The Cloud Wars: making sense of the Cisco/Google partnership

Cisco and Google announced a new hybrid cloud partnership 2 days ago, and I wanted to know why. As I dived deeper, the strategic goal became visible pretty quickly. Let’s briefly examine the evolution of application architectures, review the investments in companies/technologies that Cisco and Google have made in recent years, and understand how synergy will be realized through this partnership:

Application development and delivery trends

Apps that are built today have very different scaling and management needs compared to the apps built 5 years ago because of the massive surge in users. Building a massive system like Netflix, or Spotify needs a completely new way of building, delivering and scaling applications.

Application development ecosystem evolution

If an app is a car, think of a microservice as a navigation system, or the steering system, or the internal temperature control mechanism. A microservice is a small functional system that coordinates with other systems to deliver the end product. Typically a team of 3–5 engineers owns all aspects of a microservice. The explosion of container technology (led by Docker) certainly accelerated the adoption of microservices. Containers let you get more value out of your hardware investments, as they can be created and destroyed quickly to accomplish critical tasks related to your application, in both test and production environments.

The containers were amazing but in production environments, scaling became difficult in the face of spikes in end-user traffic. Google built and open-sourced a tool call Kubernetes which solves that problem better than any other alternative (this is critical and we will talk more about it soon).

In addition, large enterprises are comfortable today with the scaling capability and the security of public and private clouds. The way business critical applications are delivered is changing and instead of just happily living inside the isolated datacenter, these apps now live on all sorts of combination of the cloud to accomplish conflicting goals of achieving scale, and saving money at the same time.

Now, let’s take a look at what Google and Cisco have been up to recently

Cisco

As mentioned in it’s 10-K, Cisco’s strategy is to help customers in their digital transformation. Cisco is also slowly moving away from its old perpetual hardware licensing model to a recurring revenue model (SaaS) from software sales.

This strategy is driven by its acquisitions. The company has lot of cash on hand (on July 24 2016, over $65 billion in cash and cash equivalents) and this cash has been put to use in acquiring some great companies to enhance Cisco’s cloud portfolio:

(Build Private Clouds) September 2014 — Acquired MetaCloud to help global organizations operate private clouds

(Integrate containers across data centers) August 2016 — Acquired ContainerX

(Monitor performance of applications) January 2017 — disrupted what was going to be the first tech IPO of 2017 by acquiring AppDynamics days before its IPO

(Machine Learning analysis of streaming performance data) October 2017 — Acquired Perspica to enhance the product capabilities of AppDynamics

The last 2 acquisitions listed above are interesting even if considered in isolation because by bringing data from these products together with the monitoring data from its routers/switches, Cisco can become a significant competitor in the Application Performance Management Industry (that’s a deep dive for another day).

In October 2016, Cisco end-of-lifed Cisco Intercloud, its $1B public cloud offering. While it must have hurt, it did allow Cisco to pursue a coherent strategy designed around the hybrid cloud. As seen in the graph below, the revenue from Cisco’s Data Center product category has shown significant growth over last 5 years and one of the success tracking metrics for this partnership with Google would be — how far this trendline moves to the top right in coming years.

Google

Google cloud was late to the cloud party and by the time it entered, Amazon Web Services was steering a mature product backed by a big community of cloud adopters that had built processes around the tools offered by Amazon. Product differentiation has been hard in this market.

Cloud

Back in the day, the Cola wars were watched closely and there is similar intense war in the tech world for cloud leadership. The Q3 2017 report from Synergy Research Group shows that AWS owns about 35% of the market. Google at 4% and Microsoft Azure at 8% have a long way to go to become leaders in the Cloud Provider market.

Now if you are making decisions at Google, one of the big questions is how do you go against the 800-pound gorilla in the market (AWS) that in addition to growing, because of the network effect benefits from its large community of users, is also continuously innovating?

In this competitive marketplace, you can’t just compete on product capabilities but what you can do is place bets on larger technology changes of the future to make the competition irrelevant. Google’s investment in a technology like Kubernetes and it’s partnership with Cisco seems to be one such major bet.

Strategy

Imagine Laura as the CIO of a large corporation who is responsible for applications that directly drive revenue. If the app is down or slow, she can say hello to customer churn. Laura wants full control over application infrastructure, in addition to 99.99% + uptime and security for the critical applications.

When adoption of the cloud became mainstream a couple of years ago, Laura finally became comfortable with the idea of hosting some of her applications in the cloud instead of the datacenter.

She found herself in a beautiful world that was built by Amazon — a large set of tools to deploy, scale and monitor applications in production. And this cloud was cheaper than operating a datacenter.

Now try selling a new cloud offering to this CIO. Here are some concerns:

1. It is costly in terms of developer hours to change processes to host the applications on this new cloud and this initiative increases the risk of missing the next major product release deadline.
2. A downtime of even 30 minutes during transition can adversely impact users which is any CIO’s nightmare.
3. Trust has not been established yet with the new cloud vendor and the value proposition is not clear.

Now if you are Google, what you offer is cutting-edge technology backed by excellent customer support but Laura has significant resistance to change as the switching cost is too high. What are some of the ways in which you can lower this resistance to change?

1. Eliminate the switching cost for Laura by creating partnerships with vendors who have already signed a contract with Laura.
2. Ride the next big technology wave and use it to acquire new customers, so Laura’s friends will convince her to use this new cloud.

This is where the opportunity lies to take the lead. If the CIO has already decided to adopt containers and the company is already investing time in building applications using Kubernetes, the selling pitch becomes more impactful -

“Dear CIO,

You are already using Kubernetes, and we can eliminate the scaling pain for you. We have a partnership with Cisco whose networking products are already being utilized in your data centers. With a push button install, we can help you leverage the benefits of both public and private cloud for your application.

Thanks”

This pitch works really well for a fortune 500 company that still hasn’t adopted the cloud for security and privacy reasons — read banks, financial institutions. So new customer acquisition now becomes easier. This pitch also works for AWS customers who already rely heavily on Cisco for its private/hybrid cloud deployments.

The partnership brings together the networking and security prowess of Cisco and the software expertise of Google together to offer hybrid cloud solution. The net result is more than the sum of the individual entities in this case — synergy will be achieved.

Should Amazon worry?

The competitive move may do little to steal its current customer base but the new customer acquisition strategy presents a challenge if Kubernetes ends up becoming the de-facto standard for the apps being built using containers. In that case, even though the software is open-sourced, Google will have leverage.

Google built and open sourced Android but by acquiring phone manufacturers like Motorola and HTC, it is keeping the relationship with Samsung in check — because it might choose to prioritize it’s own hardware when shipping out software updates.

Is the same relationship dynamic possible with Kubernetes if that ends up becoming the standard — yes, why not. What if the apps leveraging Kubernetes on the hybrid cloud offered by Cisco+Google end up getting security updates, customer support and patches before the ones running on AWS?

What’s in it for Cisco?

Adoption of public cloud is a threat to Cisco’s hardware business as the customers who adopt the cloud end up buying less hardware for their data centers. By offering products as a result of the partnership with with Google and Microsoft, Cisco is bringing the cloud to it’s customers, thus increasing the cost for the customer to switch to a new vendor.

Why not Microsoft Azure + Cisco?

Turns out there already exists an alliance , so Cisco is indeed working on partnerships to develop capabilities on the software side.

What might go wrong with the partnership?

A large potential customer might be perfectly happy working with 2 different vendors for it’s needs and not see value in managing a relationship with just 1 vendor. It is also possible that a new product will see greater adoption than Kubernetes, which may diminish the synergy of this relationship.

In addition to focusing on their core businesses, Google and Cisco will have to manage the new relationship dynamic with each other, with hardware suppliers, vendors, and customers. This overhead might not be insignificant.

Amazon’s options

A few come to mind:

1. Wait and watch — doing nothing at this point could be a perfectly valid option. Amazon could choose to focus on its current roadmap and make changes based on market feedback.
2. Sign a similar deal with Cisco — Amazon has a formal relationship with Cisco today and unless exclusivity clause has been added to the Cisco-Google contract, there is nothing that should stop both tech companies from negotiating a new contract.
3. Make Kubernetes irrelevant — Amazon could acquire a Kubernetes competitor and invest its resources in driving up the adoption of this new software as the default product to manage containers.
4. Partner with a Cisco data center competitor

What next?

Amazon is working to add more security to it’s offerings and earlier this year quietly acquired (?) harvest.ai . More investments/partnerships in coming months on the security side for Amazon might weaken the Cisco-Google pitch.

The Cisco-Google collaboration product will become generally available in mid-2018. If those early adopters end up being financial institutions who adopt the solution, the cloud market share numbers in a few years are likely to look significantly different.