Let me break it down and at the same time preserve the “gestalt” of what’s going on in cloud today. I’m leaving a lot out, but I’m taking a lot in at the same time….
API Management / SOA Capability
One of the major forces in enterprise today is the need to move to a “platform” approach to services. And a technology that may assist with the discovery, management, security, and ecosystem around a platform, is API management. Some firms you may have heard of include Layer 7 Tech (recently acquired by CA), as well as Apigee, and Mashery. Regardless of whether you expect to deploy in the public cloud, a key capability to enable developers to program to a set of standard services is the ability to catalog, monitor, and build a standard API even where existing systems and services may not have an intuitive API, or an API at all.
An example of a company that has been very successful leveraging APIs to change the market data business is Xignite. Xignite’s SOA API for market data is well worth examining. I know Stephane Dubois, CEO personally. He presented several times to xxxxx, xxxxxx, and xxxxx and he does a great job describing the architecture and capabilities of his SOA market data platform. And it’s a serious, critical availability and performance platform that delivers over 5 billion requests per month, which is an application enterprise understands. And it runs in AWS.
Private Cloud in AWS
My experience over the past three years has been with deployment and also defining the infrastructure and web services required to succeed in the Cloud. When I say Cloud, I mean AWS, but not Public Cloud. Rather we deployed Production applications with PII data to Virtual Private Clouds and ensured that all PII data is encrypted, and also that none of our instances was directly accessible via the public Cloud. In some cases we deployed the web and application components in a virtual private cloud, but for latency reasons we ran the database at a managed service private with a fiber optic link to AWS.
Third Generation Configuration Management
Cloud and the demand for an agile, resilient infrastructure has carried forward true innovation in the domain of application configuration management, often referred to as DevOps. Firms like xxxxx and xxxxxxx and hedge funds are leveraging tools like Puppet and Chef to substantially simplify the management of their on premise and off-premise infrastructure.
My goal is to identify the simplest solution and make it work–like the process that we’ve had in place since around xxxxx. However, if you or others have questions about current practices and tools I’ve got direct experience with these tools and practices, as well as ties to the DevOps/Configuration management community.
Big Data / Hadoop / Columnar Databases / RedShift
At xxxxxx we found the ability to load billions of records into RedShift compelling, as well as the ability to query large datasets, and to manage the data with existing tools like Informatica, Microstrategy compelling.
(courtesy of blog.aggregateknowledge.com)