The lifeblood of the financial services industry is information, and I have found financial services firms to be early, innovative technology adopters. Â Electronic trading exchanges have delivered business services â€˜in the Cloudâ€™ for many years, and financial specific compute grids deliver actionable market intelligence (e.g. piCloud, RiskMetrics).Â As Cloud computing moves beyond the definition phase and technology infrastructure vendors deliver viable products, advanced Cloud computing concepts are starting to Occupy Wall Street.
The Wall Street Technology Association held a “Cloud Computing in Financial Services” seminar in Boston on 27, October 2011.Â Open discussion and presentations illustrated how Cloud 101 concepts can be applied to the financial industry.Â The content described Cloud benefits, adoption challenges, viable use cases, and recommended solutions.
Cloud computing has an ambitious mandate (e.g. agility, speed, lower cost) achieved through the following transformational strategies:
- Lower cost by using commodity hardware and improving hardware utilization
- Self-service automation delivers agility, rapid deployment, and reduced time to market.
- Standardized environment images and services enhance solution recovery
- Reduce operational cost through automation
Charlie Babcock talked about a â€˜Cloud Revolutionâ€™ based on owning little IT infrastructure and lowering IT barrier to entry.Â How many organizations are using Cloud to transform their software application and service development?
Complex legacy environments, immature processes, and stringent enterprise requirements hold back many organizational transformations.Â Barriers to cloud adoption are perennial IT pain points:
- Performance unpredictability
- Security: data confidentiality and auditability
- Regulation, compliance, controls, and privacy
- Data transfer bottlenecks and economics
- Service levels and quality of service
- Existing legacy investment
- Integration limitations
- Complex applications
- Sophisticated operational models
- ITIL processes
- Data and vendor lock-in
Seminar participants voiced not simply abdicating responsibility, and participant concerns focus on cultural change and internal discipline:
- Change development model
- Extend development and operations governance
- Enforce security policy across distributed and heterogeneous environment
- Develop joint operations models
- Manage diverse portfolio
To successfully pursue a Cloud strategy and quickly realize benefits, teams must carefully choose Cloud entry points, architecture, and products.Â Participants noted a few real world concerns:
- Public cloud economics are often negative when compared to using existing infrastructure assets in a high, long-lived compute usage scenario.
- Application interdependencies can prevent migration to Cloud
- Promoting standard environment and gaining adoption (lag time)
- RAM is the biggest scaling and sizing constraint.Â For example, Java heap stack or OS size.
Use Cases and Entry Points
Wisely choosing a first project and demonstrating benefits is extremely important. Speakers mentioned the traditional entry point list:
- Improving agility within development and test environments
- Improving agility within transient, non-mission critical application environments. For example, Development/Test environments, marketing, training and e-learning
- Enabling scalable web application/content services with bursty workloads. For example, rich media encoding/decoding and consumer facing mobile applications
- Supporting high performance computing, Big Data processing, and complex analytics
IT stakeholders will become service brokers who focus on executing the following use cases:
- Optimize workloads and capacity
- Manage service lifecycle
- Ensure compliance
- Maintain quality of service
- Measure, report, and audit environment
and Cloud infrastructure must be tuned to fit workload parameters:
- Technical architecture
- Service levels
Optimum tuning requires applying self-service, automation, flexible provisioning and configuration, multi-tenant segregation, dynamic workload balancing and resource pooling, granular consumption models to existing heterogeneous environment. Â The WSO2 Stratos platform auto-tunes today; reducing Cloud complexity and lowering barriers to success.
Heterogeneity will remain, and enterprise organizations must create a hybrid cloud environment.Â How can a hybrid cloud model reduce cost, simplify administration, and provide a flexible, future profit platform?Â Â Teams require a consistent, standardized operations and development model, and integrated services.
Speakers from IaaS vendors Eucalyptus (David P. Butler), Riverbed (Joe Ghory), and RackSpace (Andrew Schroepfer) described using IaaS Web Services to overlay virtual cloud resources (i.e. compute, network, storage, identity).Â Additional required core capabilities include physical resource management, virtual resource system, and virtual+physical resource administration.Â The Cloud automation environment must deliver provisioning, resource pooling, and policy enforcement capabilities across resource groups and resource zones (i.e. cloud burst zone, primary cloud zone)
Today, many solutions deliver â€˜push a button, get a boxâ€™ provisioning.Â Standardized, simple infrastructure environments (i.e. isolated compute, network, storage) can be quickly provisioned.Â The solutions commonly lack tooling integration, centralized login, development and operations workflow, and grouping resources into a complete solution (e.g.Â compute, network, storage, identity store, security controls, master data, business process, user interfaces, and integration connections).Â A telling quote from the seminar, â€œ9 minutes to provision a box, 9 days to configure the application platformâ€. Â With WSO2 StratosLive, teams can provision and configure an application platform in 5 minutes.
The end goal offers a Cloud self-service portal and sophisticated solution catalog backed by automated provisioning and automated service management.Â Infrastructure authority and cloud orchestration concepts are emerging.Â I had interesting conversations with RackSpace and Eucalyptus about tightly integrating the WSO2 Stratos Platform as a Service (PaaS) offering with open Infrastructure as a Service (IaaS) controllers.
A unified solution delivering multi-tier and complex service provisioning (security, network connectivity, context, application and data platform) with policy-based intelligent deployment (e.g. associate workloads and use cases with service catalogue offerings and available use cases) will provide a firm foundation for enabling ecosystem platforms based on Vertical PaaS. Â During a seminar presentation, Jonathan Marsh of WSO2 described how new web business models rely on platforms and APIs instead of web sites. Financial service firms are interested in delivering business capabilities as a service and creating an ecosystem platform interconnecting partners, suppliers, and customers. Vertical PaaS capabilities address key management, service versioning, access control, lifecycle management, service level agreements (SLA), monetization, and consumer tracking issues commonly encountered when delivering vertical service platforms.
To the stratosphere and beyond,