gave a presentation


Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

gave a presentation

  1. 1. Managing the Impact of SOA and Enterprise 2.0 on Financial Services IT Hugh Grant Director Global IT R&D Credit Suisse
  2. 2. Where we’re coming from: Enterprise 1.0 <ul><li>Heavily siloed; built with familiar models: </li></ul><ul><ul><li>Database-centric; Client-server / n-tier; Messaging / pub-sub </li></ul></ul>DB Client App Server Client Client Client BUS App Server Client Client Client Client App Server App Server DB DB Client DB DB App Server App Server App Server Mainframe App Server
  3. 3. Where we want to be: Enterprise 2.0 <ul><li>All the SOA goodness: </li></ul><ul><ul><li>client-side mashups allow end users to tailor access to shared services aligned to business needs </li></ul></ul><ul><ul><li>business logic tied in to workflow and service orchestration </li></ul></ul><ul><ul><li>back-end services hosted in a utility computing fabric or cloud </li></ul></ul>Service Container Client Mash-up Service Service Workflow / Orchestration Layer Service Container Service Service Service Container Service Service Service Container Service Service Client Mash-up Client Mash-up Client Mash-up Client Mash-up Utility Compute Cloud AJAX AJAX Composite Service Composite Service Composite Service AJAX AJAX AJAX AJAX AJAX AJAX AJAX AJAX
  4. 4. Where we are: Enterprise 1.1 <ul><li>Starting to see web services used within individual silos </li></ul><ul><ul><li>But is this really SOA? It still looks a lot like Enterprise 1.0 </li></ul></ul>DB Java Client J2EE Server Web Client Web Client Client BUS App Server Java Client Java Client .NET Client .NET Client .NET Server J2EE Server DB DB .NET Client DB DB .NET Server App Server .NET Server Mainframe CORBA Server Java Client J2EE Server .NET Client .NET Server .NET Server SOAP SOAP SOAP CLASSIC CLIENT-SERVER .NET Server J2EE Server CORBA Server SOAP SOAP WRAPPER SOAP AS INTEGRATION GLUE Web Client Web Client App Server SOAP SOAP SOAP AS WEB GLUE
  5. 5. What’s Happening? <ul><li>Clearly, broad adoption of shared services is a good thing, so what’s holding it back? </li></ul><ul><li>Line-of-Business groups will no longer own the whole front-to-back process, so we need to solve several really hard problems to ensure they can still deliver to their customers: </li></ul><ul><ul><li>Technical issues: </li></ul></ul><ul><ul><ul><li>Security, Architectural Governance, Development Lifecycle, etc. </li></ul></ul></ul><ul><ul><li>Non-technical issues: </li></ul></ul><ul><ul><ul><li>Culture, incentives for cross-silo cooperation, risk management, etc. </li></ul></ul></ul>
  6. 6. Enterprise 2.0 and SLAs <ul><li>Service Level Agreements — SLAs are essential to managing business-critical functions </li></ul><ul><ul><li>Without SLAs, how can you know if the SOA will meet business needs? </li></ul></ul><ul><ul><ul><li>Funny thing is, SLAs aren’t mentioned in most SOA discussions </li></ul></ul></ul><ul><li>Existing systems tend to have implicit SLAs between components </li></ul><ul><ul><li>Easier to manage the end-to-end SLA when you stay inside the silo </li></ul></ul><ul><li>Problems arise when refactoring systems as services </li></ul><ul><ul><li>Implicit SLAs in existing apps need to be identified and made explicit </li></ul></ul><ul><ul><ul><li>Tight coupling gives way to loose coupling between client, workflow, and services </li></ul></ul></ul><ul><li>Mashups and workflow magnify the problem </li></ul><ul><ul><li>Can no longer make assumptions about how a service is consumed, or by whom </li></ul></ul><ul><ul><ul><li>In effect, each end user desktop can drive a unique application </li></ul></ul></ul>
  7. 7. Modelling the Enterprise SOA with SLAs <ul><li>Use SLAs to model the behaviour of each level in the SOA </li></ul><ul><li>At each boundary in the SOA, we should have an SLA </li></ul><ul><ul><li>Top-level SLA will be expressed in business terms </li></ul></ul><ul><ul><ul><li>transaction throughput, availability, number of concurrent users, etc. </li></ul></ul></ul><ul><ul><li>Underlying services will have SLAs stated in more technical terms </li></ul></ul><ul><ul><ul><li>message throughput / latency, number of concurrent connections, etc. </li></ul></ul></ul><ul><ul><li>Model client behaviour too </li></ul></ul><ul><li>One size does not fit all </li></ul><ul><ul><li>Different users will require different SLAs (at different price points!) </li></ul></ul>
  8. 8. SLA… MIA? <ul><li>Today there isn’t much support for modelling and managing SLAs throughout an SOA development lifecycle </li></ul><ul><ul><li>Focus tends to be more on interface and topology </li></ul></ul><ul><li>SOA is about service </li></ul><ul><ul><li>Current Web Service management tools are focused on deployment </li></ul></ul><ul><ul><ul><li>No linkage between client development tools and deployment fabric </li></ul></ul></ul><ul><li>Need to model SLAs throughout the lifecycle </li></ul><ul><ul><ul><li>From mash-up through to service deployment in the cloud </li></ul></ul></ul><ul><ul><ul><li>Govern consumers as well as providers </li></ul></ul></ul>
  9. 9. SOA and SLAs <ul><li>We end up constructing a sort of SLA graph </li></ul><ul><li>For each end-to-end flow we can analyse the graph </li></ul><ul><ul><li>Identify mismatched or missing SLAs before going into production </li></ul></ul>Service Container Client Mash-up AJAX Service Workflow / Orchestration Layer Composite Service Service Container Service Service Container Service Service Container Service Service Client Mash-up Client Mash-up Client Mash-up Client Mash-up Utility Compute Cloud AJAX Composite Service Composite Service Service Service Service Identify end-to-end flow AJAX AJAX AJAX AJAX AJAX AJAX AJAX AJAX AJAX Composite Service Composite Service Service Service Service SLA SLA SLA SLA SLA Check SLAs at each point
  10. 10. From SLAs to Policy-Driven Computing <ul><li>Quality of Service SLAs are only one part of the puzzle </li></ul><ul><ul><li>Also need to consider: </li></ul></ul><ul><ul><ul><li>Security issues (Authentication, Identity, Encryption, Entitlements, etc.) </li></ul></ul></ul><ul><ul><ul><li>Risk / Compliance issues (Auditing / Logging, Data Protection, etc.) </li></ul></ul></ul><ul><ul><ul><li>Etc. </li></ul></ul></ul><ul><li>Moving to a policy-driven model will help to contain complexity </li></ul><ul><ul><li>But still have to embed it explicitly into the SOA lifecycle </li></ul></ul><ul><ul><li>Future tools should walk the policy graph and identify policy mismatches </li></ul></ul><ul><ul><ul><li>Alert to problems before they are running in production </li></ul></ul></ul><ul><li>WS-Policy gives us a framework, but </li></ul><ul><ul><li>We need common standards for QoS and other policies and SLAs </li></ul></ul><ul><ul><li>Policy support must become a standard part of Enterprise SOA platforms and tools </li></ul></ul>
  11. 11. SLAs and Cloud Computing <ul><li>A lot of effort going into building clouds of utility computing power </li></ul><ul><ul><li>On demand computing, next-generation service fabrics, etc. </li></ul></ul><ul><li>SLAs and Policies become ever-more important the closer you get to utility computing </li></ul><ul><ul><li>basis for pricing models; chargeback </li></ul></ul><ul><ul><li>SLAs give confidence that cloud model is manageable </li></ul></ul><ul><ul><li>SLAs can be mapped onto infrastructure and support tiers </li></ul></ul><ul><ul><ul><li>e.g., automatically deploy services onto appropriate h/w based on SLA </li></ul></ul></ul><ul><li>The bottom line: </li></ul><ul><ul><li>Senior management won’t outsource to the cloud unless they are totally sure of what they are getting </li></ul></ul>
  12. 12. Thank You!