Amazon re:Invent 2018: Say Goodbye to Best-of-Breed and Hello to Best-of-Need Applications

B. Shimmin

Summary Bullets:

  • At its annual re:Invent 2018 conference, Amazon Web Services rolled out a blinding number of micro-specialized solutions that emphasize ‘best of need’ over ‘best of breed.’
  • But, there’s a danger lurking in this seeming freedom to adopt and combine capabilities at will, namely a new form of vendor lock-in.

At Amazon Web Services re:Invent 2018 this past week, attendees were treated to an avalanche of product launches and pre-release announcements. On display were three new data management services, four new Internet of Things (IoT) capabilities, eight new storage offerings, and thirteen new machine learning (ML) libraries — all designed to encourage developers to build and deploy solutions on the Amazon Web Services (AWS) platform. And that’s just the software dealing with big data and analytics.

Why is Amazon inundating its enterprise customers and development partners with such a blind array of products? As it turns out, the answer has little to do with Amazon itself and more to do with changes in the way software is produced and consumed.

Once upon a time in the high tech industry, it was said that “he with the most APIs wins.” The notion behind that adage made complete sense back in the early noughties, when large-scale, monolithic apps roamed the confines of on-premises data centers. At that time, when updates were few and far between, APIs were the preferred means of making a given application do what the customer needed, connecting the application with outside systems or adding new functionality, for example.

These APIs are often the lifeblood of independent software developers, value-added resellers, and systems integrators, all of which earn their living by making square pegs (monolithic apps) fit into round holes (unique businesses). But, from the vendor’s perspective, APIs are difficult to build, document, and maintain. Unfortunately, they also slow innovation, making it difficult for the vendor to evolve the application without breaking (or deprecating) any mission-critical API calls currently in use among customers.

With the global adoption of server and process virtualization (and with that, containerization), the continuing success of open source software (OSS), the rise of utility-based pricing, and the use of application development techniques tuned to a bi-weekly update cycle, the very nature of application software itself has changed. Monolithic applications are officially on the way out and are being replaced, not by a suite of products, but rather by a loose affiliation of like-minded capabilities that can be readily arranged and connected to complete the task at hand.

For Amazon and other hyper-scale public cloud platform providers (Microsoft, IBM, Google, SAP, Oracle, et al.), the trick isn’t in creating best-of-breed software that can be enriched via API, but rather the creation of as many distinct ‘capabilities’ as possible. Often this rising tide of capability includes best-of-breed software from other vendors, even rivals. This makes the platform and its partner ecosystem the actual product. Customers and partners buy into the vendor, not a given solution.

As an example, at re:Invent 2018, Amazon rolled out Amazon Managed Blockchain, a managed service intended to help developers create and maintain blockchain networks using their choice of OSS frameworks, Hyperledger Fabric or Ethereum. The product here isn’t the ledger itself but instead the ability to run a ledger like Ethereum or Hyperledger on top of Amazon with just a few clicks. More importantly, customers can do so in conjunction with synergistic capabilities like Amazon’s Amazon Quantum Ledger Database (QLDB), a fully managed ledger database, which was itself announced at re:Invent 2018.

Obviously, if it only takes ‘a few clicks’ to stand up a hyperledger on AWS, it is just as likely that a similar number of clicks is all that will be required to move said hyperledger to a competing cloud platform. That’s the danger of this new best-of-need approach to software. It lowers the barriers to both adoption and abandonment equally.

Where will all of this lead? We will continue to see rapid innovation and expansion from a very small handful of hyper-scale platform players. And we will continue to see innovation coming from smaller players (e.g., the popular OSS database, MongoDB), but those players will find the most success running atop that small cadre of vendors. The downside to all of this is ironically a lack of choice. Customers may find it easy to abandon a service, feature, or capability, but abandoning one platform for another will not be so easy.

What do you think?

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.