GCP Anthos and BigQuery Omni
A Showcase of Disrupting Technology in Cloud
A Bit of History of the OS War in Personal Computing
Thirty years ago, in the early ’90s, there were multiple technology wars among different software vendors to determine who would the king of the personal computing world. First, the war between PC DOS trade-marked by Goliad IBM and MS-DOS introduced by David Microsoft in its early days. Then the war between Windows of Microsoft and Macintosh (MAC) of Apple.
In these wars, each vendor spared no methods or weapons in the battles to secure the victory as retold by Andrew Orlowski in OS/2 a quarter century on: Why IBM lost out and how Microsoft won. The victory of Microsoft over IBM is enlighted even more with the details of TRUE STORIES: The Agreement that Catapulted Microsoft over IBM by Sean Braswell. Then the war between Windows and MAC, not less dramatic, and again, Microsoft of Bill gates emerged triumphantly.
Let’s consider one of the main reasons of Microsoft victories in both well-known OS wars:
… Firstly, Microsoft ruthlessly went after the market and didn’t care about being liked, or being right — just about driving lots of copies of Windows into the market. The market didn’t like Microsoft, but Microsoft made them money. Software vendors, hardware vendors, peripheral makers — they all made more money with Windows. The software insulted you daily, but the “ecosystem” looked a lot healthier. IBM had no idea of what it took to build a new ecosystem. The suits thought that if they’d built it, and it was good (it was, eventually), then the customers would come — because of the magic letters IBM. The IBM sales teams had acquired many dark skills, including leveraging other IBM products and services. These were no use in a knife fight, on an open plain, with Bill Gates. … (Andrew Orlowski, 2012)
And
… Hundreds of thousands of IBM computers were sold with MS-DOS, but more than that, Microsoft became the maker of the crucial connection that was needed between the software and hardware used to operate computers. Company revenue skyrocketed from $16 million in 1981 to $140 million in 1985 as other computer-makers like Tandy and Commodore also chose to partner with them … (Sean Braswell, 2019)
It is clear that Microsoft won both wars proudly because the company served two groups of personal computing users — software application developers and end-users — very well.
Hardware systems, including storage components such as floppy disks or hard drives and networking items like NIC cards, can be manufactured by a great number of vendors, the more, the better for competition. However, no software application developer wants to spend money and time on moving from one vendor to another one. Customers have even less appetite to do anything with different formats and configurations created by different hardware vendors. All — software application developers and end-users want to move from one vendor to another freely and smoothly.
Current Cloud Computing: A Mirror of ‘70s and ‘80s Personal Computing
In the early days of the personal computing, ’70s and ’80s, end-users could select and purchase products from several different vendors like Radio Sheck TRS-80, APPLE II, or Commodore PET, to name a few. The biggest problems with these mini-computers were that users had to deal with all kinds of differences among them, from the UI (user interface), features, hardware components, to setting up, configuring, and using.
Do we find the same situation with cloud computing right now? The answer seems to be a big YES!
Each personal computer requires three critical parts: computing power (CPU), storage (RAM, floppy disks or hard drives), and networking ( NIC cards). Similarly, each public cloud service provider at present offers a cloud computing infrastructure of three crucial sub-systems: Computing resources, storage, and networking.
Let’s look at the top three public cloud service providers: AWS (Amazon), Azure (Microsoft) and GCP (Google).
Let’s consider only one sub-system: Computing resources. It seems that the concepts are similar for these sub-systems: EC2, Virtual Machines, and Compute Engine. However, how they are actually set up, configured, deployed, and used in reality is a totally different story. To be professionally competent in setting up and managing computing resources offered by a particular vendor requires special training programs of a significant amount of time and effort.
If, in the early days of the personal computing, no software application developers and end-users wanted to deal with differences between various mini-computer vendors, now, none of the users likes to handle differences between the sub-systems offered by AWS, Azure, and GCP, either. However, they have to use these services because they have no choice.
Let’s consider the current public cloud service environment:
For the cloud application developers, maybe, the same developer has to create two different versions of the same application: One runs on AWS, and another runs on GCP because there is no guarantee that a cloud application created for AWS would run perfectly on GCP, also to take advantage of specific features of each provider. And more, these cloud applications may not run on an on-premise server in a private corporate data center at all. In the same manner, an end-user has to use different applications to run on Azure and IBM Cloud, although these two applications may try to accomplish the same tasks.
It is clear that the current public cloud environment was exactly the same as that existing in the early ’70s and ’80s when computers from different vendors required different developing and employing approaches.
Get Thuan L Nguyen’s stories in your inbox
Join Medium for free to get updates from this writer.
One interesting study case was the recent $10 billion Department of Defense (DOD) JEDI cloud computing contract. One of the key detail of the contract is that DOD aimed to offer it to ONLY ONE service provider, which caused much debate, protest, and even lawsuits from various tier-2 cloud providers like IBM and Oracle. Then, Amazon AWS sued DOD after the contract had been forwarded to Microsoft. Definitely, DOD recognized serious risks with such a large contract offered to ONLY ONE service provider. However, DOD never wants to spend time and money to deal with multiple UI’s, mismatched configurations, different hardware and software systems, and many other headaches handling incompatible services or systems from two or three service providers.
Anthos: Cloud “Operating System”
Let’s have a flashback to over 25 years ago when Microsoft’s Windows 95 had just been released and displayed on the shelves of retailers. From that moment, almost all non-MAC users could use the same UI, share the same features, of a common operating system to work with a personal computer, most likely, a desktop, then, no matter which computer vendor made it.
So, we need to ask this question: What should be the common layer for the cloud computing as Microsoft Windows operating system does for personal computing?
Here Comes Anthos
Google Cloud Platform (GCP) Anthos is a managed hybrid and multi-cloud platform. Based on Kubernetes and cloud-native technologies, Anthos can be used to run Kubernetes clusters not only in GCP but also in AWS, Azure, and on-premise VMWare vSphere, and potentially many more in the future. It is not very far to state that a cloud user can run Anthos-compatible applications “everywhere” in the cloud, no matter who is the service provider, no matter whether public or private cloud.
Thomas Kurian (GCP’s CEO) promised to meet customers’ needs where they are, and Anthos delivers that. Anthos literally served customers in their environment, no matter where it is, even in the competitors’ cloud
With Anthos, Google brushes a masterstroke in setting its foothold in enterprise cloud computing environments. GCP offers a cloud service specific for Kubernetes technology: Google Kubernetes Engine (GKE). By embedding Kubernetes as the heart and mind of Anthos, GCP can bring GKE to the multi-cloud environment, the strategy most used by enterprises while deploying cloud computing (IDC: Multicloud Management). The influence and demand of Kubernetes has skyrocketed for last several years (Kubernetes Demand). As Kubernetes has become the de-facto standard (since 2017) for running modern, cloud- native applications in the enterprise IT environment, Anthos enables GCP to offer other managed services that can run on top of Kubernetes to users of any other cloud service providers.
So, maybe, we now have an answer to the above question:
In other words, Anthos has been designed to be a “cloud operating system.” Its role in the cloud computing environment is exactly the same as that of the Microsoft Windows operating system in the personal computing paradigm.
BigQuery Omni: A Showcase of Anthos
First, let’s find out what is BigQuery
BigQuery is an enterprise data warehouse that solves this problem by enabling super-fast SQL queries using the processing power of Google’s infrastructure. Simply move your data into BigQuery and let us handle the hard work. You can control access to both the project and your data based on your business needs, such as giving others the ability to view or query your data. (Google Cloud Platform)
So, briefly, BigQuery is a data warehouse cloud service offered by GCP. It is very powerful and popular for enterprise customers. Its speed and performance are so good that many customers of GCP competitors transferred data from other cloud environments to GCP to run data analysis with BigQuery and then transferred the results back to their original storage.
It is clear that transferring data back and forth between GCP and some other cloud service provider environments is very costly: both AWS and Azure charge around $0.09/GB (07/2020) for transferring data outbound. Additionally, transferring data back and forth costs more regarding latency: Moving data into GCP, loading into BigQuery, running data analysis, getting results, transferring results back — every step costs time and money.
What is BigQuery Omni? And what can it offer?
BigQuery Omni aims to provide solutions to both the above problems — the cost of transferring data and the cost of latency. Essentially, it can access the data where it is stored, run the data analysis on the data, and save the results right there. In other words, with BigQuery Omni, GCP comes to where the customer stores the data and provides the service right there. It is similar to a food catering service: Customers do not have to go to a restaurant. The restaurant offers a service that can serve their customers anywhere they want. Customers do not have to spend time and money driving to the restaurant, no waiting time at the restaurant, only relaxing and enjoying the food served right at their home or any other selected place, by the same well-known chef.
One of the most powerful features of Anthos may be that it can take advantage of the raw resources (computing, storage, networking) offered by any cloud service providers, including its competitors, to serve its users. If AWS can offer some powerful features of EC2/S3/VPN, Anthos can rely on them to deliver very good performance for the user. If Azure can provide any wonderful feature of its VM/Blob Storage/Vnet, Anthos will get that and provide it to the user. In other words, with Anthos, cloud users always likely to get the best ratio cost/performance anywhere at any time they want.
As a result, Anthos enables GCP to enlarge its global cloud footprint multiple times immediately, similar to the global presence of Microsoft Windows OS: Anywhere a personal computer is located, very likely Microsoft Windows is found there. With BigQuery Omni, customers of any other cloud service provider can quickly “order” a BigQuery service from GCP for their data in the same cloud region, zone, and network where the data is stored. It is interesting to think that GCP has actually converted the competitor’s regions and zones into their own ones.
Anthos: A Platform as a Cloud “Operating System”
GCP introduces BigQuery Omni to the cloud computing world as a new service to support its multi-cloud strategy. BigQuery Omni is the first of multiple major GCP services that GCP will build on top of Anthos. Much more importantly, as a new service that can access a large pool of customers, no matter where the customer data is stored, BigQuery Omni is actually a showcase for Anthos. It will be a demo for cloud application developers to see why and how they could and should build cloud applications for Anthos platform.
In other words, GCP aims to make Anthos a cloud development platform, a “cloud operating system,” exactly the same as Microsoft did make Windows a dominant one in the personal computing environment.
Built on top of Anthos, any cloud application can potentially run in any cloud service environment, either public or private, in cloud or on-premise. The developer does not have to be concerned about the specific features of each particular provider, not be bothered to develop various versions for different environments, saving time and money immensely. They will only have to make it Anthos-compatible.
That’s the aim of Microsoft for its flagship product: Windows operating system, 30 years ago. Maybe, Alphabet (or Google) has the same target for Anthos now. Who knows?
