Managing Applications Status of Standards Today February 5 2003 Open Group Members Meeting San Francisco Karl Schopmeyer
Jan 17, 2016
Managing ApplicationsStatus of Standards
Today
February 5 2003
Open Group Members Meeting
San Francisco
Karl Schopmeyer
An Application Manageability Group
Objectives Today Determine what such a group can contribute
to application management Determine if we want to make such a group
work. Determine what the basic charter for such a
group is Determine who we want involved.
User side Supply side
The changing IT environment
Business Awareness Shift Tactical: Management of IT resources Strategic: Optimization of Business
Fundamental architecture shift From
Distributed, n-tiered topology Difficult and expensive to manage Big footprint Marginally reliable computing Many I/O media
To Distributed, n-tiered topology Rich, policy-based management Reduced footprint Highly scalable, reliable computing Converged I/O medium
Application transparency
Network-based, loosely coupled tiers enabled bydistributed computing and management technologies
Loosely coupled tiers enabled bydistributed computing and management technologies
Today’s Challenges
New web-based applications introduce unpredictable loads on IT infrastructure, increasing cost and problems
Complex, heterogeneous, n-tiered application environments are difficult and costly to manage
Lacks insight into application resource inventory and utilization, and how the state of application infrastructure affects business performance
Unable to meet user expectations for Application Quality of Service Systems management offerings have failed to deliver on their
promises Business priorities are not reflected in IT infrastructure. Management is not important until after a system is developed and
installed.
Fundamental OS Shift
• Limited instrumentation• No control• Limited visibility
Result:1. Configurators and Monitors2. Ineffective Policy Mgt
• Rich instrumentation• Complete control• Great visibility
Result:1. Measure, analyze, and Affect2. Powerful Policy Mgt
What is Application Management?
Means different things to different people Deployment Management Configuration Management Fault Management Resource Management Performance Management Service Level (QoS)
Management Business Service
Management Operational Control
Different Views Of Application Management Business Management Service Management Fault Management Etc.
The Changing Requirements The user requirements are changing faster than
the solutions technologies. Users want business and Service management The suppliers are still trying to instrument Users want to integrate applications across
platforms Suppliers don’t really have a model for applications
management Architectures are changing
Dynamic, Components, runtime integration Integrating WEB Services
Manage the business, not the technology But we cannot manage the business until we
manage the technology. Modeling the “management” components of the
Business is even more difficult than managing the technology
Applications are becoming largely integration We still cannot instrument the base, thus cannot
manage the integration
BusinessManagement
ServiceManagement
ApplicationManagement
DeviceManagement
En
d-t
o-E
nd
IT
Man
agem
ent
Climbing The Management Mountain
Resources-Devices
Resource - Device Management
Systems Management
Applications
Business
The Users
SLA, QOS Business
Managmenet
The User requirements are growing
The suppliers are significantly behind the user requirements
We cannot build on “empty air”. Management functionality must be a growing infrastructure.
Solution Requirements
Measure, analyze, and affect the entire application environment
Affect and visualize in real-time, dynamic infrastructure changes and understand their effects on the business
Provide integrated dashboard to measure business productivity against objectives
Install, maintain and operate easily Be non-disruptive Standards-compliant Secure Reduce management TCO; deliver fast and measurable ROI Common and interoperable information.
Why Standards?
Common Instrumentation Difficult to instrument for management
Nobody works on management until the end of the project
Manage across components Need common data
Integrate management information from multiple components into common management views The user wants to manage the environment, not
islands Instrument for what we want to manage
Service Level, Business Management, etc. Models must drive instrumentation
Why is modeling Important?
Common Information The right Information
Ex. Service Level management needs information
Putting Semantics on information Create common Semantics, not simply
common syntax Provide interoperability of information Providing a management “abstraction”
Some people simply do not understand the value of data models
We have not convinced the world why they are important yet.
Management info Models In Context
Applications
Applications need management APIs
ManagementApplicationsManagementApplicationsManagementApplicationsManagementApplicationsManagementApplications
ManagementModels
And Information
Manageability InterfaceCommon protocols and
Information
Common understanding of protocols, syntax,
and semantics
ApplicationsApplications
Applications
Manageability/Management Interface
Why is Application Management Important?
Managing systems is not managing the applications that run on them
Getting information from applications and controlling them is the key to the next step, managing services and managing the IT business.
Applications are becoming inherently “multi-system” Users buy IT to do work, not to look at their OS. OS performance, quality, etc. is often not a direct
indication of application performance, quality, etc.
Some inhibitors There are so many of them They are different, every one of them Everybody has a different view of what it means to manage
applications Applications are becoming highly dynamic with very late binding. Management is an afterthought There is very little “long-lasting” instrumentation support for
applications. Many key applications are legacy – We will never touch them
again but they will run for years.
Some Laws of Nature•You can’t push on a rope•Water flows downhill•Management only becomes important when it’s too late
Separating the Problem
Management vs. Manageability We want to define manageability so that
applications can be managed in a common open manner
Good manageability will drive good management
Components of managing Applications Lifecycle – Deployment management RunTime Management
The Standards Players Today
The Open Group APIs
ARM API AIC API XSLM – License Mgt API
QOS DMTF
Modeling Application deployment App Server Runtime Database model (based on SNMP
MIB) Application runtime management Unif of Work Metrics Policy
SNIA SAN based open management
IETF SNMP
App based MIBs XMLConf Polichy Mgt
Global Grid Forum Generalized view of
resources and their management
Oasis Web services management
W3C Web services management
architecture Java JCP Process
JMX Interface JMX – CIM model mapping Etc.
TMF QOS Modeling
And yet today we really cannot effectively manage applications, much less services or business in
an open interoperable manner.
Management Lifecycle
initiallife cycle
deployable installable executable runningappstatus
multiplicity
sub-model
1:n 1:n 1:n
transport
setup
installation
runtime(structure)
A Runtime Model
initiallife cycle
executable runningappstatus
sub-model
installation
runtime model
function
systems
structure
dataexternal systems
workflow = default+rollback+exception
best practices (tasks):routine (daily/weekly/...)
configurationanalysis
workflows
config
ura
tion
indica
tion
sh
istory
methods
Application Architecture: Views and Elements
SystemSystem
ProcessingElements
ProcessingElements
CodeComponent
CodeComponent
ResourceResource
ActionAction DataData
External SystemExternal System
OS/HostOS/Host
Logical
Structure
OS
DataFlow
Scenario
LifecycleTime: shipping -> running
SoftwareService
SoftwareService
Application View Application ElementApplication Element
Models and Instrumentation
Measuring Unit of Work Performance ARM Unit of Work Metrics
Interfacing General Instrumentation models to CIM AIC – Application Instrumentation and Control JMX Characteristics of the models
Dynamic management objects Application defined management objects (defined by the
business) Requirements
Strong on semantics Dynamic creation of “definitions” Metrics and operations.
Key Interoperability Interfaces
CIM Object Manager
CIM Object Manager
CIM ProvidersCIM Providers
ApplicationApplication
Manageability to Manager•Multiple management systems•Common open manageability
Object Manager / Providers• Multiple Providers•Encourage common providers
Management System
Management System
EnterpriseManagement
Console
EnterpriseManagement
Console
ApplicationApplicationApplicationApplication
ApplicationApplication
Provider / Resource Interface•Protect Applications•Make application management easy
Some of the Issues
No clear understanding of the importance of models to management
Instrumentation and models are not connected ARM and Unit of work metrics are different
App management is two different worlds Technical management Management of the services delivered
Application management has to be built on lower layers
There are REAL Opportunities here
APIs Service Level / QOS TOG is unique in having a real user input
Requirements Generation Integrating other Standards Defining application management End-End
with specific goals Ex. Perforamance, availabiliy, etc.
NOTE:
However we will only draw interest if we: Do something Do something that is of interest to some of the
supplier community Do something so that there are results in a
short measureable time Work with the other standards groups in this
area, not by ourselves.