Reflections From a Week with AWS in Las Vegas

AWS re:Invent is Amazon’s premier, annual cloud computing conference held in Las Vegas. Kingland has attended for three consecutive years and during that time we have witnessed the explosive growth of conference attendees – from the 30,000 range in 2016, to 40,000 in 2017, and clocking in at approximately 53,000 attendees in 2018, per AWS CEO Andy Jassy’s tally during his keynote speech. And although I didn’t previously believe my name to be all that common, there were enough attendees that upon onsite registration during this year’s conference, I had to choose the correct “Matt Good” from a listing of three. The incredible growth in attendees is certainly parallel evidence of AWS’ dominance in the worldwide infrastructure-as-a-service (IaaS) public cloud services market, as detailed by Gartner. At about 52% market share, AWS completely dominates the next four on the list (Microsoft, Alibaba, Google and IBM).

Kingland has invested significantly in both cloud computing and development-security-operations (DevSecOps) capabilities over the past three+ years, as part of our Kingland Platform investment and in our partnership with a number of clients on specific solutions. AWS has been our primary cloud partner during that timeframe, and each year at the re:Invent conference we look forward to a number of new release announcements and takeaways that will benefit our Platform and our clients. Our attendance at the 2018 conference was no different and yielded the following key takeaways:

• With the release of Amazon Textract, AWS continues to move up the chain of functionality within the Text Analytics space. Kingland’s primary investment in the overall AI space continues to be strongest within the Text Analytics range of capabilities for extracting data and insights out of unstructured content, and with this new release from AWS, we’ll be testing its capabilities through the AWS preview process. Through our Text Analytics investment we’ve seen an incredible variety of content, from the easiest-to-process text and HTML to the most-difficult-to-process grainy, scanned images of documents with complex tables embedded within, so it will be interesting to test Textract on a variety of content for assessing its extraction accuracy. As noted in my blog regarding our Scholar componentry within our Text Analytics capabilities, we’ve invested in a variety of open source usage and proprietary capabilities for accurately extracting data from unstructured content. Stay tuned as we conduct our Textract preview and compare its accuracy against our existing capabilities and look for complementary ways to take advantage of Textract where beneficial.
• My blog detailing our Collector componentry within our Text Analytics capabilities discusses our Data Lake architecture and leverage of AWS S3, Elasticsearch and DynamoDB. With the release of AWS Lake Formation, AWS continues to make the process of setting up and securing a Data Lake even easier. We’ve invested in perfecting the architecture, deployment, scalability and security of our Data Lake over the past year, and presumably the concept of Lake Formation is an attempt at accelerating those past efforts of our experience going forward for any AWS customers looking to more rapidly set up a Data Lake. As with any AWS release that appears to be in parallel with capabilities we’ve built, we’ll be evaluating it through the preview process to see where Lake Formation could provide complementary capabilities that benefit our Data Lake architecture.
• DynamoDB holds a special place within Amazon history as Amazon CTO Werner Vogels reminded the keynote audience during the 2018 conference. He shared a story regarding his worst day at Amazon 14 years ago when a massive load of Christmas shoppers on Amazon.com brought their previous Oracle infrastructure to its knees and caused significant downtime across the site. This great challenge led to their innovation with DynamoDB and the future of highly resilient, scalable, and even purpose-built persistent data stores. AWS’ announcement of DynamoDB’s Support for Transactions extends its capabilities even further, allowing for a level of transaction management that previously was more difficult to attain with the key-value nature of DynamoDB. As we evaluate this capability, we’ll be looking into how this announcement allows Kingland to broaden our usage of DynamoDB within our transaction-heavy, microservice-based data domains such as Legal Entity (i.e. company, organization data), Natural Person (i.e. people data), and Financial Instrument (i.e. stock, bonds, funds data).
  • On the topic of purpose-built persistent data stores, AWS’ announcement of the new Amazon Quantum Ledger Database and Amazon Timestream continues to expand the breadth of offerings within the AWS portfolio of persistent data store options. While we don’t actively invest in QLDB and Timestream capabilities as part of our Kingland Platform and client solution use cases, knowing that these capabilities are available expands the art of the possible for meeting the needs of new use cases, and the concept of purpose-built data stores continues to align very well with our Platform’s microservices and data domain strategy.
• The announcement of AWS Security Hub should excite the security teams and information security officers of AWS’ customers worldwide. At Kingland we’ve invested in a significant amount of tooling and capabilities for managing and monitoring all things security with our Platform and client solutions, including compliance with a variety of regular security audits. Security Hub promises to simplify aspects of this management, monitoring and compliance, and we’ll be looking forward to the capabilities it provides through its preview. Bringing these capabilities to the forefront of a significant AWS announcement is a great move by Amazon, in light of the public challenges in security exposed within other top global tech and consumer companies over the past couple of years.

While Textract, Lake Formation, DynamoDB and Security Hub announcements represented the most significant announcements as aligned with Kingland’s Platform investment and client solution development work, we continually keep a close eye on the additional AI/ML capabilities enhanced and released by AWS. With the release of DeepLens in 2017, the continued enhancements to SageMaker, and the release of DeepRacer in 2018, AWS strengthens their focus on "builders" (i.e. developers), easing the barrier of entry for learning and leveraging AI/ML capabilities. The DeepRacer in particular was such a hit that my colleague Patrick Rice and I waited over two hours in line for an opportunity to learn more about reinforcement learning as related to autonomous vehicles, including earning a free DeepRacer car in advance of its public release, at the completion of the workshop: 

At Kingland we continue to invest in our Kingland Platform and client solutions in service of our clients’ use cases, and AWS continues to be our strongest partner for cloud-optimizing our Platform and solutions. However, we’ve also heard from our clients that the ever-expanding market share of AWS in cloud is bringing about the desire for more cloud diversification among some of our clients, particularly those within the retail industry. To that end we have additionally worked with Google Cloud Platform (GCP) to accommodate better cloud diversification. Clients can count on Kingland to leverage the best of either AWS or GCP for solving some of the toughest problems and use cases in Text Analytics and Enterprise Data Management!