The Case for Hybrid and Multi-Cloud Computing

There is an ongoing buzz around cloud computing as organizations seek to speed their digital transformation efforts. From new enterprise digital business apps to new products and services for consumers, cloud-native development has sky-rocketed. Simultaneously, many enterprise legacy workloads are being migrated to the cloud. So, what justifies cloud-native development or enterprise workload migration? And what role does hybrid and multi-cloud play in a strong cloud adoption strategy? Today we share some key considerations and how they can influence your strategic approach to cloud. 

Figure 1: Hybrid Multi-Cloud Model, Illustrating workloads across On-Premises, Private and Public Cloud

Hybrid cloud

While there are many definitions of hybrid cloud, for purposes of our article, we define hybrid cloud as an operating model where an organization leverages both its on-premises, private cloud, and public cloud infrastructure to run its workloads. In this model, certain workloads can be orchestrated across private clouds hosted on on-premises and public cloud. Meanwhile, other workloads may be suitable for either on-premises or the public cloud. 

While many workloads can be migrated to the cloud, some may be better suited to run on-premises both in terms of cost and functionality. We recommend organizations weigh the following questions as they consider their need for a hybrid cloud strategy:

• Do you have a need (regulatory, legally or otherwise) to exercise full control over a data center and its physical security?
• Do you have latency-sensitive workloads that should be kept close to operations?
• Do you have workloads that may not benefit from the elasticity and scalability of the cloud?
• Are your licenses not cloud friendly or portable to the cloud model?
• Have you already invested in on-premises infrastructure such that a cloud migration will provide unsatisfactory ROI?  
• Do you have stable, computationally intensive, and long-term workloads which are better suited for on-premises infrastructure and would be cost prohibitive on a pay-as-you-go model?
• Do you possess legacy technology that is incompatible with the cloud, but provides good business value and technical quality?

If you answered yes to any of these questions, a hybrid cloud strategy may be a fit for your needs. Yet, you may also have workloads that present a compelling business case for migration. Let’s take a look at the common factors we recommend you assess for here:

• Do you have digital business applications with exponential data and user growth that are better suited for the elasticity and scalability of cloud?
• Do you have workloads that leverage specialty services like object storage, cold storage, analytics, machine learning, and other feature-rich cloud-native services?
• Do you rely on regions and edge locations to set up globally available infrastructure?
• Do you have elastic and on-demand needs like cloning test environments for projects?
• Do have a need to scale and expand into cloud for burst capacity?
• Does your enterprise need disaster recovery and business continuity for its on-premises applications?
  
• Do you aspire to re-engineer legacy applications into cloud-native micro services to bring in scalability and speed-to-market?
• Do you want to leverage the best of on-premises based software and cloud based managed services to orchestrate business applications across on-premises and public cloud? 

If you answered yes to any of these questions as well, you could see why so many organizations are taking a hybrid-cloud model approach. Using both on-premises infrastructure and the public cloud on a case-by-case basis, this approach allows them to optimize for IT objectives and business goals. 

Multi-cloud 

Next let’s take a brief look at a multi-cloud strategy. For this article, we define multi-cloud as using multiple private and public cloud providers to run IT workloads. This may involve a combination of private on-premises cloud and other public cloud providers. Common questions to determine if a multi-cloud strategy might be a fit for you, include:

• Could you better mitigate capacity risk by spreading workloads across multiple clouds?
  
• Is your existing suite of COTS products offered in a SaaS model by a cloud provider?
  
• Could you benefit from moving workloads based on geographical proximity and latency?
  
• Do you want to leverage multi-cloud infrastructure to build disaster recovery capabilities? While availability zones and regions provide levels of redundancy, running applications or replicating data across multiple cloud providers provides the next level of redundancy.
  
• Do you have specific business use cases and requirements that would benefit from different public cloud vendors and their products?
  
• Could you benefit from the ability to exploit pricing, quality of service, and support differentials?
• Would your teams benefit from the autonomy to choose what’s best for their specific requirements?

If you answered yes to one or more of these questions, read on as we dive a little further into seven key cloud considerations and how they may impact how your enterprise approaches the cloud journey. 

Figure 2: Considerations for Cloud Adoption

Automatic and on-demand scalability of compute and storage

It’s critical that IT infrastructure should not bottleneck business growth. As the business grows, workloads steadily increase and create natural cycles with peaks and valleys in the load. The need to automatically scale infrastructure to meet these business cycles is critical. Cloud services provide the ability to incrementally scale storage and compute, scaling services automatically based on metrics such as CPU and memory utilization, thereby ensuring business needs are met.

On demand and automated provisioning of infrastructure

Cloud provides unlimited on-demand and automated provisioning of infrastructure. Because infrastructure is provisioned through code, all resources can be provisioned and integrated through code. This use of automation grows system productivity, repeatability and reliability which, in turn, reduces risk while growing system efficiencies. 

Moreover, within this structure, several managed services can run on top of large clusters and cluster administration is automated and simplified. Serverless helps run and scale applications and databases without provisioning infrastructure thereby freeing EC2 instance and cluster sprawl and driving higher utilization rates.

Network of regions and edge locations

A network of regions and edge locations for compute/storage would be cost prohibitive and capital intensive for many organizations if it weren’t for the public cloud. Luckily, its network of regions and edge locations can help deliver compute and content storage for globally diversified customer bases. Compute resources can scale horizontally across regions and deliver content through multiple edge locations mitigating network latencies and ensuring customer satisfaction and loyalty.

In built security 

As organizations look to continually shift security left and pursue opportunities to build security in, cloud computing offers a DevSecOps based model with security provided across the stack. From AES-256 encryption for data-at-rest and TLS 2.0 encryption for data-in-transit for many services to key management and Hardware Security Module as a service, and dedicated firewalls (like WAFs) cloud offers a holistic security regime for monitoring, security information and event monitoring (SIEM) for intrusion detection, and dynamic/static code scanning for identifying vulnerabilities and more. Further, security as code can help ensure continuous compliance to security controls.

Disaster recovery and high availability

Cloud helps instrument disaster recovery through a hierarchy of redundancies like multi-Availability Zones (AZs), regions, and multi-cloud. Multi-AZ deployments prevent application downtime due to a failure of an AZ. Cross-region global databases can be used to create redundancy and reduce network latency. Automated backups and snapshots help point-in-time-recovery with the least impact to RTO/RPO thereby ensuring greatest possible uptime to customers, partners and employees. 

Managed services across technology domains

Meeting a wide variety of application and infrastructure needs, cloud services range from compute services like EC2 instances, managed clusters and serverless infrastructure to data services like managed/serverless databases. Data analytics services like big data analytics, data-warehousing, serverless ETL, machine learning services and data integration services can help integrate applications, provide security services for data encryption, and centralized logging and monitoring for across services. While this is a short synopsis of many services offered, you can see how a multi-cloud approach that marries the strength of specific cloud services to IT needs on a case-by-case basis can be a real opportunity to optimize efficiencies.

Choice of managed database services for various use cases

Speaking more specifically about databases, cloud provides a range of database services from the traditional hosted model to managed and serverless solutions like RDS for RDBMS, NOSQL Key value stores, document stores, remote caches, and columnar databases. Managed database services come with a host of features enabling automated backup and point-in-time recovery, scaling using multiple writers, read replicas and sharding, increasing availability using Multi-AZ deployment and cross-region snapshots for disaster recovery. There are tools available to convert schemas and migrate homogenous databases from on-premises to cloud or heterogenous databases from one database engine to another.

Can you prioritize how these considerations will most impact your IT operations and business goals? Let us help. Our team across advisory, implementation, and managed services can help effectively guide your cloud journey.