Looking toward 2016, I believe that hyper-converged architectures will be the primary mode of deploying virtualisation in any new Software-Defined Data Centres, and also that private cloud installations will start including hybrid topologies.
[easy-tweet tweet=”In 2016 third-party performance #analytics are going to be required to ensure 360- latency inspection” user=”xangati”]
We can expect that third-party performance analytics are going to be required to ensure 360- latency inspection, because hyper-converged is a software-centric architecture that tightly integrates compute, storage, networking and virtualisation resources, and other technologies from the ground up in a commodity hardware system supported by a single vendor. Without a trusted-source analysing all aspects of performance, the default answer for any degradation issues would be to recommend that you ‘scale out’ with more systems from that single vendor. Service assurance metrics are complicated even further when a hyper-converged infrastructure is extended to the cloud and its resources are shared.
[easy-tweet tweet=”The Dell-EMC merger will drive other #IT systems vendors to promote multiple #hypervisors ” user=”xangati” usehashtags=”no”]
In 2016 I think that the Dell-EMC merger will drive other IT systems vendors to promote multiple hypervisors in addition to VMware’s vSphere; hopefully, this will in turn create growth headroom for KVM (open source), Citrix XenServer and Microsoft Hyper-V. If this happens as envisioned, 2016 will see concepts that transcend the hypervisor-virtualisation orthodoxy, such as containerisation, become much more palatable. Another trend that almost certainly will be spawned by the Dell-EMC alliance is the notion of a custom hypervisor pre-integrated with the converged stack, similar to what Nutanix has done with Acropolis.
Enterprises will explore and pilot many different permutations of containers in order to derive greater agility and scale, as well as operational flexibility. With or without OpenStack, in both private and hybrid-cloud environments, ‘microvisors’ for containers will start to gain traction. From a managed, hosted or cloud service provider standpoint, the advantages of running container instances will help streamline the goal of multi-tenant management. Microvisors, (which are a thin hypervisor with the native intelligence to use only those necessary components of the OS to migrate software to cloud-centric infrastructure) will prove especially useful in migrating legacy applications to cloud platforms.
2016 is going to see key network functions increasingly virtualised
Virtualisation and Automation
2016 is going to see key network functions increasingly virtualised; thus, Software-Defined Data Centres will create more performance challenges when storage, networking and hybrid cloud are added to the virtualisation mix. This will lead enterprises to explore more progressive frameworks for service assurance delivery, which will be increasingly automated and integrated at the orchestration layer. Traditional silos have isolated performance metrics, but these will now be required to share conversations about workload and traffic flows, and take into account all of their inter-dependencies, including v-storm contention and device conflict.
Microservices, which break down IT systems and applications into smaller, more granular elements; Containers, which take applications and services down to a self-contained, component level; and DevOps providing the framework for the IT infrastructure and automation to develop, deploy and manage the virtualised environment with greater agility will begin to see an intersection, which will rapidly increase. The goal will be to create a more seamless management process while automatically tracking, reporting and maintaining minimum thresholds for exceeding service-level targets.