Will Web 2.0 applications break the cloud?
Upcoming SlideShare
Loading in...5
×
 

Will Web 2.0 applications break the cloud?

on

  • 577 views

Computing in the cloud is fashionable and in many cases extremely cost-effective. But - considering a flawed execution model of rich Web 2.0 applications - will Web applications in the cloud fail to ...

Computing in the cloud is fashionable and in many cases extremely cost-effective. But - considering a flawed execution model of rich Web 2.0 applications - will Web applications in the cloud fail to live up to the promise due to performance and security issues?

In this presentation - I discuss security and performance issues of Web 2.0 apps in the cloud and talk about the kind of mistakes people make.

I wrap up with some thoughts on the game changers

Statistics

Views

Total Views
577
Views on SlideShare
576
Embed Views
1

Actions

Likes
0
Downloads
4
Comments
0

1 Embed 1

http://www.linkedin.com 1

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Will Web 2.0 applications break the cloud? Will Web 2.0 applications break the cloud? Presentation Transcript

    • Rich Web applications 2011 Crashing in the cloud Danny Lieberman dannyl@software.co.il http://www.software.co.il/wordpress/  Copyright Creative Commons Attribution License by Danny Lieberman
    • Course Content Preface Security Performance The future Summary
    • PrefaceCloud computing is fashionable. Ralf Lauren Fall 2010
    • PrefaceBut what about performance & security? Coco Chanel circa 1920
    • Cloud and the “security problem” Why is security so hard to sell today?  Complex  Hard to understand  Economic benefit to business unclear
    • Cloud and the “security problem” Computing as a utility – Simple – Easy to measure economic benefit – Security is built-in
    • Cloud and the “security problem” The good news – The Tier 1 providers are better at security than you or me The bad news – You still have application software – Just with a bigger threat surface
    • The cloud threat surface CIO mistakes Application software
    • The top 3 mistakes CIOS make No knowing how much your assets are worth  asset.val()== undefined Writing procedures while attackers exploit your software  $p != security.software Confusing compliance with data security  $c != security.data
    • Rich Web 2.0 applications 2011 2-5 languages Server stack Message passing in the UIPC Browser Smartphone Device 3-5 languages Message passing in the UI
    • Message passing in the UI?Very bad idea. Worst dressed at BET Awards 2010
    • Rich Web 2.0 entry points DB Servers Interfaces Server stack PHP, C#, Ruby, J2EE HTML/Javascript/CSS Web servers HTML XML PC CSS Browser Smartphone Device Javascript Java Flash
    • Rich Web 2.0 attack scenarios Any kind of code injection Server or client returns invalid HTML Pages contain dead links HTML forms dont match field types expected by controllers Client side makes bad assumptions about AJAX services Server may attempt to execute invalid SQL queries Improper marshaling/un-marshaling – DB server to Web server – DB server to application tier – Web server to browser
    • Rich Web 2.0 vulnerabilities Heterogeneous stacks – Too much chewing gum PHP, Ruby, Python – Flexibility, no static type guarantees C#, Java – Static typed, but only at Web server – Code complexity increases threat surface Redundant code on servers and clients Redundant data on servers and clients Client-server latency – Slow HTTP POST attacks
    • Cloud security reference model
    • Security summary Security Control model looks great  But doesnt mitigate core vulnerabilities  Typing issues  Interface issues  Redundant code, data and tiers  Client-server latency
    • Performance - time is money Amazon.com  100 ms of latency costs Amazon 1% of sales (http://highscalability.com) Google.com  500ms delay in delivery is a 20% drop in traffic (Google VP Marissa Mayer) Competing stock trading platforms  5ms delay is $4M in losses / ms.
    • Web servers 2011Browser opens connection.Server forks a thread for each connection, using blocking IO.Ajax latency: 200-600ms
    • Hardware 2011 What about multiple-processor concurrency?  Threads dont scale well with multi-cores  Processes are necessary to scale to multi- core computers, not memory-sharing threads.
    • Threads are a bad idea The mixture of threads and modern multi- core systems add up to some serious race condition potential. http://blogs.msdn.com/b/david_leblanc/archive/2007/04/19/why-threads-are-a-bad-idea.aspx Thread-based networking is inefficient and very difficult to use. http://www.kegel.com/c10k.html and http://bulk.fefe.de/scalable-networking.pdf
    • The future of apps in the cloud The fundamentals of scalable systems are fast networking and non-blocking design— .The fundamentals of scalable systems are fast networking and non-blocking design—the rest is message passing the rest is message passing. 3 technologies will be game changers,I think... ● Web sockets ● Node JS ● Couch DB
    • The future of apps in the cloudWeb sockets Open a connection to Web server It stays open Pass messages Eliminates at least 2 processes for every connection. (Browser-Server & Server-Database) Low Latency: 20-60ms instead of 200-600ms
    • The future of apps in the cloudNode.js Javascript on client and server No threads No blocks or locks UI is HTML & CSS Asynchronous message passing with Web sockets
    • The future of apps in the cloudCouchDB Application served out of CouchDB CouchApp lives in the browser. No middle tier Javascript on client and server UI is HTML & CSS CouchDB uses Ajax to shove JSON back and forth. CouchDB replicates on smart phones
    • Summary Application vulnerabilities are expensive  100x more expensive to fix after implementation  Potential data loss in the cloud  Security controls dont come cheap Time is money  High latency applications less responsive  Your cloud provider charges per CPU cycle  Your costs go up, revenue goes down Promising new technologies  No middle/data tiers, reduced threat surface  10x lower latency  Your costs go down, revenue goes up.