Workload Characterization of Selected JEE-based Web 2.0 Applications - - PowerPoint PPT Presentation

▶
workload characterization of selected jee based web 2 0
SMART_READER_LITE
LIVE PREVIEW

Workload Characterization of Selected JEE-based Web 2.0 Applications - - PowerPoint PPT Presentation

Feb 04, 2024 405 likes •578 views

Workload Characterization of Selected JEE-based Web 2.0 Applications Priya Nagpurkar , William Horn, U Gopalakrishnan, Niteesh Dubey, Joefon Jann, Pratap Pattnaik IBM TJ Watson Research Center IISWC 2008 Introduction Motivation : Web 2.0

slide-1
SLIDE 1

Workload Characterization of Selected JEE-based Web 2.0 Applications

Priya Nagpurkar, William Horn, U

Gopalakrishnan, Niteesh Dubey, Joefon Jann, Pratap Pattnaik

IBM TJ Watson Research Center IISWC 2008

slide-2
SLIDE 2

IISWC '08

2

Introduction

  • Motivation: Web 2.0 likely to be an important emerging

workload

  • Goals

Determine if Web 2.0 workloads differ from legacy

workloads in the way they impact the systems that host them

Evaluate the efficacy of current systems in hosting these

workloads, and implications for future systems

  • This paper

Setting up applications that exploit Web 2.0 features as

benchmarks, generating Web 2.0 workloads

Characterization on IBM’s Power5 architecture

slide-3
SLIDE 3

IISWC '08

3

Outline

  • Introduction
  • Web 2.0 Overview
  • Benchmarks and Workloads
  • Results
  • Related Work
  • Conclusion
slide-4
SLIDE 4

IISWC '08

4

Web 2.0 Tag Cloud

Wiki’s

Blogs

Social Networks

AJAX

RSS Mash-ups

Social Computing

MySpace Tags

Web 2.0

del.icio.us

Flickr

REST

ATOM LinkedIn

folksonomy

Digg

bookmarking

Communities

Video Sharing

Avatars

Facebook

SOA

  • Collaboration and Social Networking, Mashups, Rich Internet

Applications, Media Sharing, underlying technologies

slide-5
SLIDE 5

IISWC '08

5

Social Networking Social Bookmarking Rich Internet Application Mashup

slide-6
SLIDE 6

IISWC '08

6

Web 2.0 Overview

Client (sophisticated browser)

application server

Networking advances

inter-server sharing:

  • server-side mashups
  • workflows via SOA

Emerging Applications:

  • social networking
  • server-side productivity

applications

  • software-as-a-service

Heterogeneous stacks:

  • LAMP
  • J2 EE/ JEE
  • Project Zero
  • Ruby on Rails

CPU disk cpu disk cpu disk

web server application server database server AJAX REST Flash Atom

Web 2.0 is data-centric:

  • user-contributed

content

  • user application

data database server web server

client-side mashups

slide-7
SLIDE 7

IISWC '08

7

Benchmarks

  • This paper focuses on Java benchmarks and the J2EE stack

Lotus Connections (BLOGS and DOGEAR): IBM’s Enterprise Social Networking

Software

Java PetStore 2.0: Reference, open-source application for building AJAX-based

RIA’s using JEE 5, developed at SUN

  • Lotus Connections BLOGS

Uses Apache Roller blog engine

Struts: editor, Velocity:

rendering

Hibernate: provides persistence Apache Lucene: powers search

  • Java PetStore 2.0

Exploits JEE 5 features like Java

Persistence API (JPA) and Java Server Faces (JSF)

AJAX components implemented

using DOJO

Apache Lucene: powers search

slide-8
SLIDE 8

IISWC '08

8

Workloads

  • Transactions

User-level actions for each application

viewBlog (blogs: total 9), createBookmark (dogear: total 14), selectTag (PetStore: total 9)

Can involve multiple html pages

  • Workload mix

Defines proportion of transactions Based on commonly observed usage patterns (in

internal deployment of Lotus Connections)

slide-9
SLIDE 9

IISWC '08

9

Gigabit

Grinder Load Generators Linux/ I ntel WebSphere Application Server 6.1.0.13 (Lotus Connections 1.0.2) SUN Glassfish Application Server (PetStore 2.0) AI X 5 .3 / 2 -w ay pow er5 , 1 .6 6 GHz TDS 6.0 (LDAP), TDI 6.1 Linux/ 4 -w ay I ntel Xeon, 2 .6 6 GHz DB2 9.1 AI X 5 .3 / 4 -w ay pow er5 + , 1 .9 GHz (ramdisk to avoid IO)

Gigabit Gigabit

Experimental Setup and Test Infrastructure

Extended the open source Grinder framework to drive benchmarks and

generate reports using data collected from different layers

Pow er5 Architecture

  • 2 cores, SMT on, 1.66 GHz
  • Per-core 64KB L1 icache, 32KB L1 dcache
  • Shared ~ 2MB L2 cache
  • Shared (off-chip) 36MB L3 cache
  • 8 GB Memory

OS AppServer Application JVM Architecture DB

slide-10
SLIDE 10

IISWC '08

10

Experimental Methodology

  • Gradual ramp-up: virtual users are added gradually until

throughput stabilizes

15 minute warmup time

  • Virtual users generate requests with zero think time

Single virtual user emulates multiple users

Randomly picks user from 100k profiles

Request type and input data for request chosen randomly

In keeping with workload mix definition

  • Measurements on the Application Server (SUT)

30 minute measurement interval following warmup HPMs: 30 second samples

  • Repeatability

Restore databases to undo writes Data represents average across three runs

slide-11
SLIDE 11

IISWC '08

11

Throughput and CPU Utilization

10 20 30 40 50 60 70 80 90 100 Blogs Dogear PetStore Trade6 % Tim e System User 10 20 30 40 50 10 20 30 40 50 60 70 80 90

Number of (virtual) users TPS

Blogs Dogear PetStore

CPU Utilization ( AppServer)

Throughput w ith increasing users.

System under test (AppServer) > 95% busy < 5% time spent in system level code

slide-12
SLIDE 12

IISWC '08

12

Increased Chattiness and Data-centric Behavior

10K 467K 191K 83K

Bytes Received/ txn

1.5 570.35 50.84 99.80

Bytes Sent/ txn

0.89 0.13 0.70

AJAX Requests/ txn

0.06 0.19 0.31 0.22

POST Requests/ txn

1.14 31.15 30.27 22.38

GET Requests/ txn

645 18.39 19.8 9.00

TPS Trade6 PetStore Dogear Blogs

  • Chattiness

Presence of AJAX requests More GET and POST requests per transaction

  • Data-centric Behavior

More data from client to server (user generated content) Frequent database accesses

slide-13
SLIDE 13

IISWC '08

13

Architecture-level Analysis

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 Blogs Dogear PetStore Trade6

CPI

Other Stall by FPU Stall by FXU LSU Basic latency LSU Reject Dcache m iss Other (GCT Em pty) Branch m isprediction penalty Icache m iss penalty Com pletion Cycles 0.5 1 1.5 2 2.5 Blogs Dogear PetStore Trade6 L1 DCache Misses per 1 0 0 instructions Serviced by LMEM Serviced by L3 Serviced by L2

Stall Breakdow n on Pow er5 DCache Misses and w here they are serviced from

  • CPI for Web 2.0 applications lower than Trade6
  • Instruction cache misses not as significant
  • Significant cycles spent in data cache miss stalls
  • Relatively more memory accesses for PetStore 2.0

increase data cache miss penalty

slide-14
SLIDE 14

IISWC '08

14

Related Work

  • Workload characterization studies

Web 2.0

Empirical Evaluation of a Collaborative Web Application [Stewart et al

IISWC08] Enterprise Java applications

Memory system characterization [Barroso et al ISCA98] Architectural evaluation of TPC-W [Marden et al HPCA01] Performance studies of commercial systems on multi-cores [Tseng at al

IISWC08]

Characterizing a complex J2EE workload [Shuf and Steiner ISPASS07]

Java applications

Characterizing the memory behavior of Java workloads [Shuf et al] Study of allocation behavior of SPECjvm98 [Dieckmann and Hoelzle

ECOOP99]

  • New Benchmarks and characterization methodology

General methodology for characterizing commercial workloads [Kunkel et al

00]

The DaCapo benchmark suite [Blackburn et al OOPSLA06]

slide-15
SLIDE 15

IISWC '08

15

Conclusion

  • Characterization of three JEE applications that incorporate Web 2.0

features

User-level characteristics capture increased chattiness and data-centric

behavior

Stalls from instruction cache misses not significant Server-side behavior does not otherwise vary significantly or consistently

from Trade6

  • Future work

Varying workload mixes Further analysis of data centric behavior

Database usage patterns, impact on JPA layer

Additional workloads

Scripting languages

slide-16
SLIDE 16

IISWC '08

16

THANK YOU!

QUESTI ONS?