modernize rmf/cmf & smf interpretation for superior z/os systems performance management

47
Modernize RMF/CMF & SMF Interpretation for Superior z/OS Systems Performance Management Brent Phillips – Managing Director, Americas Todd Havekost - Senior Performance Consultant

Upload: intellimagic

Post on 13-Apr-2017

21 views

Category:

Data & Analytics


1 download

TRANSCRIPT

1

Modernize RMF/CMF & SMF Interpretation for Superior z/OS Systems Performance Management

Brent Phillips – Managing Director, AmericasTodd Havekost - Senior Performance Consultant

2

Agenda1. Why Modernize2. Technical Demo with Focus on Emerging Data Sources

‒ Rated Metrics and Exceptions‒ SCRT analysis from SMF 89‒ zEDC hardware compression‒ TCP/IP and OSA‒ Capping & Capping Groups‒ Transactions from SMF 72‒ zIIP SMT

3

z/OS Mainframe• Processes hundreds of billions of transactions per day

• RMF/CMF & SMF are an underleveraged strategic advantage of the z/OS platform‒ Far richer source of performance and configuration

data compared to distributed systems and cloud

• “Big Data” before it was cool‒ But analytics on other big data are far more

advanced. SMF report processes are 4 decades old!‒ So Modernize! It is time to dynamically

create better intelligence out of the data

4

1980’s Technology• Data -> PDB -> Static reports • Often requires:

‒ Programming skills for each new report or report change‒ Difficult, manual interpretation after problems occur‒ Different silos of expertise to piece together the big picture

• Does not:‒ Avoid problems by monitoring root causes for early warning‒ Rate metrics as good or bad based on context‒ Correlate physical and logical views in an integrated data model‒ Show related metrics to the issue, allow intelligent drilldowns‒ Answer: What is different? Is it safe? What will break?

5

Modernize for Better Intelligence To Improve IT Operations Decision Making

Decision Stage Status Quo xMF Reporting Modernized Interpretation with AI

Problem Identification

Mostly Reactive - Firefighting Mostly Proactive - Strategic:monitor root causes to predict upcoming problems

Collect Information

Slow and Incomplete, static reportsFast, Deep, Interconnected:

dynamically generated intelligence, correlated logical and physical views, charts rated for performance risk

Design and Choose

SolutionsHarder, Longer, More Mistakes Safer, Faster, using Dynamic Availability Intelligence:

A.I. shows related issues, includes recommendations

6

Modernize for Better Intelligence To Improve IT Operations Decision Making

Decision Stage Status Quo xMF Reporting Modernized Interpretation with AI

Problem Identification

Mostly Reactive - Firefighting Mostly Proactive - Strategic:monitor root causes to predict upcoming problems

Collect Information

Slow and Incomplete, static reportsFast, Deep, Interconnected:

dynamically generated intelligence, correlated logical and physical views, charts rated for performance risk

Design and Choose

SolutionsHarder, Longer, More Mistakes Safer, Faster, using Dynamic Availability Intelligence:

A.I. shows related issues, includes recommendations

7

Modernize for Better Intelligence To Improve IT Operations Decision Making

Decision Stage Status Quo xMF Reporting Modernized Interpretation with AI

Problem Identification

Mostly Reactive - Firefighting Mostly Proactive - Strategic:monitor root causes to predict upcoming problems

Collect Information

Slow and Incomplete, static reportsFast, Deep, Interconnected:

dynamically generated intelligence, correlated logical and physical views, charts rated for performance risk

Design and Choose

SolutionsHarder, Longer, More Mistakes Safer, Faster, using Dynamic Availability Intelligence:

A.I. shows related issues, includes recommendations

8

Modernize for Better Intelligence To Improve IT Operations Decision Making

Decision Stage Status Quo xMF Reporting Modernized Interpretation with AI

Problem Identification

Mostly Reactive - Firefighting Mostly Proactive - Strategic:monitor root causes to predict upcoming problems

Collect Information

Slow and Incomplete, static reportsFast, Deep, Interconnected:

dynamically generated intelligence, correlated logical and physical views, charts rated for performance risk

Design and Choose

SolutionsHarder, Longer, More Mistakes Safer, Faster, using Dynamic Availability Intelligence:

A.I. shows related issues, includes recommendations

9

Modernize for Better Intelligence To Improve IT Operations Decision Making

Decision Stage Status Quo xMF Reporting Modernized Interpretation with AI

Problem Identification

Mostly Reactive - Firefighting Mostly Proactive - Strategic:monitor root causes to predict upcoming problems

Collect Information

Slow and Incomplete, static reportsFast, Deep, Interconnected:

dynamically generated intelligence, correlated logical and physical views, charts rated for performance risk

Design and Choose

SolutionsHarder, Longer, More Mistakes Safer, Faster, using Dynamic Availability Intelligence:

A.I. shows related issues, includes recommendations

10

Modernize for Better Intelligence To Improve IT Operations Decision Making

Decision Stage Status Quo xMF Reporting Modernized Interpretation with AI

Problem Identification

Mostly Reactive - Firefighting Mostly Proactive - Strategic:monitor root causes to predict upcoming problems

Collect Information

Slow and Incomplete, static reportsFast, Deep, Interconnected:

dynamically generated intelligence, correlated logical and physical views, charts rated for performance risk

Design and Choose

SolutionsHarder, Longer, More Mistakes Safer, Faster, using Dynamic Availability Intelligence:

A.I. shows related issues, includes recommendations

11

Modernize for Better Intelligence To Improve IT Operations Decision Making

Decision Stage Status Quo xMF Reporting Modernized Interpretation with AI

Problem Identification

Mostly Reactive - Firefighting Mostly Proactive - Strategic:monitor root causes to predict upcoming problems

Collect Information

Slow and Incomplete, static reportsFast, Deep, Interconnected:

dynamically generated intelligence, correlated logical and physical views, charts rated for performance risk

Design and Choose

SolutionsHarder, Longer, More Mistakes Safer, Faster, using Dynamic Availability Intelligence:

A.I. shows related issues, includes recommendations

12

Better Intelligence using AI

AI(4) = Availability Intelligence for yourApplication Infrastructure providingActionable Insights using Artificial Intelligence.

Dynamically derive intelligence from the datathat improves application infrastructure availability

with predictive and prescriptive insights about performance and configuration issues.

13

Utilize applied artificial intelligence techniques that include embedded, z/OS-specific expert knowledge

‒ Derives new, meaningful metrics out of the raw RMF/SMF data

‒ Normalizes the data to properly correlate related metrics

‒ Automatically Rate metrics as good or bad based on context:• Internal component capacity limits (e.g., port throughput) • z/OS best practices for configuration and performance management• Workload balance and redundancy loss identification• Relationship and interaction of logical and physical resources

How to Generate Better Intelligence

14

Better Intelligence = Better Answers to Difficult IT Ops Questions

• What risky conditions exist right now across our entire environment?

• What related metrics are relevant to the context of this issue?

• Where do go next to see root causes?

• What help is there to create a solutions?

15

Better Intelligence = Better Answers to Difficult IT Ops Questions

• What risky conditions exist right now across our entire environment? (exception charts, rated metrics)

16

Better Intelligence = Better Answers to Difficult IT Ops Questions

• What related metrics are relevant to the context of this issue? (side-by-side mini-charts of related metrics that are clickable)

17

Better Intelligence = Better Answers to Difficult IT Ops Questions

• Where do go next to see root causes? (intelligent drill downs)

18

Better Intelligence = Better Answers to Difficult IT Ops Questions

• What help is there to create a solutions? (built-in recommendations)

19

Executive Support for Modernization• IT Executives are often already prepped for the strategic

benefits of modernizing with AI

‒ For example, for the second year in a row, Gartner Group named Applied AI as the #1 Strategic Technology Trend

‒ Reasons also valid for Applied AIon the IT Operations Data

20

Benefits of using Applied AI on RMF/SMF1. Availability:

‒ Avoid “Accidents” e.g., with predictive views ‒ Faster root cause analysis e.g., with prescriptive insights

2. Skills Gap:‒ Enhance cross-team knowledge and collaboration‒ Accelerate knowledge transfer for newer team members ‒ Prescriptive insights

3. New Use Cases:‒ Business application owner – App Infrastructure Status Dashboard‒ Application Dev team – compare new version infrastructure impact

21

• Good problem to solve with Software as a Service• Different roles in your organization get easy access to custom intelligence• Access to IntelliMagic experts for knowledge transfer, analysis• Solution infrastructure is managed for you, creating more focus

IntelliMagic Vision as a Service (or on premise)

© IntelliMagic 2016

22

Technical Demo

23

IntelliMagic Vision Systems Module – OverviewLegacy• CEC, LPAR, 4HRA (70)• Real Storage, Paging (71)• WLM (72)• Channels (73)• CF, XCF, FICON Dir. (74)• Page Datasets (75)• Virtual Storage (78)• Addr Space, Showback (30)• TCP/IP (119)

Emerging• zIIP SMT (70)• Transaction (72)• PCIe/zEDC (74)• SCM (Flash) (74)• SCRT/Usage (89, 30)• LPAR Topology (99)• Processor Cache (113)

24

• Display peak 4HRA values by productby billing month (2nd through 1st)

• Supports both CMP (single peak)and separate peaks by CEC

• Drilldowns into‒ CPU use during peak 4 hour periods‒ Address spaces registering each product

SCRT and Usage Data

25

zEDC Hardware Compression• New dashboard and ratings• Card utilizations & service times• Compression ratios• Request rates & throughput• Identify exploiters• Enhanced address space

drilldown capability

PCI Express and SCM Cards

26

• TCP/IP binds and connectionsby socket and port

• Aggregated TCP/IP metricsfor all interfaces and users

TCP/IP

27

• Supports both capacity group and LPAR (“soft”) capping• Displays relationships between key metrics• Visibility into capping enhanced in version 8.10• Added Capacity on Demand support

Capping and Cap Groups

28

• Newly-introduced transactionlevel reporting in RMF 72 recordsdisplays key metrics byservice class and report class

• Previously required processingtransaction level records‒ CICS 110s, DB2 101s, IMS log‒ Often 10s or 100s of millions

Transaction Reporting

29

Service Class / Report Class Transactions

• Key transaction metrics provided (by service or report class)‒ Transaction rate‒ Transaction response times‒ CPU per transaction (and total CPU)‒ I/O per transaction (and total I/Os)‒ Average number of concurrent transactions

30

• Multithreading metrics

• Metrics on relationshipsbetween zIIPs and CPs

• zIIP management metrics(similar to GCPs)

zIIP SMT and Other zIIP Reporting

31

zIIP MT

• New metrics when multithreading is activated for zIIPs‒ Productivity‒ Capacity Factor (CF)‒ Maximum Capacity Factor (MCF)‒ Thread Density (TD)‒ Core Busy

32

33

34

35

36

37

38

zIIPs and CPs

• Reports on relationships between the 2 types of processors• zIIP-eligible work executing on GCPs reported by CEC,

system, workload, and service class• New metric “zIIP Efficiency %” reflects % of zIIP-eligible

workload executing on zIIPs

39

40

zIIP Management

• Reports to enhance management of zIIP configuration and optimize processor cache efficiency

• Extends many concepts from GCPs, such as weights, RNI, polarity, LPAR overhead, and capture ratios

• New metric “zIIP Target %” reflects % of total processor workload executing on zIIPs

41

42

• 3 sets of reports aggregatingPerformance Index (PI) values‒ By Workload‒ By WLM importance level‒ By System

WLM Goals

43

IntelliMagic Vision Systems Module – OverviewLegacy• CEC, LPAR, 4HRA (70)• Real Storage, Paging (71)• WLM (72)• Channels (73)• CF, XCF, FICON Dir. (74)• Page Datasets (75)• Virtual Storage (78)• Addr Space, Showback (30)• TCP/IP (119)

Emerging• zIIP SMT (70)• Transaction (72)• PCIe/zEDC (74)• SCM (Flash) (74)• SCRT/Usage (89, 30)• LPAR Topology (99)• Processor Cache (113)

44

POC / MLC Assessment Offer

45

POC / MLC Assessment Offer

Purpose:‒ Identify unrealized MLC reduction opportunities‒ Discover visibility provided by IntelliMagic Vision into your data

Process:‒ Send IntelliMagic Data‒ Analysis by IntelliMagic Experts‒ Presentation of Results

Cost: No charge for qualified sites in North America

46

Questions?

47

Thank You

To request a no-charge POC or MLC assessment:Intellimagic.com/MLC

Contact us with any questions or feedback:Phone: 214-432-7920

Email: [email protected]