SharePoint 2016/2013/Online: How to Optimize SharePoint Custom Pages Using HTML5 IndexedDB API

In this article we will discuss another obvious performance issues with SharePoint Solutions involving large volume of data transactions surfacing SharePoint Custom Pages.

This could become more prominent if we have strict governance in place and we are not allowed to make use of advanced server side options (Custom Web Service End Point, MTA Enabled Modules etc.).

In one of the recent assignment I came across a similar scenario where I need to crawl data from an external Web Service end Point and surface data on SharePoint Pages. Since the anticipated data volume was huge and traditional caching approaches like Cookies wont’ work due to size limitations.

In pursuit of the solution I have gone through the “HTML5 Web Storage APIs” that allows you to setup an In-Browser Transactional Database System called “IndexedDB”.

Here is a quick introduction of IndexedDB for details I must recommend you to visit IndexedDB

“IndexedDB is a transactional database system, like an SQL-based RDBMS. However, unlike SQL-based RDBMSes, which use fixed-column tables, IndexedDB is a JavaScript-based object-oriented database. IndexedDB lets you store and retrieve objects that are indexed with a key; any objects supported by the structured clone algorithm can be stored. Operations performed using IndexedDB are done asynchronously, so as not to block applications.”

I also want to thanks to “Raymond Camden” for his detailed research on Storage Limits for IndexedDB and believe you must refer this link to understand the limits carefully before getting into concrete implementations.

Now let’s try to understand the implementation details by using following diagram:

Solution Architecture Diagram & Explanation

1

In this solution the SharePoint Page will try to look for the required data in Local Indexed DB created to support this page. If data is not found in local database, page will issue the request for data from SharePoint List.

Since we are dealing with “100,000” Items present in SharePoint List, I made use of “REST API + OData Continuation” data access strategy to overcome SharePoint List Threshold Limits. This mechanism will access only 100 List Items at a time and it is safe to extend this limit up to 2000 items per fetch.

Each fetch will a JSON Object that will be persisted into Indexed DB as individual record. I opt this strategy to reduce the page load time. If the number of items are not much you can add each item as separate record.

Every subsequent data call will be automatically diverted to the local database as primary source.

Additionally we can add “Auto Refresh Modules” to keep the local database fresh with SharePoint List Changes and sync the changes with Indexed DB “Asynchronously”.

Ideally speaking for a complete solution “Auto Refresh Modules” are must to have.

So this all about execution summary for this solution.

Now let’s have look at implementation details as follows-

I have created a SharePoint List with two columns and “100,000” Items added to it as shown below.

Demo

This list will be acting as data source for the page. In actual scenarios this source could be a Web Service End Point which can provide voluminous data on demand.

2

3

Before getting into code let’s see how this Page will behave on execution. Demonstrating the page in action will be helpful later when we get a deep dive in code.

If we run the page we will see this page took about “3 minutes” to get execution completed.

The first execution cycle will include the following actions:

  1. Initialize IndexedDB Database
  2. Query SharePoint List
  3. Add REST API Response to IndexedDB
  4. Load page with data from IndexedDB

Since we are adding data to the store asynchronously, overall application will remain functional even it is taking 3 minutes to complete.

4

Following screen shot showing data adding to IndexedDB asynchronously

5

We can also review the Indexed DB initialized as the part of this request using “Developer Tools or F12 Key” with in the browser as shown below-

6

We can explore each item in the each of the JSON Object as shown below-

7

Now refresh the page to see the execution again and we can see roughly “1 second” to complete the page request.

The subsequent execution cycle will include the following actions:

  1. Query IndexedDB for data
  2. Load page with data from IndexedDB

So we can see how we can trim the execution path by using a well-defined strategy.

8

Code Analysis

Let’s do the code analysis to understand the concrete implementation.

In Step 1 we are enclosing some of the literals as variables and will refer theses variables later in the code

9

In Step 2 we are checking if respective Indexed Database is initialized already or not and if not Initialize the Database. In this demos let’s call this database as “Products”

10

In Step 3 “onsuccess” event handler will get executed and database object will get stored in a global variable “SharePointOptimization.sharePointStore”. This variable will be acting as start point for all the operations on the database in future.

In Step 4 default error handling module is assigned as callback function to “onerror”, “onblocked”, “onabort” event handler

11

In Step 5 we are querying SharePoint List using REST API

12

In Step 6 we are making use of OData Continuation Techniques to overcome SharePoint List Threshold restrictions.

In this step we also call “AddDataToStore” function that will add SharePoint List Items coming as JSON Object to the Local Indexed Database.  It is important to recall that in this demo I am storing 1 JSON Object as 1 record in database and each object contains information for 100 List Items.

13

In Step 7 we are adding JSON Objects to IndexedDB. In order to do that we need to perform following operations-

  • Initialize Transaction with Read Write Operation Permissions
  • Get Handle on “Products” Database inside IndexedDB Data Stores
  • Call asynchronous “add” method to add JSON Object to “Products” Store

In Step 8 we are calling “QuerySharePoint” function to query data from SharePoint List in case data is not available in Local Database.

14

Steps 9, 10, 11 explains about “ReadSPStore” function where we will read the data from Local Data Store (IndexedDB)

In Step 9 following operations are performed-

  • Initialize Transaction with Read Operation Permissions
  • Get Handle on “Products” Database inside IndexedDB Data Stores
  • Call asynchronous “count” method to get total number of JSON Object available in “Products” Store

In Step 10 following operations are performed-

  • Check for get count request status
  • If success Initialize Indexed DB Cursor by calling asynchronous “openCursor” function

In Step 11 following operations are performed-

  • Check for get cursor request status
  • If success read the record from IndexedDB and add to the local array variable
  • Call “continue” function as long as there are items left in local store
  • Once all data is read and save to the local array pass this array to “RenderUI” function to render this data on the interface as required

15

In Step 12 we can plug any UI engine to produce more intuitive UI as applicable, for the sake of this demo I am writing out the Count of Store records * 100 (since each record contains 100 Items) to show the total number of items stored in the local store.

16

Steps 13, 14, 15 show you a helper function to check if local store contains required data or not. It helps to decide if we need to read data from Local Store or SharePoint List

“GetProductCount” function is quite similar to the “ReadSPStore” function except it perform a lesser number of operations

17

In Step 16 we will initialize Local SharePoint Store by calling “InitializeSharePointStore” function

18

In Step 17 we can see some of the UI elements to build a basic UI for this demo

19

Point of caution

Before implementing this mechanism make sure you have identified all the compatibility issues around this corner.

I would recommend you to refer the following site every now and then to make sure you are using features supported by the targeted browsers.

http://caniuse.com/#search=IndexedDb

20

Since I have made use of artifacts which are compatible with SharePoint Online Development Guidelines so we can use this approach with pages hosted in SharePoint Online as well.

That is all for this demo.

Hope you find it helpful.

 

Advertisements

SharePoint 2013: How to call OData Web Service End Point Using SharePoint Designer 2013 Workflows

In this article we are going to explore “Call HTTP Web Service” Workflow Action which is newly introduced in SharePoint 2013. In order to test this action will make use of OData Web Service End Point and target the following scenario where Customer Data needs to be auto populated based on the Customer ID using SharePoint Designer Workflow and OData Web Services.

Prerequisites:

  1. As this is going to be a List Workflow, make sure you must have a Custom List containing three columns “Customer ID”, “Full Name” and “Employer” already created on the site which you are going to select in Step 2 down the line. For the sake of this demo I have created a List with the name “Fill Customer Using Web Service” containing all the required columns.
  2. OData Web Service End Point http://services.odata.org/V2/Northwind/Northwind.svc should be accessible from the system which is running SharePoint Designer.
  3. Fiddler Web Proxy should be installed. You can get Fiddler from http://www.telerik.com/download/fiddler

Following are the steps involved in the development of Workflow Calling OData Web Service End Point:

Step-1 : Start SharePoint Designer 2013

Step-2 : Select the site in which you want to create the workflow

1

Step-3 : Go to left navigation pane and select Workflows

2

Step-4 : Select List Workflow [From Ribbon] > Select List [In our case it would be “Fill Customer Using Web Service”]

3

Step-5 : On the Workflow Designer select “Call HTTP Web Service” Action from the Action Ribbon Menu

4

Step-6 : Create the Workflow variables as shown below

5

Step-7 : Specify the Web Service End Point, and HTTP Method as shown below

6

Step-8 : Specify the variable of type Dictionary as placeholder for the incoming response data from the Web Service End Point

7

Step-9 : Now before proceeding any further lets’ analyze the response from the Web Service End Point. Please note that this step is of high importance and must not be neglected. Fiddler Web Proxy is your best friend when it comes to deal with Web Services. Analysis using Fiddler will provide us with the following Vital Information out of the response

  1. Structure of Data returned as a part of the response
  2. Properties Exposed by the Web Service End Point
  3. Data Types of the Properties
  4. Additional information on Error & Exceptions (If any)

8

Step-10 : After analyzing the data we come across the fact that we have got the required data in form of Key Value Pairs and the result set can be accessed by using the path “d/PropertyName”. So based on this knowledge we can now set the values for variable “CustomerName” & “CompanyName” using “Get an Item from a Dictionary” Action as shown below.

9

Step-11 :  As we have got the data from the Web Service End Point and stored in a set of variables, now it is time to update our list using “Set Field in Current Item” Action. We should use two instances of the same Action in order to set the value for both the fields using the variables as shown below

10

Step-12 : With this we are good to go. Now lets’ Save the Workflow Definition, Check for the Errors and if everything is fine Publish the Workflow.

13

14

15

16

17

Step-13 : In order to test the Workflow lets visit our List “Fill Customer Using Web Service” and quickly add a new item. We need to add the Customer ID only and rest of field will be populated by the Workflow

18

Step-14 : Specify the Customer ID and Save the item.

19

Step-15 : Start the Workflow on the newly added item.

20

21

Step-16 : And sure enough you will get the desired output with “FullName” and “Employer” Columns populated by our workflow using OData Web Service Call.

22

Though the scenario I mentioned here is quite simple, but this new add-on to SharePoint Designer 2013 is very powerful and could be utilized to cater complex business requirements without using any need of writing modules.

Hope this will help someone in need…