Background of Google Cloud Next '17
As you may have read in other blogs,
Google Cloud Next '17 spanned March 8 - 10 at the Moscone Center in San Francisco, California. The purpose of the conference is to bring together industry experts to share and discuss ideas for the future of the cloud particularly revolving around Google's premiere cloud products including Google Cloud Platform, G Suite, Maps, Devices, etc. Google Cloud Next '17 was a key event for SAP this year as
SAP and Google have entered a strategic partnership to make SAP Cloud Platform SAP HANA, express edition, and eventually machine learning capabilities available on Google Cloud Platform, as announced in the linked blog by Bernd Leukert.
I was fortunate enough to attend this event and have described my experience below.
Blog abstract
In hosting a tech conference, all technology companies face the same crucial obstacle: How do we show complex technical concepts and capabilities in a way that is consumable by all conference-goer personas? What Google achieved at Google Cloud Next '17 was to turn these concepts and capabilities into games hosted at stations that were scattered around the show floor. I address the game stations that I found the most creatively simple in the sections below.
What I suggest in this blog is that we, at SAP, adopt a similar framework for visualizing our technology. In this way, we can merge our external mission "Run Simple" with an internal mission (made up by me, on the fly) to "Show Simple" and make the information interactive and relatable to all.
Game station : Kubernetes Whack-a-pod
At this game station, the audience member is faced with a game console reminiscent of whack-a-mole. The audience member, now a player, is expected to "whack" objects as they rise from their hole.
Each time an object is whacked, it retreats back into its hole. Yet while the player was preoccupied whacking that object, five new objects have risen to take its place. A monitor above displays which objects have been whacked and which objects have regenerated.
After a few rounds, the player comes to realize that no matter how many objects they whack there is always at least one object standing to take its place. And therein lies the meaning behind the Kubernetes Whack-a-pod game station.
What this station visualizes is the high availability of the Google Cloud Platform and Kubernetes, Google's open source container cluster manager. Within the Kubernetes architecture lies a
Node Controller and its minion nodes. These minions are worker machines, either virtual or physical, that include services like a Docker container runtime,
kubelet and
kube-proxy to define and run pods. A
Kubernetes pod holds a group of one or more containers, the shared storage for those containers and options about how to run the containers. To ensure the availability of their application, the developer can create one or more
ReplicationControllers when configuring their Kubernetes cluster. When a pod fails, the ReplicationController kicks in using a
stored pod template to automatically replace it. When creating a ReplicationController, the developer can define the number of replica pods to be maintained. In the Kubernetes Whack-a-pod game, the controller ran 10 copies of the replica pods to ensure that when the player whacked one pod there was always another available to take its place.
Game station : Pi Day
At a second game station, the audience member stands in front of a display with a giant plastic π symbol on one end and a television monitor on the other. The player approaches a console with a dial and a button. The first prompt on the monitor requests the player to input the desired processing power of a Google Compute Engine (GCE) instance to be spun up. The second prompt requests the player to input the desired memory to be allocated to the instance.
The Google Cloud Platform then takes over and, while its processing, the game station attendant explains to the player that it's working to output the first 100,000 decimals of π. Within a minute, the monitor displays the value of π as well as the following :
- How long it took to request the GCE instance
- How long it took to spin up the requested instance
- How long it took the instance to compute the value of π
What this station visualizes is the end to end robustness and agility of Google Compute Engine instances, for example the ability to choose my own virtualized hardware settings through
custom machine types. A GCE instance is simply a virtual machine hosted on Google Cloud Platform. In under a minute, I as the user had been provisioned an operational virtual machine with the processing power to compute π out 100,000 decimal places. That was with 30 (Steph Curry's number) processing units and 98 (my street number) gigabytes of RAM. With the scaling capabilities of GCE's custom machine types I could have chosen as little as 27 GB or as much as 195 GB of memory for this virtual machine instance.
In conclusion
Google Cloud Next '17 did a great job of taking abstract technical topics and presenting them in a format that is relatable and memorable to anyone, in the form of games. While at game stations like Kubernetes Whack-a-pod and Pi Day I was, without even realizing it, discovering and interacting with capabilities of Google Cloud products like customized machine types and highly available infrastructure. From participating at this event, I have drawn the following algorithm:
Relatable + Memorable = Simple
We, at SAP, could take this framework and run with it. As a part of the Product Management team for SAP Cloud Platform, SAP's flagship business solution platform in the cloud, I imagine a TechEd booth where SAP Cloud Platform Internet of Things is a round of Mario Kart or SAP Cloud Platform Workflow is described using the "hot potato" party game. Visualizing business solutions at a tech conference is not an easy feat, but I see an opportunity to show them in a simple manner through games.
As for SAP Cloud Platform, I left this conference exhilarated thinking of the possibilities that customers will have with this kind of infrastructure underpinning our platform whether provided by Google, Amazon, Microsoft or any IaaS provider. I guess we will have to wait until
SAPPHIRE in May to see what comes next.
About the author
Colin Kraczkowsky is a Product Specialist within the Product Management team for SAP Cloud Platform. His domain expertise includes Cloud Platform extensions to SAP SuccessFactors, SAP's cloud HR solution. In his spare time, Colin enjoys learning new things, long hikes and Turkey Trots.