Enterprise Mobility – Performance Know-How

Enterprise Mobility : Business Importance

• Tablet industry growing from 233 million in 2013 to 1 billion  in 2017 for a CAGR of 36%, according to Yankee group
• According to a recent study, By 2015, the world’s mobile worker population will reach 1.3 billion, representing 37.2% of the total workforce. According to an updated forecast from International Data Corporation (IDC)
• Mobile analytics gives enterprise organization the ability to analyze, collaborate, and act within a single workflow for better decisions – across lines of business and in any industry, which involves high volume of business data in the backend
• Utilizing mobile devices requires that by considering limited CPU speed, reduced memory and storage, narrow bandwidth and high latency connections, and limited battery life

• Complex device/user traffic

• High user experience (Mobile ≠ Desktop !!!)

• Devices with different hardware configurations
• Diverse network connectivities/technologies

It is important to achieve good performance and scalability with Mobile Enterprise Applications.

Enterprise Mobility – Architecture

Important Performance KPIs for Mobile Applications

• End 2 End Response : Very high priority KPI in the customer perspective. This will be the combination of response time from each layer(Device to Middleware to backend to device).

Remark : E2E response also includes the network latency. LAN network latency b/w different application server back ends with middleware and WiFi/3G/4G network latency, where device operates. Since control over network latency is limited, tight coding standards to achieve better performance is important.

• Roundtrips : one network roundtrip per user interaction step.
• Transferred data volume : Keep the data volume as low as possible (approx. 10 ~ 20 KB per user interaction step).
• CPU Time : Minimum device CPU consumption.
• Memory : Minimum memory consumption.
• Memory Leak : Zero memory leaks

Performance Characteristics – Offline Application

Data Volume and Throughput Characteristics

• App performance is determined by device persistence performance
• Typical number of total BO instances on a single device is dependent on the platform
• Typical throughput of 100 uploaded CRUD messages per device and day

Resource Consumption on Server Side

• Low performance impact on backend system
• DOE is the multiplying component
• Typically two synchronization peaks per day (Morning and Afternoon)
• Sizing of landscape is critical because of idle vs. peak time

Performance Characteristics – Online Application

Data Volume and Throughput Characteristics

• Each device is fetching the business data from the backend
• Low number of screens/objects
• Low number of CRUD messages per device and day
• Recommendation: not more than 100 objects per screen/list

Resource Consumption on Server Side

• High performance impact on backend system
• One device user is creating a backend load equivalent to one online user
• App performance is determined by server stacks and connectivity

Performance Recommendations :

App Implementation

• Smart caching mechanism is essential for good performance. Reuse configuration/customizing data
• Think about doing asynchronous prefetching of business data
• Splitting the data that is retrieved into multiple requests. Load important data first, display it and then retrieve the remaining data e.g. pictures asynchronously
• No more than two round trips in parallel, because GSM/3G networks will add because of serialization and low bandwidth a lot of response time
• Keep data volume at a minimum, recommendation is 10 KB compressed data volume per screen

Architectural Design

• App performance is determined by server stacks and network connectivity
• For optimal performance the app design has to follow the backend applications
• Keep the resource consumption in backend at a minimum to ensure good response time

Performance Tools

Tools for iOS

Instruments can be used for analysis of most of performance related issues

• Memory Leak detection
• Memory allocation
• Method Profiling
• Frame Rate of views with animations
• Battery usage
• Test Automation

Fiddler can be used for analysis of HTTP communication

For details see https://developer.apple.com/Instruments

Tools for Android

For details see https://developer.android.com/