Google vs Azure vs AWS Comparison. Part 1: Overview, Scale, Ethics of the Cloud

???????Introduction

It is not possible to judge per se which cloud solution will fit best to the needs of a specific customer situation. The author of this essay sustains a neutral position. All of them are very convincing, extremely scalable and will fit any technical challenge as the examples show. However, there are a few topics where the author could detect – let us call it “beauties” – that shine.

In a series of small articles, the author will elaborate first on a general overview for customers who want to find some orientation. In the next parts of the essay, more specific comparison of the three big cloud providers will be done.

2????????How Hyperscalers scale on the money-side

Hyperscalers are driving cloud transformation at enormous pace. The industry principals reached gigantic size whilst still growing at tremendous rates. Microsoft (Azure Cloud), Amazon (AWS) and Alphabet (Google Cloud) have piled up market capitalization of around $2 trillion at the times of writing this essay. The profits regularly yield more than $1bn every week - at least Microsoft and Alphabet do so (see Statista 2021 [I], [II], [III]). With that these enterprises rank among the most precious worldwide.

Figure 1: Hyperscaling - financial and technical facts; Sources see below

3????????How Hyperscalers scale on the technical side

These big three certainly drove different business areas long before their cloud computing engagements. There is one thing in common: to cope with the incredible growth on their platforms, they invented their own technologies. This was necessary, as such size was unknown before. Let me give you some examples how the companies use their own technologies to build up incredible scale for daily operations.

3.1????????Amazon

Amazon uses its own AWS technology to run the yearly events Prime Day, Black Friday, and Cyber Monday. The technical metrics are mind-blowing and become more hefty year over year. So, for instance the AWS own product DynamoDB is managing several trillion API requests per day without performance impact sustaining single digit milliseconds response-times. 2021 the load peaked at 89,2 million database requests in just one second. Furthermore, the AWS block storage solution fleet EBS handled 11.1 trillion requests transferring 614 Petabyte in one day. The Content Delivery Network Cloudfront served more than 600 bn http-requests per day [IV].

3.2????????Google

The estimations of how many Google searches there are done hourly vary. Let me quote here a figure of around 250 mn queries per hour, what is around 2 trillion searches per year (Skai 2019 [V]). Youtube serves more than 5 bn videos per day. The user base is estimated to be 1,3 bn (Prater 2021 [VI]). Services like Gmail and Google Drive are known to deliver short latency-services around the globe. Google states that products and technologies used to fuel that platforms are also the ones offered in GcP (Donchev 2021 [VII]).

3.3????????Microsoft

Microsoft has its origin in providing Windows as operations platforms for servers, PCs, and devices. Furthermore, the Office Suite is since decades one of the most often package for desktop applications. Over the time Microsoft’s products took an evolution towards Software-as-a-Service backed by huge infrastructure to clients worldwide. So, the products morphed into Microsoft 365 what users can purchase on subscription base. Microsoft announced that in the meantime more than 50 mn. subscribers worldwide use the platform. Also, Microsoft states, that the technology is backed on Azure (Ziberg 2021[VIII], Choffani 2016 [IX]).

3.4????????How the tech companies morphed to cloud providers

The start of scalable Cloud Computing was done with the launch of Amazon Elastic Compute Cloud (EC2) in 2006. It was accompanied by Simple Storage Service (S3). This was an Infrastructure as a Service offering. Google started from another corner, namely Platform as a Service with its App Engine in 2008. Another two years later, in 2010, Microsoft launched Azure (compare Srinivasan 2018 [XI]).

4????????Where are the specific beauties of the Clouds?

In the next sections the author summarizes some selected specifics of each cloud providers. For sure, the points made in this article do not claim to be exhaustive.

4.1????????The beauty of AWS

AWS is the clear market leader in the times of writing this essay. Focusing on IaaS at first, AWS gave a natural entry point very early. In the first phase of cloud computing that was exactly what customers wanted to start with. The comprehensive storage and compute extensions gave the possibility to get rid of the CAPEX race. So, over the years, AWS developed the most granular portfolio of all the providers. Especially there are many different options for hybrid connect. This lets customers configure the right workload balance between public and private cloud.

Nowadays, enterprises which have a big legacy, i.e., use various elderly platforms, will find a lot of technology enabling a smooth slide into the cloud. A very broad scope of managed services (PaaS, SaaS) help customers to morph in a step-by-step approach to “cloud native” companies. There are examples where the range of possible solutions stand out the scopes of the competitors. Take as examples the broad set of managed relational database services (RDS) variants, caching solutions (f.i. Redis Managed Service), or holistic solution packages like AWS Connect, what is the call center solution.

In the following figure you will get an overview over the Services of AWS. Well, the structure as such is prototypical for all the clouds. However, when you analyze further you will see that behind these, there are even more granular services.

Figure 2: Fine granular setup of AWS Services (Baron 2017 [XII], p45)

There is a point that some customers of AWS begin considering, especially in the retailing space. In the last years, Amazon expanded from one category to the next (Schmidt 2019 [XII]). Hence, Amazon is in many cases not only the supplier of IT but also the competitor when it comes to the web shop. That means, Amazon offers its own footage which stands side-by-side to the products offered by retail clients. Those ask: “Well, what recommendations Amazon will give to a client when it comes to a choice between the products? Will the Amazon products be pushed more? Will Amazon place my product to the customer in a fair manner?”.

4.2????????The beauty of GcP

Google’s strength lies in its cloud purity. The inventor spirit of the company shines in a lot of products which are setting the standard of today. Following the App Engine mentioned above, Google released Kubernetes as Open-Source product in 2014 (Google 2014 [XIII]). Microservices-architectures democratized from a niche concept for tech evangelists to normal developers’ tool: “Kubernetes is a Greek word meaning ‘captain’. Like the captain is responsible for the safe journey of the ship in the seas, Kubernetes is responsible for carrying and delivering those boxes (comment of the author: containers are meant) safely to locations where they can be used” [XIII]. Anthos enables Kubernetes clusters partly distributed between AWS and Google to work in a service architecture (Google 2021 [XIV]). With Istio you can distribute Kubernetes clusters across regions (Google 2021[XV]).

Figure 3: Stunning microservices architecture based on Istio Service Mesh [XV]

Another example is the invention of TensorFlow by Google Brain in 2015. Google released it to the Open-Source community. Today, this is a leading machine learning, deep learning platform. It is also used for statistical and predictive analytics workloads (see Vaughan 2018 [XVI]). The AI products and suits have further developed and have become very popular among developers and scientists. The tool suite is accompanied by Google’s specific hardware designs of Tensor Processing Units, TPUs. These machines optimized for the crunching of compute work which is necessary to build sophisticated AI models. The cost per processing was decreased significantly by this incredible scale (Google 2021 [XVI]).

The strength in Artificial Intelligence is a big asset for the future. The analyst company ARK Invest headed by star investor Cathie Wood expects huge growth in the next decade (compare Phan 2021 [XVII]). Essential to unfold the full potential, so Cathie Wood, is the decrease of the computing cost which is associated with AI to pursue better and better Deep Learning. During the time of writing, Google still might have the smallest market share for cloud computing among the big three. However, it can expect that the AI skills is a good foundation to grow fast.

Figure 4: Tensorflow Processing Units TPU (Google 2021 [XVI])

Legendary is the protest of 3.100 Google staff including senior engineers in a letter to Sundar Pichai, the company’s chief executive: “We believe that Google should not be in the business of war,” says the letter and requests to announce a policy that it will not “ever build warfare technology.” (Shane 2018 [XVII]). Out of it came Google’s A.I. principles which among others refrain from weaponized AI (Conger 2021 [XVIII]). Instead, Google is using its AI to optimize its datacenters with regards to decarbonization and says to have done the most. You need to consider that all the big three state to be CO2 neutral already for a long time. Well, the financial capabilities are as limitless as their technical scaling. So, they easily achieve the status by acquiring the pollution certificates whilst still running operations via plants that run on combustion.

Since 2017 the company states to have achieved 100 % renewable energy consumption across all its operations, including data center. Google it thus getting the greenest rating among the big three (Oberhaus 2019 [XIX]).

4.3????????The beauty of Azure

Microsoft Azure is said to be the cloud with the most industry specific compliance certifications. It has more than 50 which are specific to global regions and countries. This includes US, EU, Germany, Japan, UK, India, and China. 35 compliance offerings which are specific to the needs of industries like health, government, finance, education, manufacturing, and media (Morgan 2019 [XX]). This is an important feature when it comes to run SAP workloads. These systems are in most cases the ones holding financials, balance-sheets, assets etc. Hence, these kinds of services underly strict regulatory requirements.

The fit of Azure was very soon also optimized for SAP workloads. Microsoft was the first enterprise out of the big three to enable HANA Large Instances (HLI) which were fully certified by SAP. These are physical servers also known as “bare metal servers”. The machines soon could scale up to 24 TiB Memory size. This is an essential requirement for some big OLTP installation of SAP S/4 in the ERP-area. Note, that OLTP S/4 workload keep all the data in the memory while they cannot scale horizontally but only vertically. OLAP workload like SAP Business warehouse however can scale horizontal. ?Azure introduced an architecture with 5 horizontal nodes adding up to 120 TiB total memory size. This supports big warehouses with sharding (splitting to pieces) of databases.

The fleet of Virtual Machines was expanded to the M-Series, which also got certification from SAP. These machines can be provisioned in the typical cloud style when available in a region. The sizes go up to 11,4 TiB (M416_v1). This serves a lot of necessary use cases – both for OLAP and OLTP.

This impressive development was also driven by Microsoft who migrated themselves to S/4 already in February 2018 ([XX]).

Microsoft also offers a specialized certification “Azure for SAP Workloads” where all kinds of migration paths and architecture setups are elaborated. All the setups are deeply interwoven with SAP preferences. See the following figure also to show how you can combine your cloud move with S/4 transformation.

Figure 5: Options to migrate SAP on Azure (Morgan 2019 [XX], p. 30)

In the meantime, AWS also is up to than excellent level. As a result, most analysts see Microsoft and AWS as market leader in the segment of public cloud with quite a distance to competition ([XXI]). However, the author considers Microsoft still as thought leader in this area.

5????????Summary - Comparison of the specific beauties

As you can see – all the three big ones really can cope any imaginable workload. None of them is the dominant provider in all aspects. They are rather all very impressive. When customers go for their specific evaluation, they do a lot of sandboxing, business case comparison, technical comparison etc. In this essay, the author wanted to give some “soft beauties” which shine through the products. The goal is to give an overview of the state-of-the art. The author will not give up his neutral position. Anyway, he summarizes the findings of his research as shown in the following figure.

Figure 6: Soft evaluation of the clouds [M. Braun 2021]

In the next parts of this essay series, the author will elaborate more specifically on the services the providers supply in the various areas. Part 2 gives an overview of basic functionalities in the clouds whilst part 3 will focus on cloud native services. Part 4 elaborates on the right path to move from ECC 6.0 to S/4 on the one side and from private to public on the other side. The SAP strengthes of the providers will be analyzed in more depth.

About the author:

Dr. Marc Braun works as Vice President Service Delivery Management at T-Systems since 2009. He helps customers to migrate and run their SAP workloads on “their” right cloud. Most recently, combining some of the moves with the customers’ S/4 transformation. Variants of SAP-Workloads on Azure, AWS, HEC and more and more also Google combined with private hybrids are building a really multi cloud scenario. Marc achieved Professional Architect certifications of AWS, Google, and Azure as well as SAP certifications.

Marc earned his PhD in 1999 with a research on Component Business Architectures based on Microsoft versus SAP for small to medium sized enterprises (SME). In his thesis he elaborated rule-based structures (algorithms) to match the disperse requirements of SMEs with adequate software functions.

Keywords: #SAPonAzure #SAPonGCP #SAPonAWS #peoplemakeithappen

6????????Annotations

I.????????Statista 2021, Net income of Alphabet from 1st quarter 2014 to 3rd quarter 2021, https://www.statista.com/statistics/540115/alphabet-quarterly-net-income/ [2021.12.27]

II.??????Statista 2021, Microsoft's quarterly net income from fiscal year 2008 to 2022 https://www.statista.com/statistics/272747/net-profit-by-quarter-of-the-microsoft-corporation/ [2021.12.27]

III.?????Statista 2021, Gewinn bzw. Verlust von Amazon weltweit vom 1. Quartal 2009 bis zum 3. Quartal 2021, https://de.statista.com/statistik/daten/studie/204386/umfrage/gewinn-von-amazon-quartalszahlen/ [2021.12.27]

IV.?????Jeff Barr 2021, Prime Day 2021 – Two Chart-Topping Days, https://aws.amazon.com/de/blogs/aws/prime-day-2021-two-chart-topping-days/ [2021.12.27]

V.??????Skai 2019, How Many Google Searches Per Day Are There? | Useful Search Metrics for Marketers, https://skai.io/monday-morning-metrics-daily-searches-on-google-and-other-google-facts/ [2021.12.27]

VI.?????Meg Prater 2021, 25 Google Search Statistics to Bookmark ASAP, https://blog.hubspot.com/marketing/google-search-statistics [2021.12.27]

VII.???Google 2021, Why Google Cloud, https://cloud.google.com/why-google-cloud/ [2021.12.27]

VIII.??Danny Donchev, 2021, 37 Mind Blowing YouTube Facts, Figures and Statistics – 2021, https://fortunelords.com/youtube-statistics/#:~:text=Almost%205%20billion%20videos%20are%20watched%20on%20Youtube%20every%20single%20day. [2021.12.27]

IX.?????CHRISTIAN ZIBREG 2021, Microsoft 365 Now Boasts Over 50 Million Subscribers, https://www.makeuseof.com/microsoft-365-passes-50-million-subscribers/ [2021.12.27]

X.??????Chouffani, Reda 2016, Why Microsoft Azure is the backbone of the Office 365 subscription, https://searchcontentmanagement.techtarget.com/tip/Why-Microsoft-Azure-is-the-backbone-of-the-Office-365-subscription [2021.12.27]

XI.?????Vitthal Srinivasan, Janani Ravi, Judy Rai, Google Platform for Architects, Birningham 2018.

XII.???Joe Baron and others 2017, Indianapolis, AWS Certified Solutions Architect – Official Study Guide Sarah Schmidt 2019, Amazon's Competitive Advantage and How Retailers Are Fighting Back https://blog.marketresearch.com/amazons-competitive-advantage-and-how-retailers-are-fighting-back [2021.12.27]

XIII.??Google 2014, What is Kubernetes, https://cloud.google.com/learn/what-is-kubernetes [2021.12.27]

XIV.??Google 2021, Anthos clusters on AWS architecture, https://cloud.google.com/anthos/clusters/docs/multi-cloud/aws/previous-generation/concepts/architecture [2021.12.27]

XV.???Google 2021, Building a multi-cluster service mesh on GKE using multi-primary control-plane multi-network architecture, https://cloud.google.com/architecture/building-a-multi-cluster-service-mesh-on-gke-using-replicated-control-plane-architecture [2021.12.27]

XVI.??Google 2021, Cloud TPU, https://cloud.google.com/tpu [2021.12.27]

XVII. Trung Phan 2021, Ark Invest believes deep learning will create $30T of value over the next 15-20 years, https://thehustle.co/02022021-deep-learning/ [2021.12.27]

XVIII. Jack Vaugahn 2018, TensorFlow, https://searchdatamanagement.techtarget.com/definition/TensorFlow [2021.12.27]

XIX.??Scott Shane and others 2018, ‘The Business of War’: Google Employees Protest Work for the Pentagon, https://www.nytimes.com/2018/04/04/technology/google-letter-ceo-pentagon-project.html [2021.12.27]

XX.???Cate Conger and other 2021, Google executives tell employees it can compete for Pentagon contracts without violating its principles, https://www.nytimes.com/2021/11/15/technology/google-ai-pentagon.html [2021.12.27]

XXI.??Daniel Obrhaus 2019, Amazon, Google, Microsoft: Here's Who Has the Greenest Cloud, https://www.wired.com/story/amazon-google-microsoft-green-clouds-and-hyperscale-data-centers/ [2021.12.27]

XXII. Nick Morgan and others, Birningham 2019, SAP on Azure Implementation Guide.

XXIII. See f.i. Provider Lens 2020, https://d1.awsstatic.com/partner-network/partner_marketing_web_team/solution-brief/SAP-HANA-Infrastructure-Services-AWS-Leader-Page.pdf [2021.12.27]