First spring
Download as PPTX, PDF0 likes319 views
This document outlines the goals and implementation of an Android app for displaying statistical data diagrams. The key goals are to scale charts to mobile screens, optimize data transmission and display, and allow zooming capabilities. Technically, it involves communication between the Android app and a backend server using REST APIs. Tasks completed include scaling algorithms, caching mechanisms, and non-interactive zooming. Remaining tasks involve improving interactive zooming. Technologies used include a Java servlet container, REST framework, ORM framework, and MySQL database.
First spring
- 1. 1.1. Concept 1.2. App Goals Introduction and Goals 1.3. General Goals 2.1. Communication Android-Server General Diagrams 2.2. Entity-Relationship Table of Contents 3.1. Tasks done: Scaling 3.2. Tasks done: Optimization 3.3. Tasks done: Cache Tasks 3.4. Tasks done: Non-Interactive Zoom 3.4. Remaining tasks: Interactive Zoom 4.1. Technologies used: REST Imp. Technologies used 4.2. Technologies used: ORM Imp. 5. Questions 1
- 2. 1 Introduction and Goals 2
- 3. » Android App for graphical display of statistical data. » Diagrams should have zoom and scroll capabilities. » Different months, different series and different qualities can be selected. » Each user can enrich the application adding comments to the diagrams. Concept 3
- 4. » In any Mobile App, some factors to take into account, » Computational process is limited Heavier tasks must take place on the server. » Memory is limited Avoid heavy objects, or keep them in cache and instantiate them only when is strictly necessary. » Don’t waste bandwidth Avoid unnecessary, heavy or repetitive requests to the server. » Save battery Don’t make an intense use of the computational processor and of the internet connection. App Goals 4
- 5. » Scale the charts to the Mobile screen resolution. » Optimize the transmission and the display of the charts. » Create an optimal mechanism to interact with the server in the zoom-in process. » Create a cache mechanism for retrieve previous graphs » Have different users with different access levels (chars, categories). General Goals 5
- 6. General Diagrams 6
- 7. Web Server: Handle the request; transform it in a SQL query; process it and creates the JSON representation for the response. Database: Basic DB behavior, response to the SQL queries. Client: Performs an HTTP Request; converts the JSON response to Java Object and display the results and/or performs some data processing. HTTP/JSON REST Web Server Mobile App Database TCP/IP Communication Android-Server 7
- 9. Tasks 9
- 10. » Scaling (average algorithm) Fix Output = Screen resolution Tasks done: Scaling 10
- 11. » Optimization (deleting, or nulling the duplications) Variable output Tasks done: Optimization 11
- 12. » Why cache? Persistence == Requests == Battery consumption » Two level cache implemented: Disk & Memory Cache. » MemoryCache Offers fast access at the cost of taking up valuable application memory. » DiskCache Persist valuable data retrieved from the Server. Fetching and writing process is slower than in MemoryCache. » Both Caches are HashMap, and the key will be the HASH of the URL Request. » Both caches implements the algorithm LRU. Tasks done: Cache 12
- 13. 1. Create HASH of the URL Request. 2. Try to obtain the key from the Memory Cache. If works, go to 5. 3. Try to obtain the key from the DiskCache. If works, go to 5. 4. Retrieve the Chart from the Server. 5. Save the chart in MemoryCache (Object) and DiskCache (JSON) Tasks done: Cache 13
- 14. » Choosing between the following procedures, » Requests on demand Request to the server every time we zoom (slow and intense use of Internet). » Request when higher resolution needed Request more than you need (higher resolution), but not in each zoom. » Requests in advance Request in background other resolutions, that can be loaded directly when the user requires them (requires cache implementation). » For the zoom, last procedure has been chosen Tasks done: Zoom 14
- 15. » Request in advance for non-interactive Zoom » Non-interactive Zoom Zoom using the contextual menu on Android. Used for moving among Years/Months/Weeks/Days. Tasks done: Non-interactive zoom 15
- 16. » Request in advance for non-interactive Zoom » Interactive Zoom Two fingers and scroll zoom. » Whenever the zoom match the menu selection (Year/Month/Week/Day) the view is automatically changed to the match, and the corresponding action is performed (previous slide) Remaining Task: Interactive zoom 16
- 17. 4 Technologies used 17
- 18. » Server Side » Use of Apache Tomcat as a Java Servlet Container » Use of Jersey Framework for implementing REST. » Use of Hibernate Framework for implementing ORM. » Use of MySQL as a Relational Database Management System. Communication Android-Server 18
- 19. REST: The predominant Web Service Model Software system designed to support machine-to-machine interaction over a network. Clients requests to servers; servers process requests and return responses in a well known format. » Transport layer is HTTP/MIME. » The URI defines the resource. » The operations are defined by the HTTP methods (GET, POST, PUT, DELETE) » The response is presented in JSON format (indicated thought Media-Type). REST Implementation 19
- 20. Advantages » Simplifying the communication among independent and foreign systems. » Organize an application into simple resources . » No problem with firewalls and proxies (port 80, usually open) » JSON is a lightweight text-based open standard and the Android parser offers good performance. » Possibility of HTTP cache and proxy server to handle high load. Examples » GET http://example.com/charts Obtain List of Charts. » GET http://example.com/charts/1?month=1 Obtain Chart info for January » PUT http://example.com/charts/1 Modify Chart with id=1 REST Implementation 20
- 21. ORM: Mapping technique for converting DB structure Object structure Advantages » Simplify the development of a service. » Overcomes vendor specific SQL differences - the ORM knows how to write vendor specific SQL so you don't have to. Disadvantages » Developers loose understanding of what the code is actually doing the developer is more in control using SQL. » ORM has a tendency to be slow, if is not used carefully. ORM Implementation 21
- 22. Hibernate @Entity @Table(name = "categories") public class Category{ private int id; private String name; private String description; private Set<Chart> charts; @Id @GeneratedValue @Column(name = "category_id") public int getId() { return id; } public void setId(int id) { this.id = id; } public String getName() { return name; } public void setName(String name) { this.name = name; } public String getDescription() { return description; } public void setDescription(String description) { this.description = description; } @OneToMany(mappedBy="category", fetch=FetchType.EAGER) @LazyCollection(LazyCollectionOption.FALSE) public Set<Chart> getCharts() { return charts;} public void setCharts(Set<Chart> charts) { this.charts = charts; } } ORM Implementation 22
- 23. ? Questions? 23
Editor's Notes
- #5: http://tomayko.com/writings/rest-to-my-wifeA resource is a concept that refers to this thing that you want it refers to a noun.A operation it refers to a verb.So we can GET, POST, PUT and DELETE charts, categories, comments….Parameters AdjectivesThe other, are just different representations of the concept.
- #13: Least Recently Used (LRU): discards the least recently used items first.
- #14: Least Recently Used (LRU): discards the least recently used items first.