Modeling and Analysis: System Model

Transcription

1 by Gerrit Muller University of South-Eastern Norway-NISE Abstract This presentation uses a web shop service as example system to construct a system model. The caching of s of the products in the shop is modeled to analyze performance, robustness, scalability and reliability of the system. Distribution This article or presentation is written as part of the Gaudí project. The Gaudí project philosophy is to improve by obtaining frequent feedback. Frequent feedback is pursued by an open creation process. This document is published as intermediate or nearly mature version to get feedback. Further distribution is allowed as long as the document remains complete and unchanged. status: preliminary draft load in seconds 1 2 quantified mid office example th = 0.02 ms tm = 2 ms na = 1000 to be measured hit rate also! Load(h) = 1000 * 2[ms] * h * 1.98[ms] Load(h) = * h [ms] utilizable capacity Hit rate is context dependent. Life cycle changes or peak loads may degrade hit rate. 0.5 Hit rate of well designed system is ample within working range (e.g. 95%) 0 th order formula is valid: Load = 0.12 * na [ms] working range hit rate 1

2 Content of Modeling and Analysis System Model What to model of the system content Stepwise approach to system modeling Non Functional requirements (NFR), System Properties and Critical Technologies Examples of web shop case 2 Gerrit Muller MASMcontent

3 What to Model in System Context? usage context system enterprise & users NFR's: performance reliability availability scalability maintainability... (emerging?) properties: resource utilization load latency, throughput quality, accuracy sensitivity (changes, inputs)... critical technologies caching load balancing firewalls virtual networks XML for customization and configuration... creation life cycle business life cycle context 3 Gerrit Muller MASMwhatModeling

4 Approach to System Modeling 1. determine relevant Non Functional Requirements (NFR's) 2. determine relevant system design properties 3. determine critical technologies 4. relate NFR's to properties to critical technologies 5. rank the relations in relevancy and criticality 6. model relations with a high score 4 Gerrit Muller MASMapproach

5 Web Shop: NFR s, Properties and Critical Technologies system NFR's: performance browsing initial cost running costs reliability/availability scalability order rate maintainability effort product changes effort staff changes security (emerging?) properties: resource utilization load, capacity memory load, capacity response latency redundancy order throughput product data quality product definition flow staff definition flow security design compartimentalization authentication encryption critical technologies caching load balancing pipelining virtual memory memory management data base transactions XML for customization and configuration firewalls virtual networks... 5 Gerrit Muller MASMwhatWebShop

6 4. Determine Relations system NFR's: performance browsing initial cost running costs reliability/availability scalability order rate maintainability effort product changes effort staff changes security 4 (emerging?) properties: resource utilization load, capacity memory load, capacity response latency redundancy order throughput product data quality product definition flow staff definition flow security design compartimentalization authentication encryption 4 critical technologies caching load balancing pipelining virtual memory memory management data base transactions XML for customization and configuration firewalls virtual networks... 6 Gerrit Muller MASMrelationsWebShop

7 5. Rank Relations NFR's: performance browsing initial cost running costs reliability/availability scalability order rate maintainability effort product changes effort staff changes 5 security system (emerging?) properties: resource utilization load, capacity memory load, capacity response latency redundancy order throughput ranking will be discussed in Modeling and Analysis: Reasoning product data quality product definition flow staff definition flow security design compartimentalization authentication encryption 4 critical technologies caching load balancing pipelining virtual memory memory management data base transactions XML for customization and configuration firewalls virtual networks... 7 Gerrit Muller MASMrankingWebShop

8 Purpose of Picture Cache Model in Web Shop Context screen screen screen client client client response time browse products data base web network network exhibit products product descriptions required capacity required capacity logistics ERP financial cache customer relations 8 Gerrit Muller MASMwebShopPictureCache

9 Zero Order Load Model zero order web load model Load = n a * t a n a = total requests t a = cost per request 9 Gerrit Muller MASMloadZeroOrder

10 First Order Load Model first order web load model Load = n a,h *t h + n a,m *t m n a,h = accesses with cache hit n a,m = accesses with cache miss t h = cost of cache hit t m = cost of cache miss n a,h = n a * h n a,m = n a * (1-h) n a = total accesses h = hit rate Load(h) = n a * h* t h + n a * (1-h) * t m = n a * t m - n a * h * (t m - t h ) 10 Gerrit Muller MASMloadFirstOrder

11 Quantification: From Formulas to Insight t h = 0.02 ms quantified mid office example load in seconds 2 t m = 2 ms n a = 1000 Load(h) = 1000 * 2[ms] * h * 1.98[ms] Load(h) = * h [ms] 1 utilizable capacity working range 0.5 hit rate 1 11 Gerrit Muller MASMquantified

12 Hit Rate Considerations load in seconds 2 quantified mid office example t h = 0.02 ms t m = 2 ms n a = 1000 to be measured hit rate also! Load(h) = 1000 * 2[ms] * h * 1.98[ms] Load(h) = * h [ms] Hit rate of well designed system is ample within working range (e.g. 95%) 0 th order formula is valid: Load = 0.12 * n a [ms] 1 utilizable capacity working range Hit rate is context dependent. Life cycle changes or peak loads may degrade hit rate. 0.5 hit rate 1 12 Gerrit Muller MASMdiscussion

13 Response Time human customer press next look client web request check cache request store in cache transfer to client transfer process display data base retrieve t 0 t t t t t t t time in milliseconds in optimal circumstances 13 Gerrit Muller MASMresponseTime

14 What Memory Capacity is Required for Picture Transfers? screen screen screen client client client browse products cache mid office network exhibit products required memory capacity? network back office product descriptions logistics ERP financial customer relations 14 Gerrit Muller MASMwebShopPictureFlow

15 Process View of Picture Flow in Web Server one copy per process multiple copies per process multiple copies per process and thread mid office web mid office web mid office web web web k cache cache cache cache cache m back office access back office access back back back office office office access access access n 15 Gerrit Muller MASMmidOfficeProcesses

16 Formula memory Use Web Server multiple copies per process and thread memory = 3 * n * s + 5 * m * s + c * s + web web web web k 3 * k * s where n = # data base access threads m = # cache threads k = # web threads s = size in bytes c = in memory cache capacity in # s cache cache cache back back data office office base access access access m n 16 Gerrit Muller MASMmidOfficeFormula

17 Web Server Memory Capacity use case n m k s c MB storage type memory = 3 n s + 5 m s + c s +3 k s small shop highly concurrent large s many s all at once memory use product browsing only s only single , , ,962 9,540 98,234 L3 main main main+disk main+disk disk processor caches where n = # back office access threads m = # cache threads k = # web threads s = size in bytes c = in memory cache capacity in # s L1 L2 L3 main memory disk kb MB GB What is the performance impact of memory use on other processing? 17 Gerrit Muller MASMmidOfficeQuantified

18 We Have only Modeled a Small Part of the System... function browse/exhibit products sales, order intake, payments track, order handling stock handling financial bookkeeping customer relation management update catalogue advertize after sales support data structured (product attributes, logistics,...) program code aspect memory use response time load network use reliability any resource, any NFR aspect result = d = all data f = all functions aspect(d, f) ignoring other dimensions such as applications, users, circumstances 18 Gerrit Muller MASMsystemSpace

19 ... to Understand Some of the Systems Aspects static mostly assumptions and coarse estimates some insight in: what are key design issues what are relevant use case areas 19 Gerrit Muller MASMconclusion

20 Refinement After Context Modeling usage context enterprise & users NFR's: performance reliability availability scalability maintainability... system (emerging?) properties: resource utilization load latency, throughput quality, accuracy sensitivity (changes, inputs)... critical technologies caching load balancing firewalls virtual networks XML for customization and configuration... creation life cycle business life cycle context 20 Gerrit Muller MASMfollowUp

21 Summary Conclusions Non Functional Requirements are the starting point for system modeling Focus on highest ranking relations between NFR's and critical technologies Make simple mathematical models Evaluate quantified instantiations Techniques, Models, Heuristics of this module Non functional requirements System properties Critical technologies Graph of relations 21 Gerrit Muller MASMsummary