Can Azure and Hyperconvergence make your Enterprise more Efficient?
One of the major challenges faced by organisations is how to efficiently manage their data centres and public cloud offering infrastructures. We look at how hyperconverged infrastructure and Azure could solve this challenge.
Hyperconvergence is gaining traction in the enterprise as more organisations look to it to bring increased efficiency to their operations. This is one reason why the size of the hyperconverged systems market is set to reach $3.9bn by 2019.
According to figures from IDC, worldwide converged systems market revenue increased 10.8% year over year to $2.99bn during the third quarter of 2017 (3Q17). What’s more, it has been reported that the market consumed 1.96 exabytes of new storage capacity during the quarter, which was up 30% compared to the same period a year ago.
Hyperconvergence offers enterprises a simplified way of managing infrastructure as it enables integrated technologies to be run as a single system through a shared tool set. It is more than just converged systems that brings together discrete hardware products sold as a single SKU; hyperconverged systems are integrated to the point the separate components cannot be broken down.
The hypervisor is more tightly coupled in hyperconverged systems. As well as running hypervisors, a hyperconverged system will manage essential datacentre functions as software on the hypervisor.
Advantages and Disadvantages
As briefly mentioned above, as the technologies within a hyperconvergence system are managed from one tool, this ease of use is a primary benefit for organisations looking to simplify operations. The console can run all functions including compute, storage networking and virtualisation.
Another advantage of hyperconverged systems is the ability to easily scale up. When an organisation needs more resources, it is simply a case of adding more nodes. However, this also highlights a drawback of such infrastructure; all resources need to be increased when any one resource needs increasing. Newer hyperconverged systems overcome this by being either storage- or compute-centric.
Hyperconverged systems are designed to run virtual machines on a hypervisor that is intended to specifically work with the system. As the system virtualises all resources, these can be modified to accommodate greater or fewer virtual machines without having to suspend these. Once virtual machine capacity has been reached, more nodes have to be added, increasing compute, storage and networking resources that can be shared with the remaining VMs.
Hyperconvergence with Azure
Microsoft has improved Storage Spaces Direct in Windows Server as part of its hyperconvergence endeavours to decrease costs and ease complexity in its Azure cloud. Windows Server Storage Spaces is software-defined storage, or virtualised storage for Windows Server.
With Microsoft’s hyperconvergence solution, two identical nodes are required in each cluster at a minimum and each node has its own storage. This combines compute and storage in one cluster. This means it is great for SMEs as this lessens complexity and hardware costs.
For larger organisations, storage and compute can be separated by building a Scale-Out File Servers (SOFS) on Storage Spaces Direct. This helps Microsoft deliver private cloud solutions to smaller enterprises. Among these private cloud solutions is Azure Stack. This is an on-premises version of its public cloud that allows organisations to use Azure without porting sensitive data into a multi-tenant environment.
Azure Stack is certified to run on select hardware and with run in the same way as Azure. It also provides a shared management platform between the public and private cloud. Developers can build on-premise cloud-native applications that can run either on-premise or in the cloud elsewhere.
The smallest Azure Stack deployment is a four-server rack with three physical switches and a lifecycle management server host. These racks can scale up to 12 servers, and racks can be scaled together. An Azure Stack composed of a full-sized, 12-rack server unit can service 400 virtual machines with 2 CPUs and 7 GB of RAM, with resiliency.
Combing the hardware with the software gives rise to an offering called an Azure Stack Integrated System. Furthermore, firms can purchase hardware from one of the certified vendors (Dell EMC, HPE or Lenovo) and license Azure Stack software to run over it.
Organisations can also use existing Microsoft licenses, including Windows Server, SQL Server and MSDN subscriptions to pay for Azure Stack.
For help in optimising your investments in Azure or conversations around cloud economics, please get in touch with your local Crayon office.