The market for IT infrastructure equipment will be dominated by increased options for customers’ data management and increased demand for solutions that serve specific workloads.
Listed below are the top IT infrastructure technology trends, as identified by GlobalData.
Firms use private clouds to achieve a range of benefits, including improved IT resource efficiency, cost reductions, security, and the ability to gain more control over workload performance, security, and compliance. The use of private cloud solutions will remain strong over the next 12-24 months. Competition between private cloud vendors will also remain intense.
Underpinning edge computing is the cost in time and bandwidth to transport data generated by IoT devices over long distances to be processed at central data centres. Edge computing infrastructure will take multiple forms and will include micro data centres, dedicated edge servers, IoT gateways, and data management platforms, as well as hyperconverged infrastructure for edge deployments. 5G will be both a driver and enabler of edge computing.
High performance computing (HPC)
HPC evolved in the 1960s from early scientific computing objectives for centralised, highly scalable processing in support of singular, compute intensive workloads. Solutions from HP, Cray, Fujitsu, IBM, and many others combined traditional desktop computer CPUs with specialised storage and connectivity resources in a large computing cluster. HPC will expand rapidly over the coming year to embrace probabilistic styles of computing in response to the growing demand for complex workloads such as AI modeling at scale.
The relationship between AI and data centre technologies focuses on two broad areas: AI for IT operations (AIOps) and the introduction of data centre platforms. AI-optimised data centre platforms will become an increasingly competitive market sub-segment over the next 12-24 months. Some platforms will incorporate AI capabilities as part of the overall solution while others will leverage the latest processing technologies and hardware accelerators to support workloads with high performance requirements.
Virtualisation involves the creation of virtual pools of compute, storage, and networking resources that are linked with but decoupled from the underlying physical hardware. VMware, the pioneer of virtualisation technology, accounts for over 80% of VMs with its ESXi hypervisor and vSphere virtualisation platform. Virtualisation software providers will offer solutions to help enterprises transition from rival technology platforms to their own.
Since cloud computing began ushering in a new application development and delivery economy in the form of platform services, policy based applications have become containerised and orchestrated through Kubernetes technology.
As applications have begun to respond to continuous integration, continuous delivery (CICD), they will present boundless opportunities along with complexities associated with moving containerised apps into production. This will be helped by open source software (OSS) technologies such as Istio service mesh, Prometheus monitoring, and other “sidecar” projects.
Data centre hardware includes computer servers, storage systems, networking switches and routers, and converged infrastructure appliances. Enterprise investment in data centre hardware is strongly influenced by demand from hyperscale companies, such as AWS, Google, and Facebook, as well as from colocation providers. One major trend that will shape the adoption of data centre hardware will be investments in hardware specifically designed to support next-generation workloads including high-capacity Ethernet switching and GPU-equipped servers and storage systems.
Silicon photonics is a major trend in the networking industry, but is of increasing importance in the data centre industry as well. Today, the practical application of silicon photonics is in pluggable optics for networking where the new packaging brings manufacturing and cost reductions. Companies like Cisco, Intel, and Macom are investing in photonic circuitry for networking, and for use either on die for chips or for interconnects on circuit boards.
Both legacy back-office and modern cloud-first solutions share one common denominator: data. This is big data, historically associated with the Apache Hadoop storage framework Regardless of the underlying data storage platform, when coupled with supportive data processing technologies like Apache Spark, these big data platforms allow companies to ingest, process, and analyse tremendous amounts of data from a wide array of sources. We expect vendors to continue to invest in solutions such as Dataproc to shift discrete, splintered data storage to a unified platform.
Software defined networking (SDN)
SDN has settled into three camps dominated by Cisco, VMware, and a scattering of OpenFlow. The SDN market feels like it has stalled because there has not been a typical 2.0 moment, but it is moving quickly into new areas.
Quantum computers could open new market opportunities across security, life sciences, manufacturing, and many other industries. It will be some years yet before quantum supremacy is achieved, and many years before it is commercially available. For the next few years, we expect to see early movers focus on hardware and education.
This is an edited extract from the Tech, Media, & Telecom Trends 2020 – Thematic Research report produced by GlobalData Thematic Research.