Machine Learning and Meta-Clouds Next for Cloud Services (Part 1)
Innovation in cloud services ushers in a new era for data center IT, re-defining market landscapes
by Dr. Cliff Grossner
Part 1 of a 2-part series. Don’t miss part 2.
Enterprises are migrating applications to the cloud, to improve agility and reap cost savings. Agility means enterprises can shorten the time needed to introduce new applications and either increase or decrease compute capacity to fit business need. Upfront capital expenditures (capex) can be shifted to as-needed operating expenditures (opex) using off-premises cloud services—shifting from investments in equipment and staff, to leveraging a cloud service provider’s (CSP’s) automated data center infrastructure, supported by highly skilled data center experts.
Although the market has achieved a degree of maturity, it still lacks a commonly accepted set of cloud service definitions; thus, we define off-premises cloud services as follows:
- – Infrastructure as a service (IaaS) includes any combination of data center facilities (including colocation), servers, network, storage, database, network applications (layers 4-7), and management
- – Cloud as a service (CaaS) provides an application execution environment — including servers, network, storage, management, and data center orchestration software (cloud operating system[OS]) — purchased as a bundle, with pricing based on usage
- – Platform as a service (PaaS) provides an application development and execution environment —including application run-time and middleware (web servers, database management systems), servers, network, storage, management, and data center orchestration software (cloud OS) — purchased as a bundle with pricing based on usage
- – Software as a service (SaaS) provides a complete application with a pay-per-use pricing model — including applications like customer relationship management (CRM), enterprise resource planning (ERP), collaboration, security, management, virtual desktop, and business analytics
The growth of off-premises cloud services is enabled by shifting to a cloud architecture in the enterprise and CSP data center, with the deployment of virtualization and data center orchestration software, for virtualization of servers, storage, and network, and for automatic deployment of applications. Data center orchestration software refers to software that provides automated coordination and management of resource pools in the data center, such as servers, storage, and networks. It includes policies describing desired service levels, required workflows, and actions for adjusting resources, to achieve the desired service levels.
Virtualization and data center orchestration technology enables large-scale operational efficiencies and capex saving for CSPs, because they leverage a shared pool of resources for compute, storage, and networking. This technology fuels the expansion of all types of off-premises cloud services. Revenue for off-premises cloud services, including IaaS, CaaS, PaaS, and SaaS, was up 54 percent over 2014, reaching $93 billion in 2015. IHS expects the off-premises cloud service market to reach nearly $278 billion in 2020, with a five-year compound annual growth rate (CAGR) of 24.5 percent.
Off-Premises Cloud Service Roadmap
The cloud services market got its start with CSPs providing simple forms of off-premises IT services: physical data center facilities, colocation hosting, and SaaS. The rise of on-demand computing followed, allowing CSPs to offer public cloud IaaS, and the introduction of data center orchestration technology in CSP and enterprise data centers.
In 2014 and 2015, enterprises achieved an objective of cloud computing that had remained elusive: on-demand agile computing, allowing automatic overflow of workloads during peak demand periods from their own data centers to CSP-provided facilities.
From 2016 to 2019, the hybrid cloud architecture gives way to a distributed architecture, where enterprises consume services from multiple CSPs — called Meta-Clouds — driving additional cost and performance optimizations. We also expect to see the emergence of a wholesale market for cloud infrastructure.
Computing resources will become a commodity and will be traded on a commodity exchange after 2020. Many software and equipment vendors will switch business models, providing their offerings as off-premises cloud services, rather than following traditional software-licensing and hardware-driven models.
Exhibit 1 Meta-Clouds Coming to Off-Premises Cloud
Enterprises Creating Custom Clouds
This year enterprises will continue to extend their off-premises cloud service usage across many clouds, creating Meta-Clouds, with enterprises consuming services from multiple cloud service providers, driving additional cost and performance optimizations. The use of multiple CSPs is already well established, as we learned in the IHS 2016 Cloud Services Strategies North American Enterprise Survey, where respondents indicated they will use an average of eight different CSPs by 2018. This result may seem surprising, but it is in line with a market containing many specialized players that can meet custom requirements of off-premises cloud services users, such as the following:
- – Offer assurances that all operations (including company management) are local
- – Provide branded SaaS applications, such as SAP, Salesforce.com, and Office 365
- – Offer various flavors of IaaS that specialize in virtualized or bare-metal servers with specific OS builds, as well as specialization in containers
Part 1 of a 2-part series. Read part 2 now.
Cliff Grossner, Ph.D
Dr. Cliff Grossner is a Senior Research Director responsible for data center compute and networking, cloud services and SDN research at IHS Markit technology . He has more than 25 years of telecommunications industry experience covering market analysis, corporate and product strategy, product management and marketing, and scientific research. A recognized thought leader, he is also a frequent expert judge for industry and technology innovation awards and an invited speaker at industry conferences and is frequently quoted in technical publications such as Light Reading, Fierce Telecom, eWeek, Network Computing and Lightwave Online.
Cliff joined IHS (now IHS Markit) in December 2014, when IHS acquired Infonetics Research where he led the firm’s data center networking, cloud and SDN research coverage. During his tenure at Infonetics, Cliff launched the SDN practice and grew revenue 100%.
Prior to Infonetics, Cliff held a range of senior positions in the information and communications technology industries including heading strategic marketing for Alcatel-Lucent’s enterprise network business unit, tenures at Bell Labs, Sesame Networks, NewStep Networks, and Nortel. He has also worked as a research scientist and faculty lecturer at McGill and Concordia universities, specializing in artificial intelligence, distributed computing, and computer architecture.
He earned his Doctor of Philosophy at McGill University, and his Master of Science in Computer Science at Concordia University, both in Montreal, Quebec, Canada He holds multiple patents in computer networking, networking embedded security and telecommunications applications.