Hybrid Platforms Trends Series: Supercloud - A Journey to Supercloud  

Welcome back to part 3 of the Supercloud trends series. In Part 1, we looked at what a Supercloud is, and in Part 2, we ensured those who should not be on the journey didn’t make a grave mistake.  

Hopefully, those still here are either looking to start the journey or are looking to steal a couple of ideas that could help with a hybrid cloud deployment. There are a lot of similar concepts between Supercloud and hybrid cloud, the main difference being the level of complexity and effort needed to see results.  

A journey 

First, I want everyone to understand that it will be a journey, and how far to travel the road is critical. Defining the end goal is essential and ensures direction and clarity; it does not mean everyone must fully adopt Supercloud. It might not be in your best interest to fully abstract the underlying cloud providers.  

Ensure you leave options and make quantified decisions against a clear set of requirements. Define the trigger points for review in your strategy, and don’t be afraid to say you have reached the optimal architecture before you get to the destination. If you take an iterative approach, you can evaluate the value of each step and ensure you manage the diminishing returns that any concept like this will pose. 

Core pillars 

As you embark on this journey, I have three core pillars that anchor your decision-making processes and help focus your teams when you define that vision and stepping stones. Without these foundations, it won't take long until the wheels fall off and everyone starts wondering why you even started! 

Automation 

Automation will be one of the critical capabilities needed to make your Supercloud operational, and everything will need to be built using Infra as code principle. Ensuring that your internal maturity in this space is high is critical. Consider this at two levels: one about the platform and the other about the developers.  

• Platform - Automation for deployment and day two operations of the underlying cloud services. 
• DevOps - Automation for elements like containers, databases, and, of course, your CI/CD processes.

Visibility  

Visibility will be critical; without a clear view of what is happening operationally and commercially, you won’t have the data points to benchmark success. This will lead to either stopping too early or pushing on too far, both outcomes that will reduce the return on investment. Consider accelerating the following concepts in your current estates before starting the Supercloud journey. 

• FinOps - Look at the definition for FinOps maturity and ask if you are at the ‘Run’ level.  

• Single view - Do you have the tooling to provide that critical single view of all platforms, applications, and data?  

• Unified operations - Can one consolidated operations team service your existing platforms and applications via a single operating policy?  

Governance, risk, compliance (GRC) 

Accelerating adoption to the speed needed to realise a return will require more than aligning the technical aspects. You must bring your people managing risk on the journey, requiring a clear view of requirements. Remember part 2, when we talked about data protection and compliance? Spend time defining your policy, understanding requirements, and getting agreement from all the associated teams, such as legal. 

If you're missing some of these basics, you will gain a better return on investment by maturing in the above pillars before looking at Supercloud. Consider the technology that can span across clouds with consistent operations, as these tools will simplify everything in the long run! 

Clarity of direction  

Leadership and people, from experience, will be the two most significant factors in success. Setting a clear vision and direction and then communicating that throughout the organisation will be critical. This will require a strong leader who can steer the technical direction, manage a broad scope of stakeholders, and calm the ship when inevitable bumps occur. Consideration should also be made on how the project will survive changes in leadership and people along the way; this is why documenting this vision, the journey, and expected milestones will provide everyone with a ‘North Star’ to align with throughout the project. It will also allow new people to align quickly and minimise changes in course mid-journey.  

We must remember the people executing the specific activities, as building a solid team will be critical. With so many tasks to complete and a complex web of interdependencies, it cannot rely on one or two people to understand and execute the vision. Build pods that can perform at pace but work in partnership with others to ensure parallel activities aligned to the broader outcomes.  

Advanced network architecture 

The network will be one of the most complex elements to get correct and has the possibility of some of the highest hidden costs. Most network admins will admit that cloud networking can be challenging, not always supporting the enterprise features that have become accepted norms in the data centre and at the edge. Couple this with differing capabilities per provider, and it will take some strong network leadership to navigate to the end architecture. Please don’t underestimate the time that should be spent at the design stage of the networking; it’s the foundation of the entire architecture. This topic could easily take up several articles alone, so I want to throw three concepts out to get you thinking. 

Network design and topology: 

Implement a hub-and-spoke network design where the central hub connects to multiple clouds (spokes), which can help manage data traffic more efficiently and securely. Considering the pros/cons of direct connections (AWS Direct Connect, Azure ExpressRoute, and Google Cloud Interconnect) versus open internet; cost and latency will be essential. Then, an overlay network using SD-WAN technologies will be created to connect the different cloud environments. This setup facilitates connectivity across diverse infrastructures and provides a programmable layer into which your automation engines can interface. 

Cost management: 

Monitor network costs closely as data transfer rates vary significantly between cloud providers. This area could cause bill shock if not managed in detail. Look to optimise network traffic flows to minimise costs related to egress. This will likely need to be done with the application developers and data architects, who will ultimately control the connections. 

Security and segmentation: 

Leverage micro-segmentation and virtualised networking, allowing you to divide cloud networks into smaller distinct zones to maintain separate security policies by workload. The critical element here is the portability of those networks, which allows easier migration of services when required. Also, ensure that data is encrypted in transit between different clouds and your on-premises environments.  

Service levels: 

Clearly understand the SLAs provided by each cloud, specific to network uptime, latency, and support. Ensure these SLAs align with your application requirements.  

Multi-cloud data management 

Managing data across this Supercloud will require a strong connection between technology and the business; defining a solid data management policy will be critical. You will no longer be able to keep everything forever. 

The organisation must agree on crucial topics like data lifecycle, retention, and compliance, and outside of the governance and policy concerns, we need to be aware of the diverse technical capabilities of each provider. Storage is provisioned and managed, and costs vary very differently in each cloud, meaning we need to consider an abstraction layer. This is one area where adopting technology that can offer consistent cross-cloud operations would be critical.  

When we look at the critical features of modern data storage (data reduction, cyber defence, cloning, replication, etc.), we sometimes find the native cloud features lacking. Your application architectures will also impact this. The requirements differ depending on virtual machine-based, non-persistent containers or containers that need persistent storage. 

Consider storage abstraction solutions from vendors with a proven heritage in storage (Dell, NetApp, Pure as examples). These providers can add significant value by providing a storage abstraction layer, significantly reducing costs, adding key features, and enabling consistent operations. This leads to my three focus areas for Supercloud data management: 

Data integration and mobility 

Ensure data can be easily moved between cloud environments without compatibility issues. This involves understanding the data formats, APIs, and services different cloud providers offer and looking at storage abstraction concepts. Leverage storage APIs that can integrate seamlessly across different clouds to simplify the management of data. 

Data governance and compliance 

Ensure that cloud data storage solutions comply with relevant data protection regulations (such as GDPR, HIPAA, etc.). This includes considerations for data residency and sovereignty, and make sure you implement tools and processes that allow for regular data audits to ensure compliance and governance standards are continuously adhered to. Think about how you will provide a consistent set of restore points, retention policies and the ability to comply with subject access requests. 

Data Redundancy and disaster recovery 

Implement robust data backup and recovery solutions that cover all cloud platforms used (again, another reason for an abstraction layer to provide consistent operations. This involves strategies for data replication across different geographic locations to ensure business continuity or building applications to have native cross-cloud resilience. It is also crucial to design fault-tolerant data storage architectures to support systems that need persistent data. 

Please don’t underestimate the complexity of the storage layer in your Supercloud architecture; cloud storage is not cheap at scale and could quickly spiral if not controlled well.  

The Control Plane 

Developing a consistent Control Plane (CP) will be the core of your Supercloud architecture, traditionally known as a Cloud Management Platform (CMP). I thought we needed more abbreviations in our industry, so I added Multi- to the front to give us a Multi-Cloud Management Platform (MCMP). Your MCMP should provide the downstream orchestration of the platforms, the upstream integration to service management capabilities and the visibility to manage governance and compliance. As the central point of control, it will need several key capabilities: 

• Automation & orchestration 

• Repeatable workflows 

• Self-service capabilities  

• Service catalog 

• Provisioning  

• Analytics 

• Configuration management 

• Integration and extensibility 

• Lifecycle management 

• Governance & compliance 

Building your requirements into this tooling is where most of the engineering effort will occur, so it must be a carefully considered decision. You want to avoid getting part way into your journey and need to change MCMP solutions! Again, this a topic that could be discussed for days, but three things worth putting at the top of your considerations are: 

Unified visibility and control 

Your MCMP must provide centralised management, enabling visibility, and control over resources across all cloud environments. This includes monitoring, managing, and optimising workloads, costs, security, and compliance. The platform should provide a centralised dashboard that offers comprehensive visibility into all resources, including infrastructure, applications, and data. This is critical in monitoring performance, usage, and costs from a single interface. 

Security and compliance management 

Robust security features are critical, including unified identity and access management, compliance monitoring, and data protection capabilities. The platform should enforce security policies consistently across all environments and help maintain compliance with various regulatory standards. 

Unified multi-cloud management 

Ensure your solution provides a single platform to manage multiple clouds, including AWS, Azure, Google Cloud Platform, VMware, OpenStack, etc. Remember that while we are talking about Supercloud, it is pretty likely you will end up with a hybrid Supercloud (sorry, another set of words before the word cloud) 

Summary 

That's the end of this Hybrid Platforms Trends series on Supercloud. Hopefully, it will give you a view of what it is and what it is not, and clarify when to consider embarking on the journey. I will repeat the advice from part 2: unless you have the scale or defined need for Supercloud, stick with one public cloud and leverage hybrid cloud architectures to support workloads optimally.  

If you are going on the journey, make sure you have a strong vision and an excellent team that can stay the course. Addressing the challenges we have discussed requires a strategic approach, including adopting integrated tools, implementing comprehensive governance policies, and ensuring staff have the requisite skills across the platforms chosen. Of course, working with expert partners like CDW can help navigate the complexities of a Supercloud environment!