Cloud is no brainer anymore. Enterprises have moved or moving to cloud in an incredible speed. Recent statistics shows the public cloud service market is expected to reach $623.3 billion by 2023 worldwide, 83% of enterprise workloads will be in the cloud by 2021, 94% of enterprises already using a cloud service, 30% of all IT budgets are allocated to cloud computing. By 2025, the data stored in cloud data centers will exceed 100 Zettabytes. The onset of pandemic has enforced Enterprises reinvent with Cloud – for example how Cloud was adopted by Ed Tech, Med Tech, Online Collaboration, Fin Tech, Manufacturing etc.
However, Cloud brings its own challenges too. Firstly, contrary to the perception Cloud is not cheap, rather costlier if not managed and governed properly. Secondly, once it is in Cloud, Enterprises had limited control, hence a robust architecture and management technique will be essential. Thirdly, Cloud is always innovating, adding new services / expanding existing services. Enterprises need to keep pace with the changes with new Cloud Services. Last but not the least how do we keep a balance between leveraging Cloud Native Services vis-a-vis avoiding the lock-in with a Cloud Service Provider. Also, for our time critical applications like Patient care, Industrial automation, Video on Demand etc. we need to offload certain part of the workload from Cloud and move nearer to the user.
So, Enterprises are also innovating, modernizing and re-engineering their Cloud Roadmap. Here are the few critical areas which Enterprises needs to focus on-
Cloud Consumption Optimization:- Cost of Cloud Consumption especially for workloads which consumes lot of Cloud Computing as well as Cloud Storage is a key focus area for Enterprises. Examples could be Electronic Design Simulation (EDA) workload for Semi-Conductor Industry, Risk Evaluation & Fraud Detection for BFSI, Fluid Dynamics for Automotive, Genomics workload for Healthcare etc. There are multiple Techniques and tools like ML based Workload Orchestrator which creates models for different workloads using historical data and works along with the scheduler to allocate right sized CPU, Memory & Storage so that Cloud resources are best optimally used.
Few key thumb rules usually followed are-
o Compute:- Optimized choices for CPU usage, memory, storage, networking profile and scale horizontally or vertically based on event triggers in compute requirements. Use more of Cloud Native Services as they are tightly integrated and consume less.
o Storage:- Hybrid storage model using both disk and object storage as applicable, Robust data backup and DR mechanism with a combination of hot and cold stand by based on business criticality, use of Catching enabled distributed storage.
o Network:- Network performance monitoring and Improvement, bandwidth cost improvement mechanisms, CDN usage for latency and network cost reduction.
Reliability & Resiliency:- As we deploy on Cloud Application reliability and failover mechanisms are becoming integral part of Cloud Design and Testing. There are various techniques employed today to simulate the failure scenarios and test how the application can withstand the fault injections. Popular among them is Chaos Engineering for discovering vulnerabilities in a distributed system. This requires injecting failures and errors into software during production. Once you intentionally cause a bug, monitor the effects to see how the system responds to stress. Through Chaos Tools admin can identify weak points in the System, see in real time how a system responds to stress.
In addition to this, modern Cloud Architecture uses Observability to monitor hardware, software stack & applications to predict the failures and generating insights from hidden behavior. Observability relies on telemetry derived from instrumentation that comes from the endpoints and services in multi-cloud computing environments. In these modern environments, every hardware, software, cloud infrastructure component and every container, open-source tool and micro-service generates records of every activity. The goal of Observability is to understand what’s happening across all these environments and among the technologies, so user can detect and resolve issues to keep systems efficient and reliable and end customers happy. Once root causes are identified through Chaos and Observability, site reliability engineers use software fix and roll out so that gradually over the time system becomes fault resilient.
Cloud Native vs Cloud Agnostic:- It’s a big dilemma always. Should I only use cloud native services to build my Cloud Backend Platform & Applications. While we always desire that our platform/applications have minimal lock in so that we have freedom to move across CSPs, it will not be a wise decision to leave behind the ease of using Cloud Native Services, Time to market and Cost benefits CNS provides. Also, Cloud native applications are designed to take full advantage of on-demand dynamic allocation of computing resources, higher asset utilization etc. leading to a sustainable solution. Rather we should adopt a balanced strategy. Having a well thought out application migration strategy and a shared vision for the future of your application architecture should be a key part of your strategy which will help to balance.
Data Strategy & Automation:- When utilized efficiently, data can provide quite a valuable view into the business such as its processes and its activities. 7 out of 10 valuable enterprises are actively taking a data driven approach to business strategy. So what does it mean to be a data driven enterprise?
It means that decisions are fact based. For a fact to be true, it needs to be-
• Contextualized– what is the boundary in which the fact is true and therefore the relevance of the decision.
• Cross pollinated– is enterprise data siloed or have relationships been identified correctly across the organizational units or even across enterprises.
• Correct– is the data clean and consistent.
While Data lake is an integral part of enterprise-wide strategy, a data architecture representative of the enterprise needs to mark all the boxes above to provide good actionable insights. Data lake is seeing evolution that goes beyond conventional data catalog features by leveraging a combination of human expertise, machine learning, rules-based decision-making to provide maximized value.
This is a mammoth task and can only be implemented accurately if the data pipeline and downstream analysis processes are completely automated. The data pipeline process of capturing the data, enriched with context, cleaned and mapped across organizational units need to be designed keeping in mind an approach of feedback loop and continual improvement and optimization.
Edge is the next evolution of Cloud:- With proliferation of IOT devices, rapid progress in connectivity with 5G, need of real time decisions with zero latency like Retail, Automobile, Manufacturing Shop Floor, Hospitals Edge will be rapidly adopted. Edge also provides a fixed cost whereas Cloud is a variable cost. Edge + Cloud gives the best optimization for many of these scenarios. According to a report by ACG Research on hybrid cloud economics, there is an over 1600% difference between network transport costs at the core versus the edge. Given that network transport costs are by far the biggest share of cloud TCO, bringing the public cloud to the enterprise edge can easily save millions. Enterprise Security is also better at Edge. In the Edge, the local cloud gateway enables you to get to the cloud via a dedicated private connection, which is far more secure than the public internet.
Securing Cloud Enterprise:- Cloud security is a critical requirement for all the organizations. Especially with the latest research from (ISC)2 reporting 93% of organizations are moderately or extremely concerned about cloud security and one in four organizations confirming a cloud security incident in the past 12 months. There are several regional as well as Industry specific security regulations and most of the Cloud Service Providers adhere to that. Cloud security is a joint responsibility between CSP & Enterprise. CSPs like AWS provide multiple Security Services and it’s important to incorporate as many as Cloud Native Security Services in the overall Enterprise Cloud Security Strategy. Periodic VAPT & Penetration Testing is essential to identify the weakness in the overall Enterprise and take proactive steps to improve Security.
Technical Debt Reduction:- Last but not the least Enterprise should have a periodic and continuous process for Technical Debt Identification, Reduction and Modernize the Products / Applications keeping in pace with the innovation in Cloud as well as Technology.
List is long, but these are the few essentials Enterprises must focus for their Cloud Journey.
