Fro Ntier Stress Tests Status report Luis Ramos

  • Slides: 22
Download presentation
Fro. Ntier Stress Tests Status report Luis Ramos LCG 3 D Meeting - August

Fro. Ntier Stress Tests Status report Luis Ramos LCG 3 D Meeting - August 17, 2006

Agenda 1. 2. 3. 4. 5. 6. Objectives Test Plan Hw/Sw Setup Test Cases

Agenda 1. 2. 3. 4. 5. 6. Objectives Test Plan Hw/Sw Setup Test Cases - Results Future Work Conclusions Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 2/22

Objectives • Develop a benchmark for Frontier servers – DB schema independent • Build

Objectives • Develop a benchmark for Frontier servers – DB schema independent • Build a tool that identifies performance bottlenecks of a given setup • Performance analysis of the complete sw stack – – CORAL / Frontier plugin Frontier Client Squid Frontier Servlet Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 3/22

Test Plan • How fast are the individual components? – – Database Application Server

Test Plan • How fast are the individual components? – – Database Application Server Cache Server Network • Explore performance impact of: – Different data (content, size, storage type, compression) – Different complexity database schemas – Different caching policies • How do DB throughput, network bandwidth, payload size, # of clients or server CPU correlate? Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 4/22

Metrics and Parameters • Metrics: – – – Individual and total throughput Server errors

Metrics and Parameters • Metrics: – – – Individual and total throughput Server errors CPU consumption/load (clients, frontier server, squid server) Memory usage and disk space needs Network bandwidth usage • Parameters: – – – # of client nodes # of test clients Payload sizes Database structure and content Caching policy Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 5/22

Fro. Ntier Test Setup Luis Ramos – August 17 th, 2006 Fro. Ntier Stress

Fro. Ntier Test Setup Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 6/22

Fro. Ntier Test Setup • Hardware setup: – 1 server running Frontier & Squid:

Fro. Ntier Test Setup • Hardware setup: – 1 server running Frontier & Squid: • • Dual Intel Xeon CPU 2. 80 GHz 2 Gb RAM HD 150 Gb Fast Ethernet (100 Mbps) – 1 Backend Oracle Database 10 g. R 2 (cooldev) • Software Setup: – Fro. Ntier v 3. 1 (need to check keep_alive feature in v 3. 2) – Frontier Squid v 1. 0 rc 4 Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 7/22

Client Setup • Hardware: – Dedicated lxplus nodes • Dual Pentium III 1 GHz

Client Setup • Hardware: – Dedicated lxplus nodes • Dual Pentium III 1 GHz • 500 Mb RAM • HD 6 Gb • Fast Ethernet (100 Mbps) • Software: – CORAL_1_5_1 with Frontier. Client v 2. 4. 7 (hacked to print some metrics) – C++ CORAL/Frontier. Plugin test • Queries the server • Gathers results • Outputs measures – Python controller script • Starts a number of clients in several client nodes • Gathers measures • Generates structured data for plotting Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 8/22

Test Cases developed • Data Types analysis • Payload Sizes analysis • Network analysis

Test Cases developed • Data Types analysis • Payload Sizes analysis • Network analysis • Throughput analysis – Includes Database, Fro. Ntier and Squid analysis Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 9/22

Data Types analysis • CORAL static mapping • C++ type Oracle 10 g type

Data Types analysis • CORAL static mapping • C++ type Oracle 10 g type int NUMBER(10) float BINARY_FLOA T double BINARY_DOUB LE std: : string(100 ) VARCHAR 2(10 0) coral: : Blob BLOB If DB schema is not created with CORAL: – User may define a different mapping Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 10/22

Payload Compression • Std: : string(100) in C++ >> VARCHAR 2(100) in DB filled

Payload Compression • Std: : string(100) in C++ >> VARCHAR 2(100) in DB filled with random numbers • Network payload size in % of user data size: (for different compression levels) Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 11/22

Payload Size • Client side payload size evolution – – – (HTTP) XML ascii

Payload Size • Client side payload size evolution – – – (HTTP) XML ascii size (Frontier. Client) some bytes down to BLOB base 64 (Frontier. Client) 33% down to BIN size (Frontier. Client) up to uncompressed BIN size (CORAL/Frontier. Plugin) down to user C++ data types • With std: strings, user data is ~ the same as DB size Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 12/22

Network Analysis • Test tool that checks network performance between multiple given client nodes

Network Analysis • Test tool that checks network performance between multiple given client nodes and a single server… – Generates TCP/IP between clients and server – Each client shows a throughput and a CPU% • Done using the netcat utility Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 13/22

Throughput analysis Frontier Server • Up to 150 clients running against a single server

Throughput analysis Frontier Server • Up to 150 clients running against a single server (direct Fro. Ntier server access, no Squid involved) Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 14/22

Throughput analysis Server error rate Depending on the size of the payload, up to

Throughput analysis Server error rate Depending on the size of the payload, up to 80 clients can be served from one Fro. Ntier node (no caching) Frontier v 3. 2 keepalive feature may reduce drastically the error rates Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 15/22

Throughput analysis Network, Squid and Frontier • TCP/IP vs Frontier Server vs Squid Cache

Throughput analysis Network, Squid and Frontier • TCP/IP vs Frontier Server vs Squid Cache Hits Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 16/22

Throughput Analysis Compression level • Zipping performance (single client) • Zip 5 throughput is

Throughput Analysis Compression level • Zipping performance (single client) • Zip 5 throughput is lower because: – CPU time is higher (both on server and client) – Network is not the bottleneck (can be in other scenarios) Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 17/22

Throughput Analysis CPU% and load average • How CPU% correlates with server total throughput

Throughput Analysis CPU% and load average • How CPU% correlates with server total throughput payload = 30 k. B Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 18/22

Throughput Analysis Fro. Ntier. Plugin vs Oracle. Plugin • CORAL Frontier. Plugin vs Oracle.

Throughput Analysis Fro. Ntier. Plugin vs Oracle. Plugin • CORAL Frontier. Plugin vs Oracle. Plugin (same CORAL code, different connect string) Single client: To do: plot multi client results Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 19/22

Future work • Run multi client throughput tests changing: – DB data type, structure

Future work • Run multi client throughput tests changing: – DB data type, structure and contents – Access method (Oracle vs Frontier) – Frontier configuration (compression level and caching) • Frontier server error rate analysis – Easily identify the limits of a given setup • Squid caching – Influence of partial caching of the queries • Tests from outside CERN – Run throughput tests from outside to measure the influence of a poorer network connection Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 20/22

Conclusions • Some values already collected • Test scripts developed – Next step: turn

Conclusions • Some values already collected • Test scripts developed – Next step: turn scripts easily reusable by others • Real world data needed! – “real” data and a estimate of the ratio between cached and uncached accesses would allow to: • make much more precise statements about compression and achievable rates – Who can provide these estimate/measurement to us? • Further results and analysis – 3 D workshop in September Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 21/22

Questions? Ideas? Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests -

Questions? Ideas? Luis Ramos – August 17 th, 2006 Fro. Ntier Stress Tests - 22/22