Introduction to Cisco's Hyper Convergence Platform : HyperFlex 3.0

Today I am going to talk about the next generation compute and storage platform which are intergerated to make one platform for datacenter. According to IDC FutureScape, Worldwide IT Industry 2018 Predictions by 2020 – “60% of all enterprises will have fully articulated an organisation wide digital transformation platform strategy and will be in the process of implementing that strategy as the new IT core for competing in the digital economy”. 

When enterprise are considering to streamline structure they actually necessity a Comprehensive,  Enterprise grade, Hyperconverged platform that sets the foundation for multi-cloud/hybrid cloud forthcoming, and brings the speed, simplicity, and scale essential for the business to prosper today. Giving a cloud experience within the datacenter.

What is Hyper-Convergence ?
Hyper-Convergence is a Software centric infrastructure architecture which integrates the features of compute, storage, networking and virtualization for the resources in the enterprise network.
Starting from the beginning for the converged network where you have various systems to support various features in the infrastructure like you have servers, Storage systems, software define Infra and so on in different boxes.

Hyper Flex is the solitary HCI solution in the marketplace. When we turn complete on Compute, Storage, and Network integrated into a single solution, provided by a single vendor.  The network is the most critical component in both Data Center and ability to connect to and effectively use public or other private cloud solutions.

Fig 1.1- Hyper Flex Reference Architecture

What are the advantages of the Hyper Converged Platforms ?
Within the Data Center containing the network in the HCI platform, It provides enterprises with 4 distinct advantages.
  • It makes deployment, configuration, and expansion much frequently and easier. With HX- HyperFlex, the service profiles are pre-built templates.
  • As you know, it reduces cost. All nodes plug into the Fi’s ( 1 port per FI per node; 2 total per node for HA )  which means we only need 4 ports on the Top of Rack switch. VS 4 ports per node for any other HCI vendor. Those ports are expensive and even in a small cluster 4-5 nodes you need 20 ports, then you add a node. 4 more ports! Not the case with HX
  • It provides a dedicated network for all inter-node inter-cluster traffic as much like a SAN.
  • Last but not least, it means the entire stack: Compute, HCI SW, Storage, and Network are all provided by and supported by 1 vendor. Which means better support for you and no vendor finger pointing.
As mentioned above it is the most important factor in enabling access to cloud services outside the Data Center. As traditionally it all started with the traditional 3 tier architecture of Compute, Network and storage provided by individual companies as standalone layers. 

The next evolution came in converged platforms which brought a lot of advantages like in some cases single vendor support, speed of deployment, and a validated building block approach. 

What's new in HyperFlex 3.5 ?
Cisco has co-innovated with an entire band of technology leaders such as Intel, SAP, Citrix, and NVIDIA. These new innovations continue to change the face of hyperconverged infrastructure, transforming it into a complete platform for multi-cloud.

So Cisco have now cluster expansion for Hyper-V converged nodes.They also have Cluster expansion for Stretched Cluster compute-only and converged nodes. Apart from it Cisco HyperFlex 3.5 Integrated Hyper-V and Windows Server OS bare metal installation included as part of cluster creation workflow. It will also have Support for VMware ESXi lockdown mode to increase security of an ESXi host by limiting access allowed for the host. It has ability to create applications for AI/ML with NVIDIA Tesla V100 GPUs integration within HyperFlex nodes

Their focus was primarily on ease of use, and providing a smaller building block ( a node which is compute and Storage ), and they did a pretty good job of it. However they left a critical part out of the infrastructure and that part is the Network.

We will discuss on Cisco HyperFlex 3.5 in details in our next article.