Today we will try to draw a portrait of the main consumer of the hyperconverged systems, as well as talk about performance and touch on use cases for using the platform.
To begin with, hyperconverged systems, in principle, and Cisco HyperFlex in particular, very ergonomically fit into the concept of data center development at any level. This is conditioned, first of all, by the ease of scaling and minimization of time and labor costs for commissioning additional capacities. The key word in this solution is unification. It manifests itself at all stages: from implementation and scaling to product maintenance.
Such an approach makes it possible to simplify the design of all processes for commissioning and operating the system as well as to minimize the set of directives for processes at all levels of interaction. In other words, we can significantly reduce the costs of integration, scaling and maintenance of the entire complex due to a unified approach of hyperconverged systems to the organization of infrastructure.
But the advantages described above can be extrapolated not only to data processing centers. The solution may be interesting to business as a basis for building on-premise or hybrid infrastructure. The minimum requirements for organizing a HyperFlex cluster are 3 nodes per factory, this allows you to meet the needs of small and medium businesses without capital expenditures at the implementation stage.
In the first part, we deliberately omitted the issue of connecting branch office and now consider it in detail. It is not always advisable to implement a full-size cluster in all small branch offices, and here the HX2xx Edgeservers, created specifically for this case, come to the rescue. If you need to connect, you do not need to buy a factory – we just take a cluster of three Edge nodes and connect to the existing branch network. Starting with the fourth version of the platform, the implementation of a cluster of two nodes will be available.
Management is carried out centrally using VMware vSphere plug-in Access or through the appropriate interface from Microsoft.
Containers and Hybrid Cloud
You cannot ignore the direction that has not been the first year seeking to peak its popularity in the development, deployment and management of applications. It is, of course, about working with containers. In this area, Cisco has recently taken major steps. The Kubernetes-based Cisco Container Platform simplifies application deployment and helps erase the boundaries between the local containerization and cloud infrastructure, allowing you to flexibly configure and manage clusters, as well as interact between different application deployment environments.
In addition, Cisco offers many hybrid cloud management solutions: AppDynamics, which monitors application performance; Cisco Workload Optimization Manager, which automates the deployment of work tasks, as well as the Cloud Center, which helps simplify work with virtual machines and containers in both private and public clouds.
I would like to add that Cisco is working closely with Google on the creation of hybrid clouds, simplifying the interaction of locally deployed clusters with the Google Cloud Platform. This allows you to simplify and make transparent bidirectional exchange in the combined clouds.
Together with Amazon, Cisco is developing Cisco Hybrid Solution for Kuberneter on AWS, which is designed to solve hybrid cloud containerization management and consolidate local capabilities using Cisco Container Platform with Amazon Elastic Container Service for Kubernetes.
Highly loaded systems
Another obvious niche for using the solution is to deploy a fully virtualized SAP landscape with on-memory SAP HANA DBMS. HyperFlex was certified by SAP in October 2018 and fully meets the requirements for high-performance environments imposed by SAP (the maximum certified configuration for SAP HANA is 768 GB).
If it is necessary to use other high-load platforms in parallel with a hyperconverged system, such as a Hadoop cluster or a machine learning platform, then they can be connected to the same Cisco UCS factory, which means that they will have a unified management model for physical servers. UCS profiles.
So what is the portrait of the main consumer of the solution? Clear portrait did not work. It has many faces, because the advantages of hyperconverged platforms will have a positive effect on processes organized on deployments of any scale – from a small corporate infrastructure to distributed data centers.
We did not describe yet another obvious use of such platforms – VDI, since the unification of management and the simplicity of scaling the solution make it extremely convenient to use virtual desktops.
In conclusion, I would like to say that the hyperconvergence market – both according to Gartner and IDC – is growing and is clearly aimed, if not to oust, then to take strong positions along with the classical organization of infrastructure. This contributes to the main word that characterizes these systems – simplification. Simplify deployment, simplify scaling, simplify administration and maintenance. And this is good. There are enough difficult decisions in the world – it is much more difficult to make them a little closer.