1 / 20

A Research Agenda for Business-Driven IT

A Research Agenda for Business-Driven IT. Jeff Kephart (IBM Research) Steve White (IBM Research) Edie Stern (IBM Tivoli). Business-Driven IT. We want a world in which businesses can respond flexibly to opportunities and threats Flexible business requires flexible IT ….

naiara
Download Presentation

A Research Agenda for Business-Driven IT

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. A Research Agenda forBusiness-Driven IT Jeff Kephart (IBM Research) Steve White (IBM Research) Edie Stern (IBM Tivoli)

  2. Business-Driven IT • We want a world in which businesses can respond flexibly to opportunities and threats • Flexible business requires flexible IT … 2

  3. Is this flexible IT? No – that’s what autonomic computing is supposed to fix! 3

  4. The role of autonomic computing • Autonomic computing systems are: • “Computing systems that manage themselves in accordance with high-level objectives from humans.” • A Vision of Autonomic Computing, IEEE Computer, J. Kephart and D. Chess, Jan. 2003. • How high is “high”? • Business-driven IT: the high-level objectives are business objectives 4

  5. Business Architects Business Process Tools & Transforms Deployers and Domain Experts Business Objectives and Metrics Platform-Independent Models Business to IT Tools & Transforms Human Expertise Convert objectives Automatically provision, deploy Automated Provisioning and Deployment DCM IT Admins DB policies High-level IT metrics, objectives Storage policies Network policies Towards Business-Driven IT Monitored business data Business Objectives (e.g. KPIs) There will be continuous feedback between IT and business levels to calibrate business-to-IT transformations Business Process Models Translation of models, metrics and objectives from business terms to IT terms will become increasingly automated Human specification of low-level, platform-specific policies gives way to high-level discipline-specific objectives with tradeoffs Policies Self-managing IT system 5

  6. High-level IT metrics, objectives Server Policy Network Policy DB Policy Availability Management Security Management Service Level Management Change Management Information Lifecycle Mgmt. From siloed policies to high-level IT objectives Server experts Network experts and tools Database experts and tools Application Experts Mainframe experts Workstation experts Availability Objectives Security Objectives Performance Objectives Change Objectives • Replace resource-oriented silos with horizontal process-oriented solutions • Replace resource-oriented policies with objectives defined by management discipline; aggregate them 6

  7. High-level IT metrics, objectives Models capture human expert knowledge about dependence of high-level IT metrics on lower level system knobs and observables like demand l. They can be refined automatically. Policies Application Manager Composing utility with models yields an optimization problem in terms of low-level params that can be posed to an appropriate optimizer. Models Perf. Model Service-level utility Resource-level utility Avail. Model Cost Model RT(cpu, b; l) DT(b) Cost(cpu) NetUtil(cpu, b; l) Util(RT, DT) Optimizer eBrokerage transactions application 1 sec response time for Gold customers is OK. I don’t need faster than 0.75 sec; more than 2.0 sec is unacceptable. 50 min downtime/month is tolerable; 100 min is bad. Good DownTime is slightly more important than good RT. l cpu, b Objectives are defined as utility function for Response Time and Down Time, which captures tradeoffs On Demand Env 1 Scenario: Managing to Performance and Availability Objectives The system can now set these parameters to their optimal values, or advise a human administrator. 7

  8. Performance-Availability Tradeoffs using Utility Functions with J. Strunk, B. Salmon, G. Ganger, CMU Cost Function for Trace Processing Application Cost ($/yr/student) $15000 Outage renders student 50% effective + sys admin spends 100% time fixing; costs $45/hr $10000 $5000 $20000 $25000 $30000 Student waits for run on 27GB trace file once per day; costs $30/hr Availability Bandwidth (MB/sec)

  9. E-commerce preference elicitation methods that help consumers to express complex tradeoffs will be adapted to systems administration Need interfaces and algorithms to support elicitation of high-level objectives WebSphere XD uses templates to elicit average or percentile response-time objectives 9

  10. WAS XD Utility Function Combination

  11. WebSphereXD-TIO Data Center FreePool WebSphereXD WebSphereXD AAMAS-06: Hakodate , Japan

  12. Control parameters • Suppose we have just two control parameters • cpu = # processors • b = data backup interval (in minutes) • We want to choose (cpu, b) to optimize Util(rt,rpo) – Cost(rt,rpo) • We need to transform the utility function into control parameter space • We can do this using models that relate (cpu,b) to (rt,rpo) NetUtil(rt, rpo) = Util(rt,rpo) – Cost(rt,rpo) = Util( rt(cpu, b; l),rpo(cpu, b ; l))–Cost(cpu, b) = NetUtil(cpu,b; l)

  13. RT(msec) l = 10-3 l = 5*10-3 rt(cpu, m; l) m(sec-1) m=10 l = 10-4 cpu b NetUtil(rt, rpo) = Util(rt,rpo) – Cost(rt,rpo) = Util( rt(cpu, m(b); l), rpo(b))–Cost(cpu) = NetUtil(cpu,b; l) Models Models could come from: Analytics/Queuing, Simulation, Machine Learning m(b) rpo(b) Cost(cpu)

  14. Net utility as function of control parameters NetUtil(cpu, b; l) Util(rt, rpo) l=0.01 U Unet rpo cpu rt b rpo cpu rt b

  15. rt b*=2.05265 cpu*=8.58375 U*=75.8644 rt*=88.6853 b*=1.19931 cpu*=3.65144 U*=137.414 rt*=95.4449 b*=0.874575 cpu*=2.49134 U*=152.661 rt*=99.5775 cpu cpu cpu l=0.002 l=0.01 l=0.05 b b b Net utility vs. control parameters Util(rt, rpo) rpo NetUtil(cpu, b; l)

  16. Challenges at IT level • Elicit high-level IT objectives • Manage to them • Interactive effectively with administrators to build trust

  17. Business Process Tools & Transforms Business Objectives and Metrics Platform-Independent Models Business to IT Tools & Transforms Human Expertise Convert objectives Automatically provision, deploy Automated Provisioning and Deployment DCM High-level IT metrics, objectives Towards Business-Driven IT Monitored business data Business Objectives (e.g. KPIs) Business Process Models Policies Self-managing IT system 17

  18. LdNode1 LdNode1 LdNode1 LdNode1 LdNode1 LdNode1 Logical Topology Model Deployer VLAN1 VLAN1 VLAN1 Insert Firewall Insert Firewall Insert Firewall Connection Connection Connection Firewall Firewall Firewall DB Server APP Server WEB Server VLAN2 VLAN2 VLAN2 Developer LdNode2 LdNode2 LdNode2 LdNode2 LdNode2 LdNode2 Deployment Topology Logical Application Structure DB JSP JSP DB EAR EAR Domain Expert Model Transformations (Best practices) Automatically Combine Fine-Grained Best Practices Patterns To Transform the Logical Application Structure to a Physical Topology End-To-End Model-Based & Goal-Driven Deployment Eilam et al. (IBM TJ Watson) • Physical • Complete • Correct • Actionable Tivoli Provisioning Manager “Rainforest” Deployment Design Tool Rational Software Architect

  19. Automated derivation of thresholds and goals from SLOs Breitgand, Henis, Shehory (IBM Haifa) • Use statistical techniques to correlate SLO violations at Application Layer with monitored data in System Layer • Automatically set alert thresholds to desired false-positive / false-negative tradeoff BusinessInt Payroll App… App Server StorageServer DB2Server Disk Controller Disk Controller Originally presented at ICAC ‘05

  20. Implications • Human specification of low-level platform- and resource-specific parameters and policies will be phased out. • Administrators will specify power, performance, availability and security objectives, and acceptable tradeoffs between them. Algorithms and interfaces for eliciting high-level IT objectives will emerge, as will standards for expressing them. • Models will capture human expert knowledge of how high level objectives relate to lower-level system parameters, and they will be refined automatically via feedback. • Resources will employ models in conjunction with optimization and planning technologies to manage to multiple objectives, both for deployment and runtime operations. • The entire stack of business-driven IT will be completed, as business objectives get transformed to high-level IT objectives that drive deployment and runtime operations. Standards for expressing business-level objectives will emerge. 20

More Related