AWS Cloud Day Melbourne 2023

On 02 Aug 2023 AWS held a Cloud Day conference in Melbourne. I was quite pleased with the event’s quality (speaking as someone who over the years, has attended a large number of conferences in the ICT industry). The pacing was wonderful, the contents carefully arranged, and the speakers were engaging. I made a few notes / observations on the sessions I went to.  


Generative AI

Generative AI was definitely a hot talking point. Besides heavy featuring in the Keynote, a number of speakers touched on it in sessions that focused on other topics – people could make connections from various angles. 

The common consensus on the phrase Generative AI, or GenAI, is that it describes the broad range of artificial intelligence that can create a wide variety of data – text, images, videos, audio, code, and 3D models. ChatGPT, Google Bart and Microsoft Bing AI are several well-known examples. 

Generative AI can leverage large amounts of unlabelled data to pre-train the machine learning models developed through artificial neural networks (which got its name from the fact that its modelling algorithms used the principles of neuronal organisation, which was discovered by connectionism in the biological neural networks constituting animal brains). The details of artificial neural networks definitely deserve separate discussions – its mathematical modelling through simplified neuron processing (as in a brain) is both fascinating and complex. Here we only need to know that it is machine learning mechanisms used by all sorts of artificial intelligence, covering several main learning paradigms such as supervised learning, unsupervised learning, reinforced learning and self learning. The ‘traditional’ AI / machine learning is often for solving very specific tasks. 

On the other hand, Generative AI can do a variety of generalised tasks. You can also fine tune generative AI by training it using additional domain-based, contextual data to achieve specific AI transformations, such as generating medical notes from doctor-patient conversations, or helping in cyber security attack responses. There is an infinite number of such use cases, or such modern inventions. Data is the genesis of these modern inventions. 


Data is the genesis

Building an end-to-end data strategy has become crucial to an organisation – not just for Generative AI, but also for analytics, machine learning, business intelligence and other aspects as well. 

AWS has a range of tools and services for building an end to end data strategy, from data store and query, to data integrate and to act. Some of these AWS services include:

Aurora DynamoDB Kinesis & MSK Redshift S3 Glue    

Redshift SageMaker EMR Quicksight

Amazon DataZone (in Preview) was explained in particular. It is a data management service you use when you plan to share, search, and discover data at scale across organisational boundaries (across AWS accounts and regions). It offers a unified data analytics portal with a personalised view of all your data. Data governance and compliance policies are enforced.

Amazon SageMaker JumpStart also received a special mentioning. Amazon SageMaker JumpStart helps organisations to quickly and easily get started with machine learning by providing pre-trained, open-source models for a wide range of problem types. JumpStart provides a set of solutions for the most common use cases that can be deployed readily with just a few clicks and supports one-click deployment and fine-tuning of popular open source models. These models can be incrementally trained and tuned before deployment. It also provides solution templates that set up infrastructure for common use cases. JumpStart can be accessed through its landing page in Amazon SageMaker Studio.


Application / business model modernisation

The following modernisation characteristics were discussed:

- Scale to large number of users

- Global availability

- Respond in milliseconds

- Petabytes of data handling

- Modular

- As Serverless as possible

- Automated and standardized

- Everyone’s responsibility

- Purpose built and decoupled


The following trends are in place:

- Move to cloud native

- Move to Containers and/or Serverless

- Move to managed databases

- Move to open source

- Move to modern analytics

- Move to modern DevOps

- Organise for value 

- People drive innovation, not the technology

- Automate, enable and self-service as much as possible with continuous improvement

- Offload the heavy lifting (to managed services / providers)

Using serverless and containers to enable modern applications and benefit modern business was discussed in one of the sessions. Digital resilience strengthening has been regarded in the last a few disruptive years as a key focus to future proof business, and improve profitability, innovation and cost efficiency. Application modernisation enables digital resilience. 

Functions-as-a-Service (FaaS, such as the Serverless services, like Lambda Functions in AWS), when compared to PaaS (Platform as a Service, including containers), can enable higher efficiency in:

  • Productivity increase
  • Time to market
  • Change vs. Run ratio

In turn, PaaS can offer higher efficiency on all these aspects over IaaS (Infrastructure as a Service, such as virtual machines). 

As to choice considerations on modernisation paths, the following guidelines were given:

- When the modernisation path is to build new application, the considerations should focus more in Serverless, Event Driven Architecture

- When Re-Platforming existing applications, the considerations can focus on moving to managed services, with Containerisation on the cloud a popular choice which may only require little overheard while offering significant gains

- When Re-Factoring existing applications, consider modernising architecture, software delivery and operational model. Containers and Serverless are good candidates. 

Serverless Caching, such as caching data in Lambda Functions was discussed. Together with more traditional caching placements: in the CDN, on the API Gateway and in the caching layer in front of databases – comprehensive caching can make day and night differences to workload performance. 


Many other topics were also discussed in the many sessions through out the day. The newly launched AWS Melbourne Region (ap-southeast-4) received a lot of attention. It actually enabled a bar celebration on the day! A number of AWS partners also provided exhibitions.        

                                                                                                                                    -- Simon Wang

Comments

Post a Comment

Popular posts from this blog

Fairness Evaluation and Model Explainability In AI

Image
Artificial Intelligence (AI) relies on machine learning and its modelling to provide outcomes. Imagine a credit card company receives hundreds of thousands of applications each year and would like to use AI to do the first round of application filtering. If it is not developed properly, the AI decisions will be skewed with unfairness – such as rejecting a lot more applications from a certain age group, from a certain gender, or from a certain employment history pattern – UNFAIRLY. Please note unfairness is the key issue here – in other words, AI and machine learning inference that is not reflecting the due course in actual situations. This is typically called bias in AI and machine learning. Biases can come from various stages of a machine learning life cycle: • Biases may exist in the pre-training data -  e.g., the dataset to be used in machine learning training has biases in it • Biases may be introduced by the machine learning exercise  Below is a machine learning lifecycle char
Read more

AWS and Generative AI

Image
 Generative Artificial Intelligence, or Generative AI, GenAI, is one of the hottest topics in the recent months. You would ‘have to be living under a rock to not notice this massive topic’.  ‘Generative AI is a branch of AI that focuses on creating new data. It is a subset of machine learning. The goal of generative AI is to create new data that is similar to the data that was used to train the model.’ – texts generated by an AI tool.  Recently I experimented a couple of random text-to-image GenAI tools to observe how generative AI can create contents and how the data generated by different GenAI tools trained by different models is different. My experiment was straightforward: text-to-image GenAI tools ask you to provide some texts as a prompt, and then use this input to generate an image. So that was what I did. The prompt I supplied was: ‘female student graduation photo in front of a white building with flowers in hand in Paris’. Both GenAI tools I tested could output an image withi
Read more

Amazon CloudFront and Its Primary and Secondary Origins

Nowadays most of the online contents are web service based. A user accesses a specific content, for example a web page, by connecting their client end application (like a web browser) running on their desktop or mobile computing device to the server that hosts the web page. There are billions of websites and webpages. DNS services, online search engines and links on well known ‘portal’ websites help people to find and locate the contents they would like to visit.  A user’s web client device and the server it visits can be worldly apart: for example, let’s think of a user in Tasmania, Australia that uses its smart phone to visit a website hosted on a server that locates in a data centre in Montreal, Canada, which is more than sixteen thousand kilometres away. They are connected by the Internet. The Internet comprises of millions of inter-connected networks. Traffic on the Internet is routed between the users and the servers through those inter-connected networks in-between. Using the ab
Read more