Build Your Own Cloud with HCI: Hyper-Converged Infrastructure
Build Your Own Cloud with HCI: Hyper-Converged Infrastructure
In this article, get everything you need to build your own Hyper-Converfed infrastructure in the cloud.
Join the DZone community and get the full member experience.Join For Free
What if you can build our own tiny — prototype kind of cloud computing with something mini low-level kind of software. Something that can convert your hardware resources that can grow by intelligently virtually slicing or aggregating them as per go basis irrespective of different make and specs of underlying bare metal underneath. HCI — Hyper-Convergence Infrastructure( an Intelligent software that provisions elastic, scalable hardware compute) is the concept behind it.
Let’s see, A typical data center commonly describe as 3 -tier (compute/server, storage, networking ). Whereas modern data center has an approach of provisioning these 3 tiers via a software-centric approach that in a common we are talking about HCI. It combines data center resources ( Compute, Storage, Network ) by using intelligent software to create flexible building blocks that can plugin adds on storage, compute, etc. that replace legacy infrastructure consisting of separate servers, storage networks, and storage arrays with pre-defined specs.
Simply put — HCI — The virtualization software abstracts and pools underlying resources, then dynamically allocates them by slicing it to applications running in the form of VMs or containers, VDI ( Virtual desktops ). That provides an abstraction over the physical layer in the data center that makes them elastic and scalable along with supported upgraded capacity or spec grade hardware on the go without patching, restart, or going into a canary mode that was previously a bottleneck for a data center.
If you want to increase compute or storage or change their specification i9 to xenon for say to upgrade (that can be heterogeneous ), data center with HCI can help by just plugin new hardware in the rack and start scaling up additional capacity with intelligent software (HCI — Hyper-Convergence Infra ) without going into maintenance mode or flip, is not it sounds — Wow! We will see how to code to get a mini HCI prototype so that we can combine couple fo laptops to convert it into our very own little cloud.
Hyper-converged infrastructure allows you to easily scale up by adding or replacing drives in existing servers or scale-out by removing nodes ( Laptops in our case ) to a cluster. This usually means that even if you only need additional storage or compute, the new node will also come with only computing or storage. Nodes can be added one at a time for incremental growth, whereas storage arrays often require a shelf of drives, which is a much larger single spend for data centers.
Hyper-converged infrastructure allows you to easily scale up by adding or replacing drives in existing servers or scale-out by removing nodes (Laptops in our case) to a cluster. This usually means that even if you only need additional storage or compute, the new node will also come with only computing or storage. Nodes can be added one at a time for incremental growth, whereas storage arrays often require a shelf of drives, which is a much larger single spend for data centers.
That results in :
- World-Class Platform: Each hyper-converged server (node), includes Intel-powered x86 hardware with flash-based SSDs and traditional HDDs, along with HCI.
Unprecedented Flexibility: A single cluster can have unlimited nodes, with node types having differing amounts of storage, CPU, and memory resources, so you can run multiple workloads with maximum efficiency.
What Can You do with HCI?
- Build a private cloud: Deploy cloud-like infrastructure on-premises with lower costs, more control, and improved security. Choose from multiple hardware possibilities from our certified partners.
- Achieve true hybrid cloud: Manage a mix of VM- and container-based applications, deployed across a mix of data center and edge environments.
- Let us not get into public Cloud Space but they are relying on this magic for provisioning elastic, humongous data centers that we use as AWS or Azure.
How goes inside HCI code ?? → let’s less talking more coding. RedHat system engineering open source code : https://github.com/RHsyseng/hci
Cloud principles derived from HCI: This desire for cloud‐type consumption is directly driving the adoption of HCI, which provides a much higher level of flexibility, simplicity, and on‐demand growth than traditional data center infrastructure while delivering the on‐premise benefits of high‐performance, security, and control.
Also, next‐generation HCI solutions deliver strong support for deploying a hybrid cloud model, whether through integration with OpenStack technologies or supporting current generation workloads, like cloud‐native applications and containers.
Looking at Key Use Cases for Hyper‐Converged Infrastructure: Although not limited to these use cases, the benefits of HCI are especially evident across these common use cases:
- Hybrid cloud / Private Cloud
- Virtual desktop infrastructure (VDI)/end‐user computing
- Public Cloud Availability Zone / Region Data Centres
- Cost and maintenance optimization for on-prem private cloud or data center
This approach rewrites the old capital expenditure (CapEx) and operational expenditure (OpEx) rules for enterprise storage. Storage components no longer cost premium prices. Your IT shop is free to select the specific components that best meet your particular requirements for performance, cost, and capacity.
With HCI, storage components share compute and memory with the server infrastructure. This eliminates the need for separate storage arrays, controllers, memory, and more. All storage technologies are fully integrated into the virtualization cluster.
Please do not take it as a type of bare metal or embedded hypervisors ( that virtualizes CPU and another physical sliced resource to get us virtual machines for Virtualisation.
Cloud computing glossy for understanding a few terms mentioned above:
- Public Cloud: Public Vendor offering like AWS, Azure, GCP, etc. — Pay as you go basis subscription
- Private Cloud: On-premises cloud-based and hosted over On-prem data centers
- Hybrid Cloud: Enterprise solution approach leveraging mia x of On-prem ( Private cloud ) + Public cloud to harness benefits of security, performance, regulation, and compliance (especially Financial and Health providers )., Geographical regulation like GDPR or HIPPA compliance.
- Multi-Cloud: Emerging strategy of Enterprise solutions to with help of abstractions like containers or IaaC ( infra as code ) service to harness the best of all vendors to optimize costing, performance, and decoupling solutions from underlaying platforms. Still, this approach is getting debated and accelerated both.
Credit and Reference :
Open Source HCI: Github for code https://github.com/RHsyseng/hci
Research Paper: https://www.researchgate.net/publication/335910358_Study_of_Converged_Infrastructure_Hyper_Converge_Infrastructre_As_Future_of_Data_Centre
Opinions expressed by DZone contributors are their own.