of instruction material. The total costs of an external training can be calculated by adding the costs of consultants who organize the training, the amount of time that all employees invest in participating in the training and the costs of instruction material.

Cost of Maintenance and Modification (CostMM):

This involves the cost incurred due to changes made to the existing application or existing cloud set-up due to technical or business reasons. A major part of this cost is attributed to the Employee cost.

Cost of System Failure (CostSF):

The consequences of a system failure strongly depend on the interdependencies of services and business processes and their relevance to the business goals. Possible cost factors are, for example, loss of productive working time, contract penalties for delays or damage to the company’s reputation which is hard to evaluate. This is defined through equation [8]

---- [8]

Cost of Back-Sourcing or discarding (CostBD):

The back sourcing of the system involves costs for the porting of data from the Cloud, as well as a certain expenditure of time. However, the costs caused by the porting of data to another Cloud or by the migration to a different system are part of the Total Cost of Ownership (TCO) of the new service system, not of the TCO of the withdrawn Cloud Computing Service.

Hence the total Cost of Ownership of the cloud set-up not only involves just the hardware/software costs for cloud but also cost incurred in staffing, real estate etc. which form a major part. Each of these factors have a particular weight assigned to them and based on the importance an organization gives to these weights, the TCO is defined as below:

----- [9]

Where w1 ....... w8 >= 0

This sums up our work in determining the actual costing factors to be considered while opting for a cloud computing solution in the enterprise.

Conclusion:

In this paper we presented a novel method of calculating the total cost of ownership for a cloud solution deployed in an enterprise. Though many previous works have been in determining the cloud set-up economics, very few of them have touched on the inconspicuous costs that should be taken into account while opting for the cloud solution. The advantages provided by cloud computing technology promises to take the IT cost lower, but ignoring the one time investments and other costs as mentioned before can definitely prove to be counterproductive to the benefits provided. We hope that this paper has thrown enough light on these aspects and help people in the business to make a well informed choice before jumping to take the cloud path to solve all their IT woes.