HCI solves many issues but not suitable for all kinds of workloads, Rupjit Choudhury, Senior Manager, IS, BPCL
Moving from On-premise, be it DIY converged infrastructure / HCI to public cloud, approaches are different based on the size of the company, complexity and years of their operations
What challenges on the IT Infrastructure side are leading companies to move to the cloud and HCI model?
Business is growing at the speed of light & with this the pressure on existing IT infrastructure has also grown multifold. The onus is on the IT leaders to deliver more to business with less and at the same time manage the IT Infra complexities that arise from business growth.
Traditionally, we had 3 tier siloed architecture consisting of storage, compute and network. There were separate infra consoles and teams who use to manage such infrastructure. Such architecture supported business processes for many years. They had many benefits also, like they could scale independently, often the maintenance activities of such infrastructure were separate due to their isolated architecture.
It also came with many downsides like, handling of multiple solutions from multiple vendors, you have manage multiple consoles. Supports from vendors become very tricky when you had to troubleshoot issues across all 3 layers of compute, storage, network.
The virtualization technology has helped organizations to consolidate their IT to a large extent. In recent years, compute needs and data growth has been enormous. IT leaders needed a solution which not only scale infinitely but also provide an ease of management.
The natural progression in this regards is building a private cloud using HCI & Hyeprscale public cloud. There are also organizations who instead of going for HCI have adopted DIY kind of infrastructure.
I would like to highlight here that Cloud & HCI, both are not mutually exclusive and they only complement each other. The complexity of managing disparate solutions in traditional infrastructure was the most challenging task, which came as a major push for organizations to adopt HCI.
Primarily, HCI is an appliance based solution, where both hardware and software works in harmony. Though HCI has solved many issues, but still, it is not suitable for all kind of workloads and also it is considered a Capex investment. Now moving from On-premise, be it DIY converged infrastructure / HCI to public cloud, approaches are different based on the size of the company, complexity and years of their operations.
For start-ups, they live & breathe in cloud, DC is not affordable for them. But for other organizations,
- Applications which are not well behaved & have nonlinear compute and storage requirement should run from cloud. It will give elasticity & scalability.
- Specific computing need for niche solutions like HPC & Big data cluster are also putting a burden on existing on-premise infra. Cloud can be a solution to this.
- It’s very difficult to build & maintain cognitive solutions like Chatbot, NLP, text to speech etc. on-premises & these are very crucial components for our business applications. Try cloud for such use cases.
- With limited resources available, it’s a challenging task for I&O to maintain vast on-premise infra of Compute, Storage & network.
How do you see the potential of Hyper Converged Infrastructure?
HCI has a huge potential to address some of the emerging needs organizations are seeking. They are exploiting the benefits of cloud for elasticity, agility and TCO. Most organizations prefer hybrid cloud infrastructure to integrate private cloud with public cloud for workload portability and redundancy. Some of the HCI vendors have added cloud management capabilities into their platform but they appear basic when compared with public cloud management platform.
HCI vendors and public cloud providers are now collaborating to deliver more integrated services in the form of Hybrid cloud. This would potentially ease the customer challenges around deploying traditional and cloud-native applications in a hybrid environment. Organizations should prioritize HCI vendors with strong hybrid cloud capabilities across a broad set of parameters, including a wide choice of cloud provider support, on-demand consumption-based pricing and API support.
HCI has the potential to address both traditional and emerging needs of edge computing because of its operational simplicity and form factor. Gartner expects that in the next two years, hardware vendors will support HCI on rugged server platforms to address edge use cases, particularly in retail, mining and manufacturing. Organizations should consider HCI vendors that can deliver lower TCO through support of lean hardware.
In the mining industry, sensor data can be aggregated using gateways, and be analysed on a HCI at the mining site. This process can address issues including predictive maintenance, seismic analysis of mines for personnel safety and emergency response. In comparison to a traditional three-tier architecture, HCI has a smaller form factor and it occupies lesser rack space and consumes lesser power, all these are some of the critical design decisions for edge computing.
How do companies decide when to adopt cloud and when to continue with an on premise model?
- Company should have a cloud strategy in place to decide when to adopt cloud and when to continue with an on premise model.
- Should also consider the in-house expertise availability on cloud computing.
- Any statutory requirement that dictates the solution to run on-premise only.
- We have to compare the ROI, cost of building and running the solution on premise versus cloud. Everything boils to cost at the end.
- Why to reinvent the wheel, for services which are readily available on cloud. Its helps faster time to market.
- In many cases, a particular service may not be available for consumption for customers in a particular region. In such cases, because of various data residency mandates, you may have to consider building it on-premises.
- Mobility, Collaboration, videoconferencing , VDIs, DR , Business Continuity are few of the workloads, companies prefer to run from cloud.
Benefits from HCI, cloud in the long term?
HCI provides a turnkey Infrastructure in terms of integrated server, storage and networking. Customers get an end-to-end systems management and operations management capabilities.
HCI is flexible, both in how it’s deployed and how it is purchased. Businesses can buy HCI as self-contained systems from suppliers, build their own or even rely on software-defined systems. Software-defined hyper-converged infrastructure reduces the risks from vendor lock-in, as software-based systems can run on multiple suppliers’ hardware. It has a smaller datacentre footprint than traditional 3 tier siloed architecture.
Using HCI, you can deploy infrastructure in minutes, so IT teams can focus more to the applications powering the business.
Cloud computing is an example of an operating expense (OPEX), which has multiple benefits. These include having the power to avoid paying any significant upfront costs or making any long-term investments.
Companies that want to stay ahead of the curve must be agile, open to constant rethinking of their business and eager to embrace newer technologies like artificial intelligence and the Internet of Things (IoT) and analytics. Cloud computing is an enabler to all of these, it helps businesses to drive innovation, agility and growth. In Cloud, You can fail fast and improvise.
Cloud & HCI don’t compete but complement each other. The single most common denominator between the Cloud and HCI is the capability of virtualization. They just represent 2 different approaches to infrastructure.