Colleague Self-Service Hardware and Software Guide

Colleague Self-Service Hardware

and Software Configuration Guide

Colleague Self-Service 2.10+

Colleague Web API 1.10+

Last Updated May 26, 2016

Page 2 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Revision History

February 2016

Original publication.

May 2016

Updated to reflect new features of Colleague Web API 1.12 (DAS

Data Reader).

Table of Contents

Revision History ......................................................................................................................... 2

Purpose and goals ..................................................................................................................... 3

Colleague Self-Service components .......................................................................................... 3

Resource differences between Colleague WebAdvisor and Colleague Self-Service/Colleague

Web API..................................................................................................................................... 4

Software capacities and limitations ............................................................................................ 5

Load balancing........................................................................................................................... 6

Hardware resource guidelines .................................................................................................... 9

Software configuration guidelines .............................................................................................12

Determining your Colleague Self-Service deployment ...............................................................20

Colleague Self-Service deployment planning and settings worksheet .......................................27

Procedures reference................................................................................................................29

Page 3 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Purpose and goals

This guide provides detailed software background, hardware requirements, software

configuration guidance, and deployment planning guidance for Colleague Self-Service and

supporting infrastructure with the goal of providing the information needed for a successful

deployment of Colleague Self-Service.

Colleague Self-Service components

The Colleague Self-Service solution is a web application utilizing several other Colleague

components. The following diagram illustrates the major components of the “Colleague Self-

Service stack.”

Feature of API v1.12+

Web Browser

Colleague Self-Service

Colleague Application Server

Colleague Web API

DMI application listener DMI DAS listener

Page 4 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Resource differences between Colleague WebAdvisor and Colleague

Self-Service/Colleague Web API

While both Colleague WebAdvisor and Colleague Self-Service share some of the same

components and transactional processing traits, there are several differences within the

Colleague Self-Service stack that cause those resources to be utilized differently. In general,

Colleague Self-Service pages combine many Colleague WebAdvisor pages into one which

means more data access to display a given page. At the same time by moving business logic

and data processing into the Colleague Web API’s RESTful APIs, the point at which data

access occurs has been moved ‘up’ the stack. The following diagram illustrates these changes.

Web Browser

Colleague Self-Service

Colleague

Application Server

Colleague Web API

DMI

application

listener

Web Browser

Colleague WebAdvisor

Colleague Application Server

DMI application

listener

Web Form Process

Business logic /

data access

Business logic

Presentation logic

DMI DAS

listener

Feature of API

v1.12+

The resulting architecture now performs the majority of data access from a ‘higher’ point in the

stack using the same connection paths as before meaning that the point where overall data

request ‘split’ into smaller data transactions is at that ‘higher’ point in the stack. This results in

increased usage of the DMI application listener’s Connection Pool as more, smaller,

transactions rely on that connection path. The ultimate result is that this architecture needs a

DMI application listener Connection Pool size larger than Colleague WebAdvisor due to the

increased data access to render a single page and those data accesses occurring at a different

point in the software stack.

Page 5 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Software capacities and limitations

Each component within the Colleague Self-Service stack has upper limits on the number of

concurrent users or concurrent request that can be handled before performance begins to

severely degrade. Many of these capacity limitations are based on how the software is tuned

but a few are based on hardware and architecture of each component. The numbers given in

this section are based on the tuning guidance within this document with the individual

capacities/limitations of each component taking into account the other supporting components in

the Colleague Self-Service stack.

Colleague Self-Service

A single instance of Colleague Self-Service can handle around 700 concurrent users/page

requests before the web application begins to degrade in performance. Short bursts can be

tolerated but sustained load above this number will result in noticeably slower response times.

The capacity is limited primarily by the number of concurrent connections allowed to the

Colleague Web API (Colleague Web API Connection Limit setting). Colleague Self-Service is

primarily an I/O-bound application that relies on the Colleague Web API for all data I/O. As

such, the number of concurrent API requests allowed has a large impact on how long it takes to

service a page request. To a lesser degree the capacity is also limited by the number of open

sockets on the host OS as browsers and web application like Colleague Self-Service open

multiple persistent connections to facilitate the application functionality. The large number

sockets opening and closing takes away CPU cycles for running applications.

Colleague Web API

A single instance of Colleague Web API can handle around 400 concurrent API data requests.

Short bursts can be tolerated but will result in increased response times. The Colleague Web

API is a CPU-bound application as the majority of the workload is reading/invoking transactions

in Colleague and transforming that data into JSON responses. Besides CPU, the capacity of the

API is greatly limited by the number of concurrent connections to the DMI application listener

(Connection Pool Size setting) that can be used to interact with Colleague and the amount of

time the slowest transactions take. In order to help limit the amount of I/O performed against

Colleague, the Colleague Web API caches source data when possible where the source data

does not change often or at all under most operational circumstances. Even with caching in-

place, the number of concurrent connections allowed to the DMI application listener and to a

greater point, the speed of the transactions has a large influence on the overall performance.

Take registration, for example, which is traditionally one of the slowest transactions. As more

registration transactions consume the allotted concurrent connections to the DMI application

listener, all transactions (fast or slow) have to wait for the resources to free which has a

snowball effect that causes all request to slow.

Page 6 of 31

Colleague Self-Service Hardware and Software Configuration Guide

DMI application listener

A single instance of DMI with the application listener role can service a large number of

concurrent requests with relative ease and reliability but the type of transactions has a great

deal of influence on this reliability. In terms of Colleague Self-Service, the DMI application

listener is focused on a small number of request types, of which most simply pass-through to

Colleague for further processing. However, many of these requests have large response bodies

and thus consume more listener memory requiring more RAM to be allocated to the listener.

Besides listener memory, the primary limiting factor is the number of concurrent Colleague

Application Server connections (Connection Pool Size setting) since almost all transactions will

be passed-through to Colleague for servicing. In general having larger connection pools can

degrade performance and also presents a potential failure point if multiple Colleague Web APIs

are using the same listener. A Connection Pool Size above 100 should be avoided if possible. If

a larger Connection Pool Size is used the listener will require additional memory and could

become unstable under very heavy load.

Load balancing

Colleague Self-Service and portions of the supporting infrastructure can be easily load balanced

for load distribution and fail-over. As the expected number of concurrent users/request

increases load balancing becomes a requirement for successful operation of Colleague Self-

Service. Load balancing is critical during periods of heavy usage such as registration to limit

bottlenecks within the Colleague Self-Service stack. There are several ways to balance the load

amongst the components in order to provide the capacity needed. While Ellucian recommends

and supports Colleague Self-Service running in a load balanced environment, Ellucian does not

directly support or recommend any specific load balancing appliance/solution.

Page 7 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Colleague Self-Service

Load balancing Colleague Self-Service involves the following step:

- Install each instance of Colleague Self-Service following the standard installation

instructions.

- Set a static machine key for the first instance of Colleague Self-Service. This is an

additional step that is not normally performed when installing Colleague Self-Service.

See Configuring a Colleague Self-Service static machine key procedure within this guide

for steps.

- Fully configure the first instance of Colleague Self-Service and test as if you were

installing only one instance.

- Configure the remaining instance(s) of Colleague Self-Service by:

a. Set the same static machine key as the first instance. See Copying a static

machine key from one instance to another procedure within this guide for steps.

b. Duplicate the settings from the first instance of Colleague Self-Service

(connection settings, branding, theming and menu and page access). This can

be done manually in each instance or you can copy the following files from the

first instance to the others:

 App_Data\settings.config

 App_Data\sitemap.config

 Content\Theme.css

 Your logo images (typically placed in Content\Images directory)

c. Test each instance to ensure it is functional and configured as desired.

- Configure the load balancer:

a. Use a least connections or least activity balancing strategy

b. Do not use session affinity (sticky load balancing). Colleague Self-Service is

stateless and any node being balanced can handle the request as long as the

IIS/software configuration is correct.

Tips when load balancing Colleague Self-Service:

1. When making configuration changes though the Colleague Self-Service UI

(Student/Admin) pages do so using the URL of each instance and not the load balanced

URL then replicate the settings to each additional instance to ensure that all instances

are configured the same way.

2. The Unique Cookie ID setting must be the same for all instances.

Page 8 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Colleague Web API

Load balancing Colleague Web API involves the following steps:

1. Install each instance of Colleague Web API following the standard installation

instructions.

2. Fully configure the first instance of Colleague Web API and test as if you were installing

only one instance.

3. Configure the remaining instance(s) of Colleague Web API by:

a. Set the same static machine key as the first instance. See Copying a static

machine key from one instance to another procedure within this guide for steps.

b. Configure setting to be the same as the first instance. This can be done manually

in each instance or you can copy the following files from the first instance to the

others:

 App_Data\settings.config

 Your report logo(s) (typically placed in the Images directory)

c. Test each instance to ensure it is functional and configured as desired.

4. Configure the load balancer:

a. Use a least connections or least activity balancing strategy

b. Do not use session affinity (sticky load balancing). Colleague Web API is

stateless and any node being balanced can handle the request as long as the

IIS/software configuration is correct.

DMI application listener

DMI by itself cannot be load balanced in the traditional sense. Instead, multiple DMI application

listeners can be installed in scenarios where there are multiple Colleague Web APIs and

through configuration different Colleague Web APIs can share or use different DMI application

listeners. The need for additional DMI application listeners will be based on how many

Colleague Web API instances are in use and taking into account the total capacity of each DMI

application listener in terms of their Connection Pool Size. Using multiple DMI application

listeners is also a strategy for limiting interruption in the case of failure as if one DMI application

listener fails the entire set of Colleague Web API instances will not all fail along with it.

Page 9 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Hardware resource guidelines

The following resource guidelines represent the minimum required hardware based on the

maximum software capacities/limitations and software configuration guidelines discussed within

this guide.

Colleague Self-Service

Setting

Value

Remarks

CPU cores

Memory (RAM)

6 GB

Colleague Web API

Setting

Value

Remarks

CPU cores

Memory (RAM)

8 GB

Colleague Application Server

The following guidelines focus on the Colleague Application Server resources needed to

support the Colleague Self-Service stack and are in addition to resources that will be needed to

run other Colleague applications, such as other DMI application listeners, the DAS (for UniData

clients), Colleague UI and background processing. These values are not exact guidelines as

there are many factors that can cause the actual resource consumption to be more or less than

the values stated below; they are meant to provide a reasonable starting point based on

experience.

Setting

Value

Remarks

CPU cores

2 cores for each DMI

application listener

1 core for every 6 UniData

licenses used by the DMI

application listener

Each DMI application listener

should be allocated a

minimum of 2 CPU cores.

For every 6 UniData licenses

used 1 CPU core should be

allocated. This count does not

include the database

connection for UniData

clients.

Page 10 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Memory (RAM)

2 GB for each Colleague Web

API being used

256 MB for each UniData

license used by the DMI

application listener(s)

Each DMI application listener

servicing Colleague Self-

Service should have no less

than 2 GB of heap space

allocated for each Colleague

Web API instance pointing to

that DMI application listener.

Based on the limitations of

this guide, if a DMI application

listener is servicing a single

Colleague Web API, then that

DMI application listener would

require 2 GB. If two Colleague

Web APIs are using the DMI

application listener, then the

allocation would be 4 GB.

For each UniData license

used a planned RAM

allocation of 256 MB should

be made. This can vary if the

UniData memory allocation

settings are larger but 256 MB

is a good baseline for what a

UniData executable could

consume while running

Colleague Self-Service

transactions. This count does

not include the database

connection for UniData

clients.

For example, if the total

UniData licenses used by all

DMI application listeners is

100, then 25 GB of RAM

should be allocated (256 MB x

100 = 25600 MB \ 1024 = 25

GB). Reminder: this RAM

allocation is just for Colleague

Self-Service and does not

take into account the other

needs of the Colleague

Application Server such as

other DMI listeners, Colleague

UI, and background

processes.

Page 11 of 31

Colleague Self-Service Hardware and Software Configuration Guide

DMI DAS Listener (when using DAS Data Reader feature of API 1.12+)

The following resource guidelines are for when a separate DMI DAS listener is setup exclusively

for use with the DAS Data Reader feature of Colleague Web API 1.12+.

Setting

Value

Remarks

CPU cores

Memory (RAM)

4 GB

Page 12 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Software configuration guidelines

The following guidelines are based on the maximum software capacities and hardware resource

guidelines discussed within this guide.

Colleague Self-Service

Setting

Value

Remarks

Microsoft .NET configuration - machine.config

(32-bit, C:\Windows\Microsoft.NET\Framework\v4.0.xxxxx\Config\machine.config)

processModel element

<processModel

enable="true"

autoConfig="false"

maxWorkerThreads="A"

maxIoThreads="A"

minWorkerThreads="B"

minIoThreads="B" />

* See remarks for calculations used

to determine A and B above.

A = 1200 / CPU cores

B = Ceiling(A / 2)

Based on the hardware

resource guidelines (8 CPU

cores) A=150 and B=75.

Note: If copied from this

document ensure that quote

characters are normal quotes

and not back/forward quote

characters as this will cause

the .NET to fail initialization

Note: The processModel

setting is the ONLY setting

that should be changed in the

machine.config file! Other

Ellucian products, such as

Colleague UI, call for other

changes to this file and those

changes will negatively affect

the performance of Colleague

Self-Service!

IIS - Application Pool Advanced Settings

Process Model:

Idle Time-out (minutes)

1440

Process Model:

Maximum Worker

Processes

Recycling:

Regular Time Interval

(minutes)

Recycling:

Specific Times

None

Page 13 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Colleague Self-Service application settings

Colleague Web API

Connection Limit

400

Colleague Self-Service web.config file

(Student\web.config)

AppSettings:

Registration Mode

<add key="RegistrationMode"

value="True"/>

This is a recommended

setting that has been added

to improve performance.

Enable (by settings to True)

during planned registration

periods to shift resource

usage to favor better

registration performance.

Disable (by settings to False)

during all other times.

Page 14 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Colleague Web API

Setting

Value

Remarks

Microsoft .NET configuration - machine.config

(32-bit, C:\Windows\Microsoft.NET\Framework\v4.0.xxxxx\Config\machine.config)

processModel element

This is the default setting in

the machine.config. You

should not have to modify this

value.

IIS - Application Pool settings

Process Model:

Idle Time-out (minutes)

1440

Process Model:

Maximum Worker

Processes

Recycling:

Regular Time Interval

(minutes)

Recycling:

Specific Times

None

These recycling settings

assume that the Colleague

Web API warm-up script is

being run, at a minimum,

once every 24 hours.

Colleague Web API application settings

API v1.11 and

prior

API v1.12+

Using DAS Data

Reader

Use of the DAS Data Reader

in API v1.12+ is

recommended for both

performance and better

license utilization. If not using

this option follow the

guidance for API v1.11 and

prior.

Connection Pool Size

DAS Connection Pool

Size

N/A

Page 15 of 31

Colleague Self-Service Hardware and Software Configuration Guide

DAS Data Reader feature of API v 1.12+

In on-going efforts to improve performance, the ability for Colleague Web API to fetch read-only

data directly from a DMI DAS listener was added in version 1.12. This feature provides the

fastest path possible for reading data in the Colleague architecture and can do so using fewer

database connections when compared to using the DMI application listener for reading data.

Given that the majority of data provided through the Colleague Web API is from read-only

sources, enabling this feature removes all data read requests from the DMI application listener,

freeing those connections for data updates and login/logout activities. Some important notes to

keep in mind are:

- The Colleague Web API must be able to talk to a DMI DAS listener so firewall rules may

need to be adjusted.

- The connections can be secured.

- The DAS Data Reader feature supports connection pooling for efficient use of DAS

connections

- The connections to the DMI DAS listener use a much shorter timeout than normal

Colleague connections to prevent sessions from remaining active for long periods of

time after burst loads.

- While a separate DMI DAS listener is not required, it is suggested to use a dedicated

DMI DAS listener for this feature, especially in cases of multiple Colleague Web APIs

being used in a given environment.

- This feature is currently opt-in via the Colleague Web API Connections settings page.

Switching it on and off is as easy as checking a checkbox. Ellucian felt it important to

provide this feature to improve performance but did not want to simply force its use.

Page 16 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Colleague Application Server

Setting

Value

Remarks

DMI application listener DMI Defaults settings

Connection Pool Size

Sum of the Connection Pool Size

values from all Colleague Web API

instances connecting to this DMI

application listener, maximum of

100.

API v1.11 (and prior), API

v1.12 not using DAS Data

Reader

Based on these guidelines

this value will either be 50

(when one Colleague Web

API instance uses this DMI

application listener) or 100

(when two Colleague Web

API instances use this DMI

application listener).

API v1.12+ Using DAS Data

Reader

Based on these guidelines

this value will be between 20

and 100 in increments of 20

based on the number of

Colleague Web API instances

using this DMI application

listener.

Web Overflow Pool

Size

Pool Thread Timeout

(secs)

300

Web Exec Timeout

(secs)

300

Transactions Before

Database Refresh

100,000

Page 17 of 31

Colleague Self-Service Hardware and Software Configuration Guide

DMI application listener properties

Max Concurrent

Connections

Sum of the Connection Pool Size

values from all Colleague Web API

instances connecting to this DMI

application listener times 2.

API v1.11 (and prior), API

v1.12 not using DAS Data

Reader

Based on these guidelines

this value will either be 100

(when one Colleague Web

API instance uses this DMI

application listener) or 200

(when two Colleague Web

API instances use this DMI

application listener).

API v1.12+ Using DAS Data

Reader

Based on these guidelines

this value will be between 40

and 200 in increments of 40

based on the number of

Colleague Web API instances

using this DMI application

listener.

Page 18 of 31

Colleague Self-Service Hardware and Software Configuration Guide

JVM Arguments –d64

-d64

-Xms and -Xmx

When the DMI application

listener is running on 64 bit

hardware (which is almost

always the case) ensure that

the –d64 argument is present

so that Java uses 64-bit

memory management.

-Xms sets the starting

memory size of the Java JVM

and –Xmx sets the maximum

amount of memory the Java

JVM can use.

Ensure that the correct

amount of Java JVM memory

is allocated to each listener

using the -Xms and –Xmx

arguments. Allocate 2 GB for

each Colleague Web API

connecting to this DMI

application listener. Based on

the guidelines If one

Colleague Web API is

connecting the values would

be -Xms2048m –Xmx2048m,

if two are connecting then the

values would be -Xms4096m

-Xmx4096m. Not having –

Xms and –Xmx set the same

is dangerous and often

results in out of memory

scenarios that will crash the

DMI application listener under

heavy load!

Registry Pool Size

Page 19 of 31

Colleague Self-Service Hardware and Software Configuration Guide

DMI DAS Listener (when using DAS Data Reader feature of API 1.12+)

The following software configuration guidelines are for when a separate DMI DAS listener is

setup exclusively for use with the DAS Data Reader feature of Colleague Web API 1.12+.

Setting

Value

Remarks

DMI DAS listener properties

Max Concurrent

Connections

Sum of the DAS Connection Pool

Size values from all Colleague Web

API instances connecting to this DMI

DAS listener times 2.

The maximum for a single

DMI DAS listener should not

exceed 200 connections

(before being multiplied by 2).

Consider using multiple DMI

DAS listeners if the sum of

the DAS Connection Pool

Size values from all

Colleague Web API instances

connecting to this DMI DAS

listener exceeds 200.

JVM Arguments –d64

-d64

-Xms and -Xmx

When the DMI DAS listener is

running on 64 bit hardware

(which is almost always the

case) ensure that the –d64

argument is present so that

Java uses 64-bit memory

management.

-Xms sets the starting

memory size of the Java JVM

and –Xmx sets the maximum

amount of memory the Java

JVM can use.

Ensure that the correct

amount of Java JVM memory

is allocated to each DMI DAS

listener using the -Xms and –

Xmx arguments. Allocate 4

GB, making the values

-Xms4096m -Xmx4096m. Not

having –Xms and –Xmx set

the same is dangerous and

often results in out of memory

scenarios that will crash the

DMI DAS listener under

heavy load!

Page 20 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Determining your Colleague Self-Service deployment

The first step in determining the amount of Colleague Self-Service infrastructure to deploy is to

determine the expected concurrent user load that could occur. While the natural answer would

be the size of the student population, this in many operational scenarios is beyond what is really

needed. The best answer comes from registration periods and the maximum number of users

who can register at the same time which is often limited by Colleague registration priorities.

Once the number of expected concurrent users is determined that number is used to calculate:

- The number of Colleague Self-Service instances needed,

- The number of Colleague Web APIs needed to support the Colleague Self-Service

instances,

- The number of DMI application listener instances needed to support the Colleague Web

API instances,

- And finally the amount of Colleague Application Server resources needed to support the

transactional needs of the DMI application listener(s).

Remember that Colleague Self-Service and the Colleague Web API can be load balanced to put

in place the necessary instances to support the expected load, and multiple DMI application

listeners can be installed if one DMI application listener is not enough based on the number of

Colleague Web APIs in use.

This section is split into two sub-sections. The first sub-section bases its calculations

using the tuning guidance of Colleague Web API 1.11 and prior. The second sub-section

bases its calculations using the tuning guidance of Colleague Web API 1.12+.

Page 21 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Sample deployment calculations using Colleague Web API v1.11 (and prior)

The following table provides sample calculations based on a typical small, medium, large and

extra-large load sizes when using Colleague Web API v 1.11 and prior. Remember that each

instance or license has hardware requirements which are listed in the Hardware resource

guidelines section of this guide.

Load

Size

Number

Reg/hour

Concurrent

Self

Service

Users

(estimated

in first 10

minutes)

Colleague

Self-

Service

Instances

Colleague

Web API

Instances

DMI

Application

Listeners

Colleague

Licenses

(UniData

Licenses)

By Client

Database

Type

Small

250

125

UniData: 115

SQL/Oracle:

Medium

700

350

UniData: 115

SQL/Oracle:

Large

2,000

1,000

2, Load

Balanced

3, Load

Balanced

UniData: 330

SQL/Oracle:

160

Extra

Large

5,000

2,500

4, Load

Balanced

7, Load

Balanced

UniData: 760

SQL/Oracle:

370

Page 22 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Large Load sample using Colleague Web API v1.11 (and prior), explained

- 2 Colleague Self-Service instances, load balanced.

o A single instance of Colleague Self-Service can handle around 700 concurrent

users. Two instances will be needed to support 1000 users.

- 3 Colleague Web API instances, load balanced.

o A single instance of Colleague Web API can handle around 400 concurrent

requests. Three instances will be required to support 1000 requests from the

1000 users.

- 2 DMI application listeners.

o A single DMI application listener can handle, at most, two instances of Colleague

Web API with the recommended Connection Pool Size of 50 from each

Colleague Web API instance. Two DMI application listeners will be needed.

 The first listener will be servicing two Colleague Web API instances. At

any given time, it will only ever receive 100 requests from the Colleague

Web API instances (50 from each).

 The second listener will only be servicing one Colleague Web API

instance. At any given time, it will only ever receive 50 requests from the

third Colleague Web API instance.

 The Connection Pool Size (DMI Defaults) will be 100. The first listener will

use all 100 connections, but the second listener will only use 50 of the

100. This effectively results in 150 maximum connections made to

Colleague even though the listeners have the ability to create 200

connections to Colleague. The Connection Pool Size (Colleague Web

API) of the third Colleague Web API instance being 50 makes this work!

 The Web Overflow Pool Size (DMI Defaults) will be 5. With two DMI

listeners in use, a maximum of 10 connections can be made to

Colleague.

 The Registry Pool Size (listener properties) of each DMI Application

listener will be 5. With two DMI listeners in use, a maximum of 10

connections can be made to the database.

Page 23 of 31

Colleague Self-Service Hardware and Software Configuration Guide

- Colleague licenses (UniData Licenses)

o UniData clients will need 330.

 The sum of the Connection Pool Sizes (DMI Defaults) is 150 which are

multiplied by 2 to account for the DAS connection to UniData for a

subtotal of 300.

 The sum of the Web Overflow Pool Sizes (DMI Defaults) is 10 which are

multiplied by 2 to account for the DAS connection to UniData for a

subtotal of 20.

 The sum of the Registry Pool Sizes (listener properties) is 10 and is not

multiplied by 2 as they are direct UniData connections for a subtotal of 10.

 300 plus 20 plus 10 results in a grand total of 330 UniData licenses

needed to support 1000 concurrent Colleague Self-Service users.

o SQL and Oracle clients will need 160

 The sum of the Connection Pool Sizes (DMI Defaults) is 150.

 The sum of the Web Overflow Pool Sizes (DMI Defaults) is 10.

 The Registry Pool Sizes (listener properties) are direct SQL or Oracle

connections and do not impact the UniData license count.

 150 plus 10 results in a grand total of 160 UniData licenses needed to

support 1000 concurrent Colleague Self-Service users.

- Deployment diagram:

Web Browser

Colleague Self-Service

Colleague Application Server

Colleague Web API

DMI application listener

Colleague Web API

Colleague Self-Service

Colleague Web API

DMI application listener

Load Balancer

Page 24 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Sample deployment calculations using Colleague Web API v1.12+

The following table provides sample calculations based on a typical small, medium, large and

extra-large load sizes when using Colleague Web API v 1.12. Remember that each instance or

license has hardware requirements which are listed in the Hardware resource guidelines section

of this guide.

Load

Size

Number

Reg/ho

Concurre

nt Self

Service

Users

(estimate

d in first

minutes)

Colleagu

e Self-

Service

Instance

Colleagu

e Web

API

Instance

DMI

Applicati

Listeners

Colleague

Licenses

(UniData

Licenses) By

Client

Database

Type

Total

DMI

DAS

listener

connec

tions

Small

250

125

UniData: 85

SQL/Oracle:

Medium

700

350

UniData: 85

SQL/Oracle:

Large

2,000

1,000

2, Load

Balanced

3, Load

Balanced

UniData: 225

SQL/Oracle:

Extra

Large

5,000

2,500

4, Load

Balanced

7, Load

Balanced

UniData: 520

SQL/Oracle:

150

210

Page 25 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Large Load sample using Colleague Web API v1.12, explained

- 2 Colleague Self-Service instances, load balanced.

o A single instance of Colleague Self-Service can handle around 700 concurrent

users. Two instances will be needed to support 1000 users.

- 3 Colleague Web API instances, load balanced.

o A single instance of Colleague Web API can handle around 400 concurrent

requests. Three instances will be required to support 1000 requests from the

1000 users.

- 1 DMI application listener.

o A single DMI application listener can handle, at most, five instances of Colleague

Web API with the recommended Connection Pool Size of 20 from each

Colleague Web API instance. Only one DMI application listener will be needed as

3 Colleague Web API instances will produce a total Connection Pool Size of 60.

 The Connection Pool Size (DMI Defaults) will be 60.

 The Web Overflow Pool Size (DMI Defaults) will be 5.

 The Registry Pool Size (listener properties) of each DMI Application

listener will be 5.

- 1 DMI DAS listener.

o Each Colleague Web API instance will have a DAS Connection Pool Size of 30

for a total of 90 DAS connections. It is recommended that with this number of

additional DAS connections that a separate DMI DAS listener be used.

- Colleague licenses (UniData Licenses)

o UniData clients will need 225.

 The Connection Pool Size (DMI Defaults) is 60 which is multiplied by 2 to

account for the DAS connection to UniData for a subtotal of 120.

 The Web Overflow Pool Size (DMI Defaults) is 5 which is multiplied by 2

to account for the DAS connection to UniData for a subtotal of 10.

 The Registry Pool Size (listener properties) is 5 and is not multiplied by 2

as they are direct UniData connections for a subtotal of 5.

 The sum of the DAS Connection Pool Size from the three Colleague Web

API instances is 90 and is not multiplied by 2 as they are direct UniData

connections for a subtotal of 90

 120 plus 10 plus 5 plus 90 results in a grand total of 225 UniData licenses

needed to support 1000 concurrent Colleague Self-Service users.

o SQL and Oracle clients will need 65

 The Connection Pool Size (DMI Defaults) is 60.

 The Web Overflow Pool Size (DMI Defaults) is 5.

 The Registry Pool Size (listener properties) are direct SQL or Oracle

connections and do not impact the UniData license count.

 60 plus 5 results in a grand total of 65 UniData licenses needed to

support 1000 concurrent Colleague Self-Service users.

Page 26 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Deployment diagram:

Web Browser

Colleague Self-Service

Colleague Application Server

Colleague Web API

DMI application listener

Colleague Web API

Colleague Self-Service

Colleague Web API

DMI DAS listener

Load Balancer

Page 27 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Colleague Self-Service deployment planning and settings worksheet

Colleague Self-Service

1. Expected number of concurrent users.

2. Number of Self-Service instances needed. Divide Box A by 700, round up to the nearest whole

number.

Self-Service settings

3. Colleague Web API Connection Limit. Enter 400

400

Colleague Web API

4. Number of Web API instances needed. Divide Box A by 400, round up to the nearest whole number.

Web API settings

5. Connection Pool Size

Are you using DAS Data Reader (API 1.12+) Yes: enter 20 No: enter 50

6. DAS Connection Pool

Size

Are you using DAS Data Reader (API 1.12+) Yes: enter 30 No: enter 0

DMI Application Listener(s)

7. Total Application Listener Connections from Web API. Multiple Box D by Box E.

8. Number of DMI Application Listeners needed. Divide Box G by 100, round up to the nearest whole

number.

DMI Settings – DMI Defaults

9. Connection Pool Size. Enter the lessor of Box G and 100.

10. Web Overflow Pool Size. Enter 5

DMI Settings – Application Listener Properties

11. Max Concurrent Connections. Multiple Box I by 2.

12. Registry Pool Size. Enter 5

Colleague Application Server

UniData Licenses

13. Licenses used by the main pools. Enter Box G.

14. Licenses used by the overflow pools. Multiple Box H by Box J.

15. Licenses used by Registry pools.

UniData clients: Multiple Box H by Box L.

SQL Server or Oracle: Enter 0.

16. Licenses used by main and overflow pool database

connections.

UniData clients: Add Boxes M and N.

SQL Server or Oracle: Enter 0.

17. Licenses used by DAS Data Reader.

UniData database: Multiple Box D by Box F.

SQL Server or Oracle: Enter 0.

18. Total UniData licenses needed. Add Boxes M through Q.

System Resources

19. CPU Cores needed. Add Boxes M, N and Q then divide by 6.

20. RAM needed (GB). Add Boxes M, N and Q then divide by 4.

DMI DAS Listener(s) - when using DAS Data Reader feature of API 1.12+

21. Total DAS Data Reader connections from Web API. Multiple Box D by Box F.

22. Number of DMI DAS Listeners needed. Divide Box U by 200, round up to the nearest whole number.

DMI Settings – DAS Listener Properties

23. Max Concurrent Connections. Enter the lessor of (Box U multiplied by 2) and 400.

Page 28 of 31

Colleague Self-Service Hardware and Software Configuration Guide

This worksheet is also available as a downloadable Excel spreadsheet with built-in formulas,

Colleague Self-Service Hardware and Software Configuration Guide – Worksheet.

Page 29 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Procedures reference

Configuring a Colleague Self-Service static machine key

Important: The following steps are to be done at the web site level within IIS! Configuring these

settings at the server level or application level can lead to problems as the software is

architected by Ellucian to expect the configuration at the web site level.

1. Open IIS Manager on the web server.

2. Under the Sites node, choose the Colleague Self-Service site node to view the website

configuration options.

3. Double-click the Machine Key icon.

4. On the Machine Key form, clear all four check boxes.

5. Within the Actions pane on the right side of the Machine Key form, click on the Generate

Keys link to generate unique keys for validation and decryption. The Machine Key from

should now look like this:

6. Within the Actions pane on the right side of the Machine Key form, click Apply to save

the changes.

Page 30 of 31

Colleague Self-Service Hardware and Software Configuration Guide

Copying a static machine key from one instance to another

When load balancing Colleague Self-Service or Colleague Web API all instances of the given

product must share the same machine key. If load balancing both Colleague Self-Service and

Colleague Web API the machine key used for the Colleague Self-Service instances should be

different from the machine key used for the Colleague Web API instances for higher security.

Use the following steps to copy the machine key from one instance to the other instances of

Colleague Self-Service or Colleague Web API.

Important: The following steps are to be done at the web site level within IIS! Configuring these

settings at the server level or application level can lead to problems as the software is

architected by Ellucian to expect the configuration at the web site level.

1. Starting with the instance that has already had its machine keys configured, open IIS

Manager on the web server.

2. Under the Sites node, choose the Colleague Self-Service or Colleague Web API site

node to view the website configuration options.

3. Double-click the Machine Key icon.

Page 31 of 31

Colleague Self-Service Hardware and Software Configuration Guide

4. On the Machine Key form

a. Note the Encryption and decryption methods

b. Copy the generated Validation key and Decryption key values from the two text

fields on the form

5. For each additional instance of Colleague Self-Service or Colleague Web API

a. Open IIS Manager on the web server.

b. Double-click the Machine Key icon.

c. Set the Encryption and Decryption methods to be the same as the values noted

in step 4a.

d. Clear all 4 checkboxes.

e. Paste the Validation key and Decryption key values into the appropriate text

fields using the values captured in step 4b.

f. Click the Apply link on the far right-hand side in the Actions pane.