How to Build an Enterprise Software Delivery Platform on AWS
The decision to migrate IT operations to the cloud is not always simple, particularly at the enterprise level. Your migration roadmap must include a clear picture—not only of what you will be moving, but also where it will go, and what it will look like when it gets there. In this post, we will try to provide the outlines of that picture for an enterprise software delivery platform—revealing how it would operate, and what services it would provide to the enterprise.
Basic Enterprise IT and the Cost of On-Premises
At the most basic level, such a platform needs to provide core enterprise IT and business software services. Traditionally, of course, enterprise software delivery has been handled on-premises. Enterprise IT has often centered around a patchwork of local servers, often of varying ages and capabilities. For a long time, of course, this was the only kind of IT that was possible—the result, more often than not, is a mixture of legacy and reasonably modern resources (both hardware and software), which provide the enterprise with indispensable services, but at relatively high cost in terms of maintenance, availability, and data compatibility.
Typically, the scope of those services will extend down to the regional, district, and local office level. Ideally, those services should have a degree of integration, or at least interoperability at the level of data-sharing and key business functions; the reality in traditional enterprises, however, is more likely to include fragmentation of services and partial interoperability mediated by ad hoc patches.
Moving to the Cloud
For an increasing number of enterprises, the IT upgrade path means full or partial migration to the cloud, rather than an investment in new on-premises resources. Cloud-based IT provides some clear advantages: elimination of most or all hardware upgrade and maintenance costs, unified (or at least highly integrated) platforms for development and deployment, and a high level of availability.
Cloud migration also provides the opportunity for (and the challenge of) complete reorganisation, integration, and modernization of enterprise IT and business software services.
Amazon's Web Services division has been moving forcefully into the enterprise IT market; AWS currently provides full or partial IT hosting for a number of enterprise-level clients, and it is continually introducing new enterprise-oriented services.
For an existing enterprise, the first step in the actual process of moving to the cloud is migration. To aid in migration at the enterprise level, Amazon provides a variety of migration tools. For large volumes of data, these range from high-speed connections to large-volume physical transfer.
AWS Snowball, for example, uses high-volume portable data-storage appliances to ship encrypted data from the enterprise site to an AWS data center. For even larger volumes of data, Amazon provides Snowmobile, a specially equipped tractor-trailer rig with massive, high-speed data transfer and storage capabilities.
AWS also provides a range of flexible services for migrating databases into the cloud. These services include migration to a cloud-based instance of your existing database platform, as well as migration to a new platform. AWS supports the majority of well-known commercial and open-source database platforms, as well as its own MySQL-compatible database, Amazon Aurora.
Amazon also has programs and services for the migration of other key enterprise applications and suites, such as SAP, to AWS. In addition, Amazon provides a full range of native AWS enterprise software services, including Business Intelligence, typically at a cost that is much lower than the equivalent services from established service providers.
Development and Deployment
AWS includes a wide range of development and deployment tools, as well as a variety of platforms, such as EC2 (for full-service, server-based cloud deployment) and Lambda (for serverless, container-based deployment). The high level of interoperability makes it not only possible but natural to combine these services—seamlessly.
AWS makes it easy to deploy a combination of standard enterprise and industry-specific software, along with applications developed in-house, in an environment which supports an extremely high level of integration. Needless to say, AWS is also an ideal environment for the full range of DevOps practices, including containerization and continuous delivery.
Putting Up with Delays
The bottom line is that migrating to AWS is likely to make modernization of an enterprise's software delivery system not only easy, but inevitable. It is not uncommon for on-premises IT systems to include major inefficiencies and bottlenecks, particularly at the enterprise level.
Typically, users not only put up with these inefficiencies, they also assume that they are the natural way of doing things, or simply become oblivious to them. The result is that such things as long turnaround times or delays in access to key services become institutionalized, and the inefficiencies of on-premises IT become inefficiencies built into the enterprise itself.
Migration to AWS and the subsequent refactoring of enterprise IT which accompanies that migration automatically expose these inefficiencies and provides ready-made opportunities and channels for remedying them. Developers, managers, and users all discover that the delays and inconveniences which they had accepted as natural and inevitable were simply artifacts of the ad hoc practices which had arisen out of the need to keep a patchwork on-premises IT system functioning.
When those inefficiencies are removed, the enterprise as a whole becomes more efficient, and key processes which had operated at an artificially slow rate (and often at considerable expense and opportunity cost) naturally accelerate.
- General Electric's Oil and Gas division, for example, has been able to reduce multiple invoices to a single per-customer invoice—a process that, while it may seem simple to the customer, represents a major increase in behind-the-scenes efficiency. GE Oil and Gas also makes regular use of the Snowball data transportation system to move huge quantities of crucial pipeline diagnostic data to AWS for analysis.
- Kaplan, Inc., a major provider of educational tools and services, was able to move its extremely data-intensive test preparation division IT center to AWS, using a combination of Oracle tools and AWS-based services to produce a highly integrated cloud-based system. In addition, this allowed Kaplan to close 75% of its data center facilities, and sell off its legacy IT equipment in the process.
- Unilever moved its extensive web-based marketing system to AWS, allowing them to consolidate a variety of disparate and incompatible hosting systems into a unified system with full interoperability. The resulting increase in efficiency allowed Unilever to reduce its launch time for a campaign from two weeks to two days.
So what does your enterprise IT system look like on AWS? Fully integrated. Fully modernized. Highly efficient. With the integrated services you need to innovate securely at speed and scale.