16
Performance Engineering Services CLOUDBEARINGS © 2015, CloudBearings CloudBearings Performance Engineering Services

CloudBearings - Performance Engineering Services, V1.0

Embed Size (px)

Citation preview

Page 1: CloudBearings - Performance Engineering Services, V1.0

Performance Engineering Services

CLOUDBEARINGS© 2015, CloudBearings

CloudBearingsPerformance Engineering

Services

Page 2: CloudBearings - Performance Engineering Services, V1.0

What is Performance EngineeringProcess and Intent to test and tune software for realizing the required performance

Broader Aspects – Speed, Scalability and Stability

© 2015, Cloud Bearings

Advantages• Increased revenues by increasing optimum system performance• Optimized utilization of infrastructure and resources• Improved availability by improvising performance and stability• Reduced maintenance cost• Proactive approach towards unexpected system failures• Safeguard against unplanned code or configuration changes

Page 3: CloudBearings - Performance Engineering Services, V1.0

Key ChallengesPerformance Engineering today

© 2015, Cloud Bearings

Page 4: CloudBearings - Performance Engineering Services, V1.0

Performance Tuning

and Optimizatio

n

Performance Testing

and Sizing

Capacity Planning

System Diagnostic

© 2015, Cloud Bearings

Key Elements Mainly four key elements

Test ExecutionsRegression Tests, Load Test and

Benchmarking Tests

Determine Prod. Capacity

System Sizing as per stands, Facilitate Capacity Management

Optimal Guidelines – Web/App/JVM/DB/OS

End to End Profiling, Recommendations towards

scalability

Page 5: CloudBearings - Performance Engineering Services, V1.0

© 2015, Cloud Bearings

Distributed Apps Key Elements towards gathering Performance Statistics

Page 6: CloudBearings - Performance Engineering Services, V1.0

© 2015, Cloud Bearings

User Experience Tracking Key User Experience Metrics

Page 7: CloudBearings - Performance Engineering Services, V1.0

© 2015, Cloud Bearings

Capacity Planning Doing the right sizing for varying system loads

Page 8: CloudBearings - Performance Engineering Services, V1.0

End to End Performance Testing and Analysis for Web Applications and Platforms

Identifying System Sizing Requirements based on Site Traffic, Average Page Views and Number of Orders.

Simulating close to real environment using process automation scripts against key goals/objectives.

Identifying all relevant base lining criteria’s

Analyzing and Profiling of the system and application to identify potential bottlenecks and validate performance SLA compliance by leveraging best technologies and tools.

Performance Tuning and Optimization based on outcome from iterative analysis cycles.

Final Report Compilation and Submission

“ The overall activity has dependency on various external factors. i.e. 3rd party services which are performing low and becoming bottleneck for overall applications; Images etc.

Applicable Conditions

CloudBearings Performance Engineering

Page 9: CloudBearings - Performance Engineering Services, V1.0

EXECUTION FOR KEY GOALS / OBJECTIVES

RIGHT SERVER SIZING Optimal no of Web Servers corresponding to number of users Optimal ratio Web Server vs. App Server Max users per server and max number of transactions per serverSERVER OPTIMIZATIONS

Performance metric on a specific / given hardware Tuning Vendor defaults for optimal performance Identifying resources which needs tuningCAPACITY PLANNING Determine current server (s) capability Identify room for growth, what h/w, s/w to leverage towards achieving optimum levels Is there excess capacity? An excess server can be removed ?

THIRD PARTY VALIDATIONS What is the current ISP and what is the capacity ? Can the ISP deliver on the SLA agreement ?

SECURITY EXPOSURE Identifying system vulnerabilities and minimizing the same System susceptibility against DoS attacks How effective are the recovery procedures ?INFRASTRUCTURE Browser/ User Profile issues – What do the user do ?, When do they do ?, How fast ?, Major geographic locations, is application browser-interface dependent ?, Identifying asynchronous communication and non-HTTP communications. Is modem, WAN, LAN emulation necessary ?

Page 10: CloudBearings - Performance Engineering Services, V1.0

EXECUTION FOR KEY GOALS / OBJECTIVES, contd ..

INFRASTRUCTURE Peripheral Systems Issues – is the LAN/WAN dedicated/shared with other applications?, are there any shared production resources?, are there any web pipes, ERP, mail servers, file systems, DNS servers etc.?, does it share databases with other systems?, does it share hardware with other systems? Etc. External Systems Issues – are there outside vendors providing CDN services?, Distributed Hosting Issues – are there multiple mirrored sites?, is any site configured for failover?, How is traffic load balanced across sites?, are there architecture components on shared WAN connections?, what is the failover and recovery behavior etc. Firewall Issues – what is the throughput capacity?, what is connection capacity & rate?, DMZ operations, throughput policies from a single IP, connection policies from a single IP etc.; Intrusion Detection Systems – is there statistical content sampling?, is there any inverse relation b/w throughput and security?, How is content filtering achieved? Etc.;

Internet issues – Peering issues with Client’s hosting/bandwidth provider Site Web-Pipe Issues – How much bandwidth site(s) have and are there multiple web pipes? Border Router Issues – What sort of load balancing multiple pipes configured for?, Does they use same inbound and outbound pipe?, is there the same outbound pipe regardless of inbound pipes?, is there equal distribution of outbound regardless of inbound pipes?, Are there multiple border routers?, failover configuration for multiple border routers etc. Load Balancing Issues – Algorithm used i.e. round robin, sticky IP, least connections, subnet based, weight based etc., timeout for LB table, Does it do any connection pooling?, is it doing any content filtering?, is it checking for HTTP response status?, are there application dependencies with LB timeout settings?, what failover strategies are employed?, connection persistence timeout, time out for critical functions etc.

Page 11: CloudBearings - Performance Engineering Services, V1.0

EXECUTION FOR KEY GOALS / OBJECTIVES, contd ..

APPLICATION Database Server Issues – have both small/large datasets tested?, connection pool configurations, and upper limits etc..

Web Server issues – How many connections, the server can handle?, How many open file descriptors/handles , processors/threads configured?, release/renew of connections happening correctly?, How large is the listen queue?, server’s page push capacity?, type of caching employed, is there any page construction done here?, is there dynamic browsing?, What type of server side scripting is done i.e. java script, ASP, JSP, PHP, Perl?, any acceleration/content caching devices in front of web server?, can server extensions and there functions be validated?, monitoring pools, connections, processes, threads, queues, sessions etc.; Application Server Issues – is there any page construction done here?, how is session management done and capacity there?, any clustered configurations?, any load balancing done?, do components have a specific interface to peripheral and external systems?

SECURITY Firewall(s) & Multiple DMZ(s) – does firewall(s) do content filtering?, is it sensitive to inbound/outbound traffic?, upper connection limits, any policies attached for specific IP in terms of num connections or throughput?, are there multiple firewalls (DMZ’s), is yes, are they sensitive to data content? Etc.; Intrusion Detection Systems – is there any content filtering?, is the system sensitive to inbound/outbound traffic?, alert thresholds, acceptable security thresholds etc;

Page 12: CloudBearings - Performance Engineering Services, V1.0

Sample – System Sizing Requirements, needed on a high level..

System Sizing Requirements

Value

Average expected MBps of Traffic per day

 

Average page views per day expected

 

Average orders per hour expected

 

Peak expecting MBps of Traffic per day

 

Peak dynamic page views per second(or lowest time increment – please specify)What is the average session length…In page requests or clicks?In time?Peak orders per hour expected

Page 13: CloudBearings - Performance Engineering Services, V1.0

CloudBearings Performance Engineering Approach

Page 14: CloudBearings - Performance Engineering Services, V1.0

Tools and Software'sPerformance Testing

 Jmeter, LoadRunner, Grinder, Silk Performer

Web Servers  Apache, IIS, App Servers  Apache Tomcat, JBoss,

WebLogic, WebSphere, Pramati

Databases  Oracle, MySQL, MSSQL, Postgres SQL

System Profiling

JProbe, JProfiler, HotJar, DynaTrace, YourKit, AppDynamics, New Relic

Scripting Sed, Awk, Unix Shell

Tools and Software's, we are used to …

Page 15: CloudBearings - Performance Engineering Services, V1.0

Custom Reports, Recommendations & Dashboards

We help optimize the performance of commerce web and native mobile applications. It is possible from high-level overview of your environment, to drill down into the nitty-gritty details, in order to identify bottlenecks that are slowing your apps down, so you can fix them before users are affected with a negative experience. Deep visibility in to the cause of

application performance issue Browser Monitoring for consistent

user experiences App-Map for External Services Custom Dashboards Custom Dashboards for Web and Mobile

App(s)1

Top Transactions and Histograms2 Transaction Drill downs3

Page 16: CloudBearings - Performance Engineering Services, V1.0

Performance Engineering Services

CLOUDBEARINGS© 2015, CloudBearings

Thank You