Today, at Cloud Next 2019 in San Francisco, Google’s annual industry conference, Google announced its Cloud Services Platform, Anthos, for managing hybrid clouds that span on-premise and cloud data centers, and across multi-cloud environments. It is a big deal. It uses Kubernetes to enable migration across environments, is hardware agnostic, supports Amazon AWS and Microsoft Azure, and is subscription-based with a starting list price of $10,000/month per 100 vCPU block.

There is a thought that Anthos is a shot across the bows of AWS and Azure – and certainly, an approach that abstracts functionality from underlying cloud architecture will impinge on the ‘data gravity’ customer retention approach being used by these vendors. But IBM is at risk with Anthos as well, as the positive reception of its recent Red Hat acquisition is rooted in the promise of a single-vendor approach to providing hybrid and multi-cloud management and orchestration capabilities.

Clearly, Anthos has been developed with large enterprises as the target segment; some enterprise accounts are already early beta customers. To ease the addition of cloud as a core infrastructure platform in these accounts (by simplifying migration across in-premise and cloud environments) Google introduced Anthos Migrate, a service which will auto-migrate VMs from on-premise or other clouds into containers in the Google Kubernetes Engine.

It’s important to note, though, that hybrid cloud management is not only a point of pain within enterprise customers – it is a challenge (and arguably, a more acute issue) within midmarket (100-999 employees) firms. Consider these stats from Techaisle study of 510 US midmarket firms:

Data for Europe and Asia/Pacific also very interesting current and planned adoption percentages for hybrid/multi-cloud.

The multi-cloud, hybrid-cloud journey began within midmarket firms much before it became fashionable within enterprises.

One of the key markers in midmarket IT’s embrace of cloud has been the evolution from deploying workloads on whichever platform best met current demands to a more strategic platform approach. This has been a progression. Early on, before embracing various cloud types, it did not make sense to invest in establishing an overarching strategy; today, it does not make sense to continue to operate multiple discrete infrastructure platforms individually, as the management overhead exceeds the effort required to develop and implement a consolidated strategy. As much as cloud pronouncements claim that hybrid-only is cloud’s ultimate destination, experience has shown that cloud users prefer to adopt a strategy that delivers consistent management of cloud (and on-premise) resources while enabling deployment of discrete workloads on the most appropriate platform. It seems clear that once all options are in play, midmarket firms find use cases for each of public, private, hybrid and on-premise infrastructure. 

With this broad deployment of capacity in many different locations, the most compelling strategic question becomes ‘how do we aggregate and harness these resources into a responsive, properly-provisioned, manageable whole?’ Each part of the question is important. Resources that are outside of a central location and software management policies – for example, as would typically be the case with public cloud infrastructure and SaaS – can only be connected into a responsive hybrid system through the deployment of orchestration software. If the systems are isolated in operational silos, they are not responsive, and in aggregate, they almost certainly are overprovisioned, as each discrete environment requires processing, storage and memory headroom. And management of these diffuse systems is very complex, with staff needing the ability to test and deploy software, apply patches, upgrade equipment, establish and maintain an effective security posture, etc., in different environments and most likely through use of different tools. IT staff requires tools that allow for ‘single pane of glass’ monitoring and consistent maintenance routines.

This is exactly where Google Anthos can step in along with its committed, initial tranche of thirty OEM partners.

But what is not clear from Google’s announcement is the partner play. GSIs are necessary and play a useful role in deploying solutions within the enterprise segment but the midmarket needs a completely different partnering strategy. If Google wants to extend Anthos to the under-served midmarket segment, it will need to match the new technologies with an effective partnering approach – and while AWS may lack a significant partnering capacity, both Microsoft and IBM have depth in this area that may counter Google’s technical advances.