build & planing

Theory Assessment
Plan and Build Computer Network (1A)

ICAITS029B Install network hardware to a network
ICAITS021C Connect internal hardware components
ICAITS030B Install software to networked computers
ICAITS116B Undertake capacity planning


Time Allowed: 3 hours
Pass Mark: 15
(Assessor’s Copy)
NOTE: the theory focus on capacity planning


Table of Contents


Question 1: What is meant by capacity planning in information technology? 2
Question 2: What are some of the factors that capacity planning needs to consider? 2
Question 3: Briefly describe the following terms 3
Question 4: What are the stages in capacity planning? 4
Question 5: Describe some of the commerce site metrics in capacity planning for a business that is moving into e-commerce. 4
Question 6: Describe some of the differences in a corporate site as against a high-end e-commerce site that directs the activities for capacity planning 6
Question 7: Describe the steps in capacity planning for your organization that has moved onto e-commerce utilization of computer systems 7
Question 8: Describe the difference between Tier 1, 2 and 3 system configurations for capacity planning for an organization that has moved onto e-commerce utilization of its computer systems. 7
Question 9: Discuss the term “persistence: in capacity planning 9











Question Marks
1 2
2 1
3 4
4 1
5 5
6 4
7 4
8 5
9 4
Total Marks 30


Table 1 Assessment Grid
UNIT Theory/Project Practical/Project
ICAITS029B (All elements) X
ICAITS021C (All elements) X
ICAITS030B (All elements) X
ICAITS116B (All elements) X
Marks 30 70
Grade : < 50 Not Yet Competent, 50 - 72 competent, 73 – 82 competent with credit, 83+ competent with distinction



Question 1: What is meant by capacity planning in information technology?

Answer Suggestion:


Capacity Planning is the activity relating to the management of the information communication infrastructure to ensure that the output of the facility is maximized and the input is minimized.

The activity involves the use of tools to measure inputs and outputs that may be competing. Constraints come in the form of labour, material availability, equipment utilization and limited space. Planning, the use of mathematical models and flexibility in resource utilization is needed
Capacity planning in information communication technology systems is needed when:
• Starting a new system, or a service organization
• Extending the operations of an existing business system.
• Considering additions/modifications to the system
• Introducing new techniques, equipment and materials



Question 2: What are some of the factors that capacity planning needs to consider?


Answer Suggestion:
• Bottlenecks that will determine the capacity for any particular stage of the process.
• Peak demand rate
• Communication Bandwidth
• Load balance within the system
• Existing Hardware and scalability
• Cost
• Technology changes in hardware and software


Question 3: Briefly describe the following terms

a. Design Capacity

Answer Suggestion:

• Target output level
• Note: It could be reached or exceeded in practice.
• It is the agreed % between vendor contractor and customer of the system

b. Effective Capacity

Answer Suggestion:

• It is the actual capacity to reflect current conditions.
• Note: could be less than or more than design capacity.
• Required because of; limited resources, increased equipment breakdowns when run at higher capacity for long periods and the impact of the environment.


c. Utilization

Answer Suggestion:
• Measures use of resources i.e actual running time.
• Measure as % of effective capacity.
• Note: measurement is often different between organizations and systems.

d. Yield

Answer Suggestion:
• Measures output of product in % against equivalent, potential or maximum products or systems’ input.
• Yield is concerned with avoidable and unavoidable losses due to downtime as a result of system or equipment problems.


Question 4: What are the stages in capacity planning?

Answer Suggestion
1. Analyse existing system capacity
2. Determine future capacity requirements
3. Plan for capacity enhancement
4. Install capacity enhancements




Question 5: Describe some of the commerce site metrics in capacity planning for a business that is moving into e-commerce.

Answer Suggestion


a. Hits: Results of multiple hits from the browser to the Web server

A hit is any request from the browser to the Web server this includes requests for dynamic content, images, or streamed video

b. Average hits per page: The average number of Web server hits per page.

The suggested configurations used stores with 1 hit per page. The test tool did not load any images, nor did it use frames. The test tool only made one HTTP request per page.

c. Commerce/non-commerce page ratio

Most pages are generated by issuing a database query and rendering the results by using JSP (Java Server Page) templates.

Typical stores also contain some static content, such as HTML page requests. The suggested configurations did not contain any requests for purely static pages. Static pages are served at roughly the same rate as static images.

Requests for static pages can be ignored in capacity planning, if the ratio of dynamic requests to static requests is fairly high. A site which serves two dynamic pages for every static page can be sized without regard for the static pages.

This is because dynamic pages require substantially more resources to serve.


d. Database size

The database size affects performance of the system. Sites that include many shoppers, merchants, or catalogue items usually cannot serve requests as quickly as those that do not.

Various factors will affect the throughput of the system, including database tuning, hardware capacity, and physical topology.


e. Megabytes of data that are served per unit of time

It is important to plan your site to have sufficient resources for serving expected network throughput. This includes sizing all firewall, routing, and other network components to be able to handle the peak workload expected of your site. The suggested configurations can be measured using a full duplex 100 Mbps Ethernet between the load generator and Web server.

A 100 Mbps switch can be used between the Web server and database server.

It must be noted that added equipment results in the Internet communication being significantly slower.

If the site is running web casts or real time audio, this will demand bandwidth. This has to be taken into account in the overall capacity planning solution.


f. Performance Test Scripts

The commands executed affect system workload.


Te set up and the commands run as part of performance tests can be like the following.

Components:

A basic store that contains a set of use cases commonly used in shopping model:

– New shopper registration
– Registered shopper logon
– Browse categories, products, and items
– Add item to shopping cart
– View shopping cart with tax and shipping details
– Modify quantity of items in the shopping cart
– Delete items from the shopping cart
– Checkout by entering billing and shipping information


A script to generate the following:

– Up to 5 000 categories
– Up to 50 000 items
– Up to 50 000 registered customers
– Up to 50 000 orders

A script to simulate various shopper scenarios
– Browse
– Add
– Prepare
– Buy

A script to define the regular shopping mix

The regular shopping mix depends on two variables:
1. Customer type
New registered customer
Existing registered customer
Guest shopper

2. Shopping flow

Browse

Visit the home page

Logon using registered customer id (existing registered customer only) or register as a new customer (new registered customer only

Select “Multiple Category Display” then “Multiple Product Display”

Browse and Add

Browse, and then randomly add one, two, or three items to the shopping cart.

Browse and Prepare

Browse and add, then enter address book information (guest shopper only) or specify which address to use (new registered customer or existing registered customer).

Browse and Buy

Browse and prepare, then display order information. Enter a credit card number with expiry date, finally submit the order.

Question 6: Describe some of the differences in a corporate site as against a high-end e-commerce site that directs the activities for capacity planning


Answer Suggestion
Typical corporate site High-end commerce site
Page Views Approximately 15,000 More than 50,000
Connection (Note 1) Single DSI Multiple DS3s
Back-end integration Optional Mandatory: inventory, logistics, and sales demand integration
Availability OK – Networks rarely fully redundant 24x7x365 – Network has no single point of failure
Scalability Poor – little content monitoring Good – Local and geographic mirrors
Impact of downtime Embarrassment, inconvenience Financial loss, customer loss
Source: Forrester Research, Inc.
Note 1: IEEE 802.6, Metropolitan Area Network, DQDB Media Access Control (PHYSICAL LAYER: Interfaces between network medium and network devices; Defines electrical and mechanical characteristics)

Question 7: Describe the steps in capacity planning for your organization that has moved onto e-commerce utilization of computer systems

Answer Suggestion

Steps

1: Complete a worksheet with calculated system capacity metrics.
2: Select the number of server tiers
3: Determine requirements for availability
4: Select hardware
5: Configure the appropriate number of disk drives
6: Configure memory for each system
7: Configure the network requirements
8: Is a separate Payment Manager server required?
9: Is a separate server required for search?
10: Will a Staging server be used?

Reference: IBM® WebSphere® Commerce Suite, Capacity Planning Guide, Version 5.1
Question 8: Describe the difference between Tier 1, 2 and 3 system configurations for capacity planning for an organization that has moved onto e-commerce utilization of its computer systems.

Answer Suggestion


1–Tier
In this configuration the Commerce Server, the database server, and the Web server are all on the same machine.




2-Tier
In this configuration the Web server and Commerce Suite is on one machine, the database server is a separate machine



3-Tier
In this configuration the Web server, Commerce Suite and the database are on separate machines







Question 9: Discuss the term “persistence: in capacity planning

Answer Suggestion
Persistence is a term that comes into play when using a Controller to load balance multiple servers at a site which is designed to have customers return to the same server during a session. Persistence is required to seamlessly process customer requests while keeping the link between your site and your customer unbroken. However, E-commerce poses its own special challenges for persistence, specifically in regard to scalability and performance of the load balancing device.



No Persistence No SSL Persistence
No Persistence: Left example shows the effects of a customer source IP address changing over the course of a session. Customer accesses server (A), and fills a shopping cart full of goods. Customer decides to visit other areas of the site before purchasing. Customer returns to shopping cart and finds it apparently empty because customer was accidentally rerouted to the wrong server (B).
No SSL Persistence: Example on right shows the results of IP users being funneled into very few IP addresses. Four customers log on to AOL (An ISP) and are directed to a Web site. All end up concentrated in a single server (A), giving customers sluggish site performance while other available resources (servers B, C and D) sit idle.

For example, a customer uses regular HTTP to select items from a catalog, and places them into a shopping cart. When the customer wishes to purchase the items, however, most sites use SSL – a secure, encrypted protocol required for transmitting sensitive information like credit card numbers. In some cases, site design dictates that customer traffic for both HTTP and SSL be sent to the same server in order to keep the buying process seamless. This can be handled by persistence settings available on Controllers

Example:
BIG-IP Controller has SSL (Secure Socket Layer) Persistence. It tracks persistence using SSL Session ID, and allows BIG-IP Controller to properly load balance SSL traffic when it is originating from a site using proxy servers or multiple firewalls. These types of traffic sources concentrate a large number of users into a small number of IP addresses, or cause the end users’ IP address to change over the course of the session. Both of these scenarios cause standard persistence based on source IP to "break". In the case of IP address concentration, source IP persistence will concentrate the load on a single server, giving end users poor performance while other available resources sit idle. When a user’s IP address changes over the course of a session, source IP persistence will no longer be able to associate the user with the server that she was sent to, thereby causing things like shopping carts to appear empty. SSL Session ID Persistence solves these problems.