50
Performance? That's what version 2 is for! Performance? That's what version 2 is for! Eduard Tudenhöfner Eduard Tudenhöfner

Berlin Expert Days 2013 - Performance Talk

  • Upload
    nas-tra

  • View
    302

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Berlin Expert Days 2013 - Performance Talk

Performance? That's what version 2 is for!Performance? That's what version 2 is for!

Eduard TudenhöfnerEduard Tudenhöfner

Page 2: Berlin Expert Days 2013 - Performance Talk

Performance? That's what version 2 is for!

Overview

► Introduction / Motivation

► Performance Issues & Solution Strategies

► How to proactively reduce risk of Performance Issues?

► Conclusion

Page 3: Berlin Expert Days 2013 - Performance Talk

Introduction / Motivation

Page 4: Berlin Expert Days 2013 - Performance Talk

Why is performance important?

StressProfit

Revenue

Real User Experience

SPEED MATTERS

Trustworthy

Performance USABILITY

Page 5: Berlin Expert Days 2013 - Performance Talk

Why is performance important?

StressProfit

RevenueEfficiency

Monitoring

Bottlenecks

Real User Experience

Maintenance SPEED MATTERS

Trustworthy

Performance USABILITY

Success Factor

Page 6: Berlin Expert Days 2013 - Performance Talk

Why is performance important?

StressProfit

RevenueEfficiency

Less Resources Stability

Monitoring

SecuritySecurity

Bottlenecks Economical

Conversion RateFASTER SLA Monitoring

Real User Experience

Maintenance SPEED MATTERS

Trustworthy

Performance

More SPEED

USABILITY

Success Factor

Page 7: Berlin Expert Days 2013 - Performance Talk

Why is performance important?

StressProfit

RevenueEfficiency

Less Resources

Higher Fault Tolerance

Stability

Monitoring

SecuritySecurity

Bottlenecks Economical

Conversion Rate

SALES

FASTERSatisfied customers

BETTER Reputation

SLA Monitoring

Real User Experience

Maintenance SPEED MATTERS

Trustworthy

Performance

More SPEED

Key Performance Indicators

USABILITY

Success Factor

Page 8: Berlin Expert Days 2013 - Performance Talk

Investing in Performance really pays off

Why is performance important?

A page that was 2 seconds slower results in a 4.3%

drop in revenue/user(Bing)

400 ms delay cause 0.59% dropin searches/user

(Google)

400 ms slowdown cause 5-9% drop in full-page traffic

(Yahoo)

Introducing gzip compression resultedin 13-25% speedup and cut outbound

network traffic by 50%(Netflix)

Source: www.stevesouders.com

Page 9: Berlin Expert Days 2013 - Performance Talk

Consequences of Poor Performance

Consequences

► Damaged customer relations

– Reputation of the company suffers

– People will continue to associate poor performance with the product, even when the issue is fixed later on

► Lost income & Delayed project schedules

– Revenue is lost

– Penalties have to be paid due to late delivery

► Increased development & maintenance costs

– Delivering features requires more time and effort if performance issues are hindering the acceptance of those features

– Additional time and resources are required if performance issues are found

Page 10: Berlin Expert Days 2013 - Performance Talk

Consequences of Poor Performance

The cost to fix a performance issue

► Is a Technical Debt (defined by Ward Cunningham)

– doing things the quick&dirty way sets us up with technical debt

– technical debt incurs interest payments (in the form of additional effort)

– The later technical debt is payed

back, the higher the interest will be

► So should we pay huge interest at the end or pay back technical debt every development cycle?

Source: Steven Haines. Pro Java EE 5: Performance Management and Optimization

Page 11: Berlin Expert Days 2013 - Performance Talk

Performance Issues & Solution Strategies

Page 12: Berlin Expert Days 2013 - Performance Talk

Performance Issues & Solution Strategies

Client

Application Server

Databases

Legacy Systems / Service Provider

Page 13: Berlin Expert Days 2013 - Performance Talk

Client

Application Server

Databases

Legacy Systems / Service Provider

Application

Performance Issues & Solution Strategies

Page 14: Berlin Expert Days 2013 - Performance Talk

Client

Application Server

Databases

Legacy Systems / Service Provider

Application

Performance Issues & Solution Strategies

Page 15: Berlin Expert Days 2013 - Performance Talk

Client (UI / Browser)

► Bloated Clients

► Too many requests required until a page is fully loaded– Time to first impression

► Very expensive DOM manipulations

► Unsuitable communication patterns

– long running synchronous calls that block the UI

► Network bandwidth

– especially in the mobile area

Performance Issues & Solution Strategies

Page 16: Berlin Expert Days 2013 - Performance Talk

Performance Issues & Solution Strategiesdone with: www.webpagetest.orgChrome / DSL (1.5 Mbps/384Kbps)

50ms RTT

Page 17: Berlin Expert Days 2013 - Performance Talk

Performance Issues & Solution Strategies

49 requests

3.9 s till page is fully loaded

done with: www.webpagetest.orgChrome / DSL (1.5 Mbps/384Kbps)

50ms RTT

Rendering starts after 3.4 s

DOM completeafter 2.6 s

two uncompressedimages

Page 18: Berlin Expert Days 2013 - Performance Talk

Solution Approaches

► Reducing RTTs by– Reducing number of resources

– Avoiding bad requests

– Combining CSS / JS resources (e.g. during build process)

► Reducing Request overhead by

– Using compression (gzip, deflate, Zopfli)

– Minifying CSS / JS resources (cssminifier.com, jscompress.com)

Performance Issues & Solution Strategies

Page 19: Berlin Expert Days 2013 - Performance Talk

Solution Approaches

► Placement of CSS and JS files– CSS at the top / JS at the bottom

– browser should start rendering as early as possible (user perceives a faster loading page)

– anything below the script is blocked from rendering and downloading until after the script is loaded (even when threads are available)→ entire page is delayed

Performance Issues & Solution Strategies

Page 20: Berlin Expert Days 2013 - Performance Talk

How to achieve that in Java (e.g. in JSF)?

Performance Issues & Solution Strategies

Page 21: Berlin Expert Days 2013 - Performance Talk

How to achieve that in Java (e.g. in JSF)?

► JAWR (jawr.java.net)– Built-in minification

– Enforced caching

– Bundling of resources

– CSS image sprite generation

– Can be integrated in Ant / Maven

– Can be used with JSF, Spring MVC, Wicket, Grails, ...

Performance Issues & Solution Strategies

Page 22: Berlin Expert Days 2013 - Performance Talk

JAWR

Performance Issues & Solution Strategies

source: jawr.java.net

What we would like to achieve

How we want to structure

our work

How we can define files bundles

Page 23: Berlin Expert Days 2013 - Performance Talk

Client

Application Server

Databases

Legacy Systems / Service Provider

Application

Performance Issues & Solution Strategies

Page 24: Berlin Expert Days 2013 - Performance Talk

Memory

► Memory leaks / OutOfMemoryErrors (but not every leak leads to OOME)

► Unnecessary creation of expensive objects

► inappropriate GC strategy / Heap sizing (for generational GC)

Performance Issues & Solution Strategies

Page 25: Berlin Expert Days 2013 - Performance Talk

Solution Approaches (Memory)

► Generation sizing (for generational GC) – -XX:NewRatio=3 → 1:3 (Young:Old) → Young takes ¼ of what was specified with

-Xmx

– Sizing proportion between Old/Young generation is important for performance

– e.g. if too many short-lived objects are created, they are moved to the old generation

– An oversized young generation can also cause performance problems

● JVM guarantees that GC runs on young generation

● space on old generation is reserved for emergencies (so that all objects can be copied)

– Memory analysis with e.g. VisualVM, -verbose:gc

Performance Issues & Solution Strategies

Page 26: Berlin Expert Days 2013 - Performance Talk

VisualVM with VisualGC Plugin

Performance Issues & Solution Strategies

Page 27: Berlin Expert Days 2013 - Performance Talk

Caching

► Wrong caching strategy

► Too much or the wrong stuff is cached

► Inappropriate cache sizes (too small / too big)

Performance Issues & Solution Strategies

Page 28: Berlin Expert Days 2013 - Performance Talk

Solution Approaches (Caching)

► What do we want at the end of the day?– Efficient caching to maximize cache hit ratio that also meets:

● Storage constraints

● Availability constraints

● Tolerance for staleness

► What to cache?

– Easy: Slow-changing / mostly read-only data (metadata, configurations, …)

– Challenging: rapidly changing read-write data

– There will always be tradeoffs

► Caching is not solution to everything

– Bigger cache → less memory available for servicing user requests

Performance Issues & Solution Strategies

Page 29: Berlin Expert Days 2013 - Performance Talk

Remote Boundaries

► Remote boundaries too fine-grained → too chatty

► remote communication often done transparently for the developer

► increased round trips

► increased serializations/deserializations

► and of course: more memory usage

Performance Issues & Solution Strategies

Page 30: Berlin Expert Days 2013 - Performance Talk

Solution Approaches (Remote Boundaries)

► Decrease number of remote calls → „The best call is the call that is not done“

► Boundaries should be more coarse-grained e.g. by using wrapper classes (of course that contain only the really required information)

► Depending on the communication parties (heterogeneous/homogeneous), the right protocol should be used

Performance Issues & Solution Strategies

Page 31: Berlin Expert Days 2013 - Performance Talk

Client

Application Server

Databases

Legacy Systems / Service Provider

Application

Performance Issues & Solution Strategies

Page 32: Berlin Expert Days 2013 - Performance Talk

Databases (from an application's point-of-view)

► More / Less data is retrieved than actually required

► Same data is retrieved over and over again (n+1 query problem)

► High normalization good for reducing redundancy, but bad for performance

► Inappropriate connection pool sizes

► Usage of O/R mappers– can lead to unexpected behavior if used in a wrong way

– possibilities of JPA framework not known or not used efficiently

Performance Issues & Solution Strategies

Page 33: Berlin Expert Days 2013 - Performance Talk

Solution Approaches (Databases / OR Mapper)

► Data Retrieval– Read-Only queries (query.setHint(“eclipselink.read-only“, “true“))

● improves performance by avoiding copying and change tracking the objects

– Fetch Joins

– Batch Reads

– Other loading optimizations

● Use projection queries where appropriate● Use pagination for large result sets (query.setMaxResults(), query.setFirstResults())● Use named queries (likely to be precompiled by provider, reusability)

► Updating Data

– Batch Update

● allows a bunch of update operations to be performed as a single DB access● reduces round trips to the database

Performance Issues & Solution Strategies

<property name="eclipselink.jdbc.batch-writing" value="Oracle-JDBC"/>

Page 34: Berlin Expert Days 2013 - Performance Talk

Fetch Joins - Example

Performance Issues & Solution Strategies

Query query = em.createQuery(“SELECT po from PurchaseOrder po WHERE po.status = 'ACTIVE' AND po.customer.address.city = 'Stuttgart'”);

List<PurchaseOrder> orders = query.getResultList();

for (PurchaseOrder order: orders) { order.getCustomer().getName();}

Page 35: Berlin Expert Days 2013 - Performance Talk

Fetch Joins - Example

Performance Issues & Solution Strategies

Query query = em.createQuery(“SELECT po from PurchaseOrder po WHERE po.status = 'ACTIVE' AND po.customer.address.city = 'Stuttgart'”);

List<PurchaseOrder> orders = query.getResultList();

for (PurchaseOrder order: orders) { order.getCustomer().getName();}

{returns N purchase orders} → 100 positions = 101 SQLs

Page 36: Berlin Expert Days 2013 - Performance Talk

Fetch Joins - Example

Performance Issues & Solution Strategies

Query query = em.createQuery(“SELECT po from PurchaseOrder po WHERE po.status = 'ACTIVE' AND po.customer.address.city = 'Stuttgart'”);

List<PurchaseOrder> orders = query.getResultList();

for (PurchaseOrder order: orders) { order.getCustomer().getName();}

{returns N purchase orders} → 100 positions = 101 SQLs

Better:SELECT po from PurchaseOrder po FETCH JOIN po.customer...

Page 37: Berlin Expert Days 2013 - Performance Talk

Fetch Joins - Example

Performance Issues & Solution Strategies

Query query = em.createQuery(“SELECT po from PurchaseOrder po WHERE po.status = 'ACTIVE' AND po.customer.address.city = 'Stuttgart'”);

List<PurchaseOrder> orders = query.getResultList();

for (PurchaseOrder order: orders) { order.getCustomer().getName();}

{returns N purchase orders} → 100 positions = 101 SQLs

Better:SELECT po from PurchaseOrder po FETCH JOIN po.customer...

{returns N purchase orders} → 100 positions = 1 SQL

→ related objects will be joined into the query instead of being queried independently

Page 38: Berlin Expert Days 2013 - Performance Talk

Batch Reads - Example

Performance Issues & Solution Strategies

Query query = em.createQuery(“SELECT po from PurchaseOrder po WHERE po.status = 'ACTIVE' AND po.customer.address.city = 'Stuttgart'”);

query.setHint(“eclipselink.batch”, “po.customer”);

...

Page 39: Berlin Expert Days 2013 - Performance Talk

Batch Reads - Example

Performance Issues & Solution Strategies

Query query = em.createQuery(“SELECT po from PurchaseOrder po WHERE po.status = 'ACTIVE' AND po.customer.address.city = 'Stuttgart'”);

query.setHint(“eclipselink.batch”, “po.customer”);

...

{returns N purchase orders} → 100 positions = 2 SQLs (one additional for each relationship)

→ subsequent queries of related objects can be optimized in batches instead of being retrieved one-by-one

→ Batch reading is more efficient than joining because it avoids reading duplicate data.

Page 40: Berlin Expert Days 2013 - Performance Talk

Solution Approaches

► Data Retrieval– Read-Only queries (query.setHint(“eclipselink.read-only“, “true“))

● improves performance by avoiding copying and change tracking the objects

– Fetch Joins

– Batch Reads

– Other loading optimizations

● Use projection queries where appropriate● Use pagination for large result sets (query.setMaxResults(), query.setFirstResults())● Use named queries (likely to be precompiled by provider, reusability)

► Updating Data

– Batch Update

● allows a bunch of update operations to be performed as a single DB access● reduces round trips to the database

Performance Issues & Solution Strategies

<property name="eclipselink.jdbc.batch-writing" value="Oracle-JDBC"/>

Page 41: Berlin Expert Days 2013 - Performance Talk

Client

Application Server

Databases

Legacy Systems / Service Provider

Application

Performance Issues & Solution Strategies

Page 42: Berlin Expert Days 2013 - Performance Talk

Legacy Systems / Service Provider

► They are often out ouf our control

► Legacy Systems– often very difficult to troubleshoot legacy systems

– running dinosaurs

– limited insight into those systems

► Service Providers– No influence on them

– SLAs

– What do you do if your PaaS provider has connection issues or is down?

Performance Issues & Solution Strategies

Page 43: Berlin Expert Days 2013 - Performance Talk

How to proactively reduce risk of Performance Issues?

Page 44: Berlin Expert Days 2013 - Performance Talk

How to proactively reduce risk of Performance Issues?

Pragmatic Solution Approach

► 1. From a general point-of-view– Define someone that is responsible for Performance Management in the

project

– Identify Performance Risks early

– Define Performance Objectives (measurable & realistic)

● otherwise there is a risk that objectives are simply ignored because too difficult to achieve

– Conduct Architectural Reviews (continually)

● to find out whether the architecture is really capable of meeting performance objectives

Page 45: Berlin Expert Days 2013 - Performance Talk

How to proactively reduce risk of Performance Issues?

Pragmatic Solution Approach

► 1. From a general point-of-view– Do Performance Tests (before application goes to production)

– Monitor your Application (especially in PreProduction & Production)

● to find out how the application is really used (application usage pattterns)

● to identify trends (important for capacity planning)

– Know your users

Page 46: Berlin Expert Days 2013 - Performance Talk

Pragmatic Solution Approach

► 2. From a technical point-of-view– Know the used technologies

– Always look out for possible performance improvements in those technical areas

● Important: analyze the effects of „improvements“ and „Best Practices“

– Pay back technical debt as soon as possible

– Add small performance tests and not just unit tests (and automate them)

How to proactively reduce risk of Performance Issues?

Page 47: Berlin Expert Days 2013 - Performance Talk

Conclusion

Page 48: Berlin Expert Days 2013 - Performance Talk

Conclusion

Conclusion

► Investing time & money in performance really pays off

► There needs to be someone responsible for APM

► Performance issues can reside anywhere in an architecture– Architectural reviews, performance tests, aware

developers/architects/testers can help in reducing the risk

► From the managements point-of-view it seems that performance engineering seems to cause initially more costs than bringing value

– Problem: difficult to demonstrate success, but poorly performing applications are clearly observable as failures

“Why do we have performance engineers if we don't have performance problems?” by Connie U. Smith

Page 49: Berlin Expert Days 2013 - Performance Talk

[email protected]@etudenhoefner at TwitterPerformance Matters!

http://goo.gl/zwML7

Page 50: Berlin Expert Days 2013 - Performance Talk

Thank you for your Attention!

[email protected]@etudenhoefner at Twitter