Why Can`t We All Just Get Along?

Download Report

Transcript Why Can`t We All Just Get Along?

Why Can’t We All Just Get Along?
Glen Daniels
Progress Software
Progress Software
Application Platforms
OpenEdge
Apama
2
Business Application Platform
Integrated platform optimized for business
Event Processing
Market leading event processing platform
Sonic
Actional
Shadow
Enterprise Service Bus
Market leading ESB
SOA & Web Services Management
Leading enterprise-class SOA management
Mainframe Integration
First unified mainframe integration platform
DataDirect
DataXtend
EasyAsk
ObjectStore
Data Connectivity
Market leader in data connectivity
Data Integration
First model-driven semantic data integration
Natural Language Search and Query
First and leading natural language query
Object Data Management
Market leading object database
Services Infrastructure
Data Infrastructure
© 2006 Progress Software Corporation
A Little About Sonic
 We’re all about Integration
•
•
•
•
•
Message-centric (async)
Highly distributed
Legacy / Mainframe
Data adaptation
Orchestration
 “Connect everything… achieve anything”
3
© 2006 Progress Software Corporation
Different Strokes
 We live in a multi-protocol world
• JMS, FTP, UDP, Jabber, etc.
 Not always necessary to switch…
• Services rarely move from one to another
 …but necessary to adapt!
• You never know who you might be talking to
• Data formats, protocols, semantics
4
© 2006 Progress Software Corporation
External or Internal?
 Sometimes integrating units within the same
company can be “external”
• Regulatory compliance requires IT diffs
• Organizational opacity
 Bringing in a new acquisition can make
formerly disparate systems “internal”.
• Newly shared security domains, registries, etc
 Shared models can really help here
• WSDL, Bindings
5
© 2006 Progress Software Corporation
The Promise
 Web Services as common “glue” for talking to
a variety of platforms
 Wire-level interoperability to avoid the “many
adapters” issue
 A composible set of specs which smoothly
take you from simple to richly-featured
 Distributed extensibility and evolution
 Interop… also WRT Skills and Tools
6
© 2006 Progress Software Corporation
The Reality
 Lack of community architectural consistency
• No standard for how to write/use extensions
• No real framework for bindings with various
capabilities
 Interoperability problems
• Whither Soapbuilders/WS-I?
 Perception as bloated, vendor-controlled
 …and some great successes too
7
© 2006 Progress Software Corporation
URIs
 URI == Endpoint?
• Metadata about resources is critical
– Content-type is NOT enough
• …but you shouldn’t require EPRs to get it
 Single URI -> single resource
• Need stronger curb on use of RefPs
• Policy, Sem Web assertions, etc.
 Benefits of URIs are well-understood, let’s not
throw them out
8
© 2006 Progress Software Corporation
Bindings
 We love bindings!
• Design a service in the abstract (types,
operations)
• Bind to particular formats, policies, addresses
 Great for managing change
• Evolution
• Apply policies
9
© 2006 Progress Software Corporation
A Common View : Benefits
 You get a lot from a common service model
•
•
•
•
•
Manageability / governance
Registries
Interoperability
Tooling (codegen, orchestration)
Composibility (intermediaries/mediation)
 Is WSDL it?
10
© 2006 Progress Software Corporation
We’re Not Highlanders
 “There can be only one” –
NOT!
 There is a place for WS next
to the Web
 Can we get to a place where
enterprise “ility-enabled” apps,
over a variety of underlying
protocols, gain from what
we’ve learned from the Web?
• URIs, Proxies, Uniform IF
11
© 2006 Progress Software Corporation
So Where Do We Go From Here?
 Collect “enterprisey” use cases
• Build taxonomy and best practices
• Demonstrate the same story with WS and
REST?
• Picking the right tool for the right job
 Push harder for architectural coherence
• EPRs/URIs – please let’s get the word out
• Is “composibility” a myth?
• Versioning
 Interop work?
12
© 2006 Progress Software Corporation