What Separates Cloud Leaders From Others

They are responsible for almost everything we see on the internet. They are the internet. A handful of cloud service providers, namely AWS, Azure and Google Cloud, rule the market. Their clientele includes billion-dollar streaming services, banks, federal institutions and many more.

Source: What Separates Cloud Leaders From Others

It has been more than a decade since Amazon dished out its cloud services, and it still tops the charts with Microsoft and Google at close second and third positions. Many cloud players have burst into the scene along the way but are yet to make a mark. What separates the top three from the rest is their constant revival of niche services. The big three forayed into ML-based services quite early. Now they have custom options to create chatbots, deploy AutoML, recommendation engines, and many other applications that power most companies.

A recent Gartner’s Magic Quadrant survey has named AWS, Azure, and GCP and IBM leaders in the cloud.

So what makes them unique? Let’s find out.

Nailing The Niche

---

(Note: CAIDS:  Cloud AI Developer Services; CIPS: Cloud Infrastructure and Platform Services)

According to Gartner, AWS services allow organisations to deploy a wide range of AI/ML capabilities into their applications. AWS offers language, vision and autoML services for developers. The company also benefits from having an extensive set of cloud infrastructure and platform services (CIPS) that enterprises broadly adopt. This makes it easy for existing AWS customers to add these services to their contract. More than any other vendor in this Magic Quadrant, AWS is very developer-centric; other providers target data science professionals, with developers being a secondary target.

Gartner attributes Google’s commitment to the model quality, accuracy and a diverse customer portfolio to be key contributors to its place at the top in the magic quadrant.  According to Gartner, Google underpins many SaaS offerings from other technology providers. The vendor actively engages open-source communities originated in Google — TensorFlow, BERT and Kubeflow, especially with the latest releases such as  Document AI, Visual Inspection, etc.

Microsoft’s Azure is closing in on AWS. The company has even managed to one-up AWS to bag a billion-dollar defence contract. With larger institutions putting more confidence in Azure and CEO Satya Nadella’s aggressive move towards the cloud, AI has started to pay off. Microsoft’s ubiquity, states Gartner, gives it an advantage in the enterprise market. Developers are used to working with the company’s tools. Microsoft embeds many of its AI capabilities into its commercial products with everyday AI features. It also contributes to developing standards and guidelines for best practices associated with AI, including AI’s responsible use.

How AWS Retains Top Spot 

“AWS is a Leader in this Magic Quadrant. Its CAIDS offering has a wide range of services and products.” 

Gartner

AWS offers a wide range of services, products and capabilities that allow developers to enhance the applications they are building. For instance, AWS AutoML services via Amazon SageMaker Autopilot makes it easy to build models for non-technical users. SageMaker can manage model workflows with a model registry, facilitating easy incorporation into a continuous integration/continuous delivery (CI/CD) pipeline for ModelOps. 

According to Gartner, AWS’s strong position in CIPS gives it an advantage when adding extra products and services for clients, as the relationship already exists. It is always easier for enterprises to add a product than to add a new vendor.

AWS AI Services

• Amazon Fraud Detector makes it easy to identify online fraud. Fraud detection models can be set up with just a few clicks and no prior ML experience.
• Amazon CodeGuru allows one to find and fix code issues such as resource leaks, potential concurrency race conditions, and wasted CPU cycles. 
•  Amazon Transcribe offers speech-to-text capability to their applications.
• Amazon Lex for transforming speech to text, and natural language understanding (NLU) to identify the intent of the text.
• Amazon Polly converts text into life-like speech, allowing users to develop new categories of speech-enabled products. 
• Amazon Textract  for quickly automating document workflows and processing millions of document pages in hours. 
• Amazon Comprehend uses machine learning to find insights and relationships in the text without the need for a machine learning expert.

How Azure Made It To The Top Tier

Product or service: Microsoft continues to enhance its cognitive services, ML and autoML portfolio, which applies to multiple personas: data scientist, developer and citizen data scientist. The Azure Machine Learning designer offers a visual flow design tool to connect datasets and modules to create, test, train and deploy pipelines and ML models. Cognitive services provide stand-alone and customisable models for speech, language and vision. 

Market understanding: Microsoft is among the more flexible providers of CAIDS in terms of deployment options. Its services are deployable in the Azure cloud, a virtual private cloud or on-premises via containers, based on enterprise customers’ needs. Microsoft also integrates services with extended systems and devices with its intelligent edge options.

Azure AI Services:

• Azure Databricks offers the most advanced machine learning capabilities to quickly build, train, and deploy machine learning models. Azure Machine Learning provides a Python-based machine learning service with automated machine learning and edge deployment capabilities.
• Azure Cognitive Search, formerly known as Azure Search, utilises the same integrated Microsoft natural language stack, which Bing and Office use. With Cognitive Services, users can spend more time innovating and less time maintaining a complex cloud search solution. It has recently added neural voice for text to speech in 49 languages and expanded its Azure Cognitive Search capabilities.

GCP Is Bullish On AI And Why It Works

“Google is a Leader in this Magic Quadrant. It offers services in all three areas evaluated — language, vision and autoML — and delivers them from its public cloud.”

Gartner

Though GCP showed significant losses in the last quarter, the company is well-positioned in AI research. Alphabet Inc. leverages the research done at DeepMind and Google Brain to power their services. “Google taps into its research technology, including Google Brain and DeepMind, to innovate across the AI stack. This has resulted in products like Cloud AutoML, Vizier, WaveNet, BERT and Federated Learning. Other innovation examples are feature store, edge inferencing, neural architecture search and purpose-built hardware for deep learning. Select customers directly engage with Google on applied research,” said Gartner in their report.

Gartner also cited GCP’s transparent and flexible pricing to reflect its interest in building a value-driven approach to building AI solutions and maximising developer productivity. Customers have a choice to pay for specialised models, capabilities, use of pre-trained APIs, and for compute and storage. There are also free tier options to ensure all customers can access Google services.

GCP’s AI services:

Users can use AI Platform Training and Prediction services to train models and deploy them to production on GCP in a serverless environment or do so on-premises utilising the training and prediction microservices provided by Kubeflow. Here are a few other services that made GCP one of the most preferred cloud providers:

• Deep Learning VM Image contains the most popular AI frameworks on a Google Compute Engine instance
• AI Platform Notebooks makes it easy to go from data ingestion to preprocessing and exploration, and eventually model training and deployment.
• BigQuery is a huge add-on to the platform. It allows users to store data and import the labelled data to AutoML and directly train a model.
• Original Postroducts/ai-machine-learning/getting-started-kubeflow-pipelines" target="_blank" rel="noreferrer noopener">Kubeflow Pipelines to build reusable end-to-end ML pipelines that one can share with other users and deploy on GCP or on-premises.

What Is Cloud Without AI

Companies don’t just want cloud service providers to store their data but generate intelligent insights on the go. AWS, Azure and GCP have dozens of services that do exactly that. While other players such as IBM, Alibaba, and Oracle offer decent services, they cannot break into the top bracket. Companies like Oracle, which were supposed to rule cloud markets, are trailing behind and facing an existential crisis when it comes to cloud. 

Developers and organisations are beginning to drift towards AutoML, federated leaning, responsibleAI and other subdomains that promise to tackle the previous shortcomings of algorithmic applications. Unless other cloud players change their AI tack, the big three will continue to have the run of the place.

Find the full report here.