Since the inception of the cloud, large monolithic infrastructures have been the norm. Azure, AWS and Google Cloud all offer almost infinite scalability and relatively low cost. However, is the dominance of the big three cloud service providers waning?

Businesses have been increasingly creating smaller hybrid cloud structures to meet their needs. By mixing on-prem and larger hosted platforms, they have been afforded greater choice and the ability to develop specific cloud infrastructures.  

However, are we moving into an era where bespoke cloud services become popular, as businesses look to create ‘boutique’ clouds – offering more personalisation and a specific set of features often linked to one service application?

Speaking to IT Pro, Nick McQuire, senior vice president of enterprise research at CCS Insights, says: «The definition of what a boutique cloud is remains open. You could argue, for instance, that a private managed cloud is also a boutique cloud, so I think we need to define what we mean.”

“I think the future of cloud services is a real mix. The hyperscalers will always be there, as will the hybrid cloud infrastructures,” McQuire continues. “Inside of these, we may see more specialised cloud services, which could be described as ‘boutique’ for specialist sectors such as financial services, or to meet specific regulatory requirements.”

The industry is already seeing a move to multi-cloud deployments: Microsoft’s Azure Arc – a rebranding of its Data Box Edge hardware – for example, focuses on the burgeoning IoT and edge computing space and is in public preview. The idea is to bring VMs and containers to any infrastructure no matter irrespective of needs of size. Enterprises with specific requirements for their cloud deployments could create a boutique cloud within the Microsoft environment.

For many industry watchers, the next battle will take place across the multi-cloud, as enterprises continue to focus on building the bespoke services they need. Already the major players are jockeying for position: Microsoft has Azure Arc, Google has debuted Anthos, IBM will run its services on multi-cloud management systems and Cisco has its CloudCenter Suite.

The adoption of cloud services will continue. According to Gartner, by 2022, nearly a third (28%) of spending on essential IT services will shift to the cloud.

Michael Warrilow, research vice president at Gartner, explains: “Cloud shift highlights the appeal of greater flexibility and agility, which is perceived as a benefit of on-demand capacity and pay-as-you-go pricing in cloud.”

Is smaller better? It all depends on the specific business need. What is certain is the cloud environment is rapidly changing. We are moving out of its first phase of development to more refined services and flexible infrastructures.

Compact and bijou

Research from Flexera illustrates how hybrid cloud adoption has expanded, with 84% of enterprises have a multi-cloud strategy. Enterprises with a hybrid approach (combining public and private clouds) grew to 58% last year according to their survey.

Flexera also revealed: “Among enterprises, the central IT team is typically tasked with assembling a hybrid portfolio of clouds. This year, while 31% of enterprises see public cloud as their top priority, a combined 45% of enterprises see a hybrid cloud or a balanced approach between public and private as the biggest focus. Only 9% of enterprises are focusing on building a private cloud, and 6% see their top priority as using a hosted private cloud.”

The multi-cloud and hybrid cloud have continued to expand to become the dominant form of cloud service infrastructure. But as we move closer to real-world deployments of 5G and edge computing, the multi-cloud may change again to become more boutique as services specialise.

CCS Insights’ McQuire explains: “If your business is a complex IT environment, then existing suppliers like IBM and Red Hat, for instance, will be able to offer you the services your business requires. The mix of on-prem and public cloud isn’t going to go away anytime soon. However, what I think we are beginning to see the first green shoots of is the large players pushing into specific industries. IBM last year, for instance, launched a cloud service aimed at the financial sector.”

Commenting on the research hic company conducted with Freeform Dynamics, Hiren Parekh, UK country leader for OVH, adds: “Our results show strong interest in working with specialist cloud providers, with 21% of organisations committed to using providers aligned to specific applications or infrastructure; 18% are committed to using specialist providers focused on particular use cases and 17% are committed to local providers who cover a specific geography. This echoes what we are hearing from our customers, suggesting demand for cloud providers of all sizes and underlining the popularity of the multi-cloud approach.”

A specific need will drive the business case for smaller cloud services. Large cloud deployments can become unwieldy with businesses often feeling they have little control. Hybrid cloud infrastructures have addressed these anxieties to a degree, but we could see more refinement in how companies buy and organise their cloud services over the short term.

A small cloud future

Smaller cloud service providers such as Vultr, Packet, UpCloud and Linode offer compact and specific services, which could define both what the boutique cloud means today and how some cloud services could be bought over the next few years, particularly by smaller businesses.

“I see a cultural shift in risk appetite which we see across the whole spectrum of technology,» says Justin Day, CEO of Cloud Gateway. «Businesses are seeking out smaller cloud service suppliers because they offer better flexibility and more agile working while also being more focused or simply better at delivering more niche services. Because at the end of the day, it’s about allowing businesses of all sizes to get the very best out of their cloud systems and leverage the very best out of those service suppliers.»

Adam Bradley, UK MD of Ekco, believes service levels are pushing companies towards smaller cloud service providers. “I think people are just sick of bad service,” he says. “They are sick of waiting for someone who knows what they are talking about to call them back, and they are tired of having to do all the hard work themselves.”

Cloud services have become somewhat horses for courses. CTOs tasked with adopting an agile cloud-based IT infrastructure have often found themselves managing what could be described as cloud sprawl – running several cloud deployments from several vendors. Businesses are rationalising their use of the cloud.

As cloud services have matured, it has opened the door for smaller service providers who can focus on specific sectors or industries. Building cloud services for these highly defined spaces is a crucial trend through 2020. In the medium term, whether these boutique vendors can remain viable faced with shifts by the large cloud suppliers towards multi-cloud and specialist cloud services, remains to be seen.

HPE has expanded its Small Business Solutions portfolio with the new ProLiant MicroServer Gen10 Plus, which the company claims provide industry-leading capabilities that will help SMBs drive growth and digital transformation.

Designed to address businesses’ budget, IT and special business needs, the system offers automation, remote management, security capabilities, as well as Intel Pentium and Intel Xeon E processors.

Customers can use the new MicroServer, which is the size of a typical hardback text book, for less than $20 per month. And HPE says the offering is as easy to set up as a smartphone. In fact, the new offering weighs in at just 10 pounds and is a third of the size of existing server market products, the tech firm added.

“We are committed to helping small businesses innovate, serve their customers, and drive growth and digital disruption by empowering them with enterprise-class technologies that uniquely address their needs for IT expertise, budget and space,” said Tim Peters, vice president and general manager at Global SMB & Mid-Market, HPE.

“The design of our latest HPE MicroServer and strategic pricing model was inspired by our SMB customers to meet their expectations for the most economical, secure and easy-to-manage solutions that supports their entire business operation.”

The new ProLiant MicroServer Gen10 Plus offers a range of capabilities that HPE said enable faster performance, data protection, automation and ease-of-management. For starters, the inclusion of Intel Pentium or Intel Xeon E delivers compute support for virtualisation and database workloads but registers at just 36 decibels for versatile placement.

The system is also the first ProLiant MicroServer – and the industry’s only server family – to provide the HPE-exclusive silicon root of trust technology, the company explained, which extends security protection at the silicon level.

Users can also use cloud-based AI management tool HPE InfoSIght for Servers, HPE Integrated Lights Out 5 (iLo5) for remote management, as well as flexible options for both in-office operations and the cloud.

“SMBs are looking for easy-to-manage solutions that can scale as needed. Solutions, such as this one from HPE, addresses this demand with small businesses by delivering enterprise-grade technologies, which combine servers, software, networking and cloud capabilities that are easier for small business to adopt and manage regardless of their in-house IT capabilities,” commented Shari Lava, research director, Small Medium Business (SMB) Research Program at IDC.

​Amazon Web Services has unveiled a free open source operating system called Bottlerocket, designed specifically to run containers on bare metal or virtual machines.

Bottlerocket is being pitched as a purpose-built operating system designed with a single-step process to make it far easier to automate updates, while also cutting out much of the unnecessary elements found in general-purpose software.

Its biggest selling aspect is its dual partition setup, running as active and inactive. When an update is issued, the inactive side is changed first, with the system then switching the positions of the partitions in order to complete the update.

The OS also uses image-based updates, which means the update can be rolled back in its entirety if necessary, helping to reduce downtime and minimise process failure. This is in contrast to most general-purpose operating systems which use a package-by-package approach.

As part of the slimmed-down design, Bottlerock takes a different approach to authentication and secure login normally found on general-purpose systems. There’s no SSH server to support secure logins, although users can use a separate container to access admin controls.

The new OS also supports all the container tools you might expect, including Docker images and anything conforming with the Open Container Initiative standard. The system is also built with some third-party components, including the Linux kernel, the container’s runtime, and Kubernetes. 

The OS is currently in a preview state, and is hosted on GitHub alongside a host of tools and documentation to support its use. Among these is a Bottlerocket Charter, which claims that the OS is open and «not a Kubernetes distro, nor an Amazon distro», adding that such a platform can only be built with the support of a wider community.

Despite its open nature, the OS is optimised to work best with AWS tools out of the box, specifically Amazon’s Elastic Kubernetes Service (EKS).

The OS is currently available in a free public preview as an Amazon Machine Image (AMI) for Elastic Cloud Compute. Once released to general availability, Bottlerocket will come with three years of support, incorporated into AWS support plans at no extra cost.

VMware has announced the launch of a number of new Kubernetes-focused products, including the latest version of its vSphere platform.

Most of the new products fall under the company’s Tanzu portfolio, unveiled at last year’s VMworld. 

Tanzu represents VMware’s efforts to integrate Kubernetes container management – which the company is betting big on as the next significant step in enterprise applications – with its existing VM management tools. 

Three new Tanzu products are being introduced; first up, Tanzu Mission Control, a tool previewed as part of last year’s announcement which is designed to help enterprises manage multiple Kubernetes clusters across a range of environments, while centralising key functions like security, configuration management and data protection. It also allows businesses to hook VMware’s other management and monitoring tools (such as its Wavefront and CloudHealth products) into its Kubernetes workloads.

Following on from this is VMware’s new Tanzu Application Catalogue, which represents a way for customers to integrate open source components and tools from Bitnami’s catalogue into their applications in a safe and secure way, by providing a curated repository of open source products that have been verified as stable and vulnerability-free.

For organisations at the start of their container projects, Tanzu Kubernetes Grid is being introduced as a ubiquitous container runtime, combining open source Kubernetes tooling, container images and registry and lifecycle management. Described by VMware as an evolution of its strategy with Enterprise PKS (which will remain as a separate offering), the aim is to make it easier to quickly start using Kubernetes in a consistent way across multiple environments, alongside existing VMware deployments.

Speaking of which, a new version of VMware’s flagship vSphere suite is also being introduced, and it includes a range of Kubernetes-friendly features. Previously teased as Project Pacific, vSphere 7 has been ‘fundamentally modernised’, according to VMware, and re-architected to put Kubernetes management at its heart.

VMs are not being forgotten about, of course – the goal is to enable VMware admins to easily run containers and VMs concurrently. vSphere 7 has also been optimised for simplified lifecycle management, allowing enterprises to manage hundreds or thousands of instances in less time, with fewer tools, including introducing REST and JSON-based APIs for automating lifecycle management tasks.

vSphere 7 also introduces greater security through remote attestation, where a trusted host is used to verify the integrity of other hosts within the network. Elsewhere, vMotion has been improved to allow for easier migration of large VMs with minimal disruption and the Distributed Resource Scheduler now runs every minute as opposed to every five minutes.

GPU virtualisation is now on offer too, thanks to the company’s recent acquisition Bitfusion, which is being touted as a particular benefit for those looking to run workloads using machine learning.

The aforementioned tools are also being rolled into VMware Cloud Foundation 4 with Tanzu, which includes vSAN 7 for managing virtualised storage.

“Kubernetes is still hard,” VMware CEO Pat Gelsinger said. “We’re democratising Kubernetes into the industry, with the most powerful platform, the most powerful infrastructure community across multiple clouds; This for us is an important day, not just for us, but for our customers, and for the industry.”

VMware Tanzu Application Catalog, Tanzu Mission Control and Tanzu Kubernetes Grid are available now, while VMware Cloud Foundation 4 and VMware vSphere 7 are scheduled to be available by the start of May.

HPE has added a suite of ‘as a service’ networking capabilities to its GreenLake portfolio designed to give customers the tools to accelerate 5G deployment.

The HP spinoff is aiming to extend its reach with telecoms firms and enterprises with its new 5G tools, which are designed to enhance existing 5G networks and ramp up the scale of infrastructure rollout.

For example, the company’s cloud-native and container-based software platform, dubbed 5G Core Stack, will provide telecoms firms with 5G tech at the core of their mobile networks.

This is to ensure that networks are embedded with 5G technology at their hearts, as well as on the edge, termed standalone 5G networks. This is against non-standalone 5G networks (5G networks running on 4G cores) which is how many operators run their networks today.

HPE hopes that mobile network operators and virtual network operators can adopt the technology at the core and the edge, and repackage these platforms to serve their own enterprise customers.

This is in addition to technology from HPE subsidiary Aruba, which has been used to launch services geared towards raising interoperability and integration between 5G and Wi-Fi 6 networks. These services are dubbed Air Slice and Air Pass.

“Openness is essential to the evolutionary nature of 5G and with HPE 5G Core Stack telcos can reduce operational costs, deploy features faster and keep themselves open to multiple networks and technologies while avoiding being locked-in to a single vendor approach,” said Phil Mottram, VP and GM for HPE’s communications and media solutions division.

“HPE has one of the broadest 5G portfolios in the market and is uniquely positioned to help telcos build an open multi-vendor 5G core, optimise the edge with vRAN, and deliver connectivity and new compute services to the enterprise using MEC and Wi-Fi 6.”

The as a service suite is all-encompassing and includes both hardware and supporting software, as well as cloud-native 5G functions such as Air Slice and Air Pass.

While HPE initially hopes to complement technology offered by the likes of networking giants Huawei and Ericsson, Mottram conceded that smaller customers may opt to replace all services offered by these firms with HPE technology.

This bold move to give network operators and enterprises an alternative has been made possible due to the open nature of 5G standards.

These standards were devised to break the stranglehold that existed previously, and allow enterprises to effectively mix and match elements of their 5G infrastructure in a way that wasn’t possible with previous generations of technology like 3G and 4G.

The technology, which includes the underlying 5G infrastructure, as well as support software at both the core and the edge, will be made available to customers on a consumption-based model through the company’s GreenLake portfolio.

With 50 conversations currently underway with prospective customers, HPE expects larger enterprises to adopt a sample of the 5G as a service portfolio, while smaller firms are more likely to adopt the full capabilities on an end-to-end basis.

In terms of cloud-native 5G functions that can work on top of the underlying 5G infrastructure, meanwhile, HPE launched a couple of enterprise-oriented services powered by Aruba’s 16.5 million hotspots.

Air Slice, for example, allows customers to carve up their networks into segments with dedicated functions to avoid crosstalk. Air Pass, meanwhile, gives individuals the capacity to join Wi-Fi networks without having to manually enter credentials.

Users’ identities are verified using their ties with another entity, such as a bank or a mobile network, in a similar way to one-click social media logins, used by Facebook or Sign in with Apple.

“As part of the foundational capability we’re talking about here is a shared data environment, so having a data model you can utilise across different functions in the capabilities and sharing it across different functions,” HPE’s chief technologist for communications and media, Chris Dando told IT Pro.

He added HPE was looking at how the end experience could be made seamless, with individuals not just tied to a physical SIM or looked on as being a phone number, but retaining their individuality. This is just one aspect of the suite of cloud-native 5G functionality the firm is hoping to build out.

“Those sorts of things are where we’ve led the way and are taking that too the next layer with regards to building out some of these core capabilities,” he continued.

“That’s going to become more and more important if you look to add on more device types for different use cases, as you get into IoT, and being able to identify groups of things as being part of a particular workflow or enterprise-type environment.”

Cisco, incidentally, outlined a similar Wi-Fi hopping technology at its flagship Cisco Live 2020 conference in January, which has already been deployed at the Fira de Barcelona.

It was expected to allow visitors to Mobile World Congress (MWC) hop seamlessly between 4G and 5G networks and the venue’s Wi-Fi networks, before the event was cancelled due to the global coronavirus outbreak.

Lloyds Banking Group has agreed on a five-year partnership with Google Cloud to further push its digital transformation work.

The deal comes under Lloyd’s £3 billion internal investment to upgrade its IT systems to compete in the increasingly digitised world of finance.

The UK-based bank, which serves 26 million customers, will collaborate with Google Cloud to help drive its overall cloud transformation programme and streamline its operations. The five-year «strategic collaboration» will see Lloyds use a number of Google Cloud services to help modernise and improve its customer experience.

This will also include Google Cloud experts on-site to enhance engineering processes, all in an effort to boost efficiency and offer innovative new services to the bank’s retail and commercial customers.

«Banking customers today expect secure access to their funds, without downtime, and delivered through the modern experiences they receive in other aspects of their lives,» says Thomas Kurian, Google Cloud’s CEO.

«We are proud to work with such a storied institution as Lloyds, which helped to create – and continues to redefine – the next generation of financial services.»

In 2018, Lloyds announced a partnership with Thought Machine, investing £11 million in upgrading its IT infrastructure. That came with announcements of 6,000 jobs cuts with a promise to create 8,000 more digital ones in the near future.

«As the UK’s largest digital bank we have made strong progress in transforming not only our systems but also how we work,» said Zaka Mian, group transformation director, Lloyds Banking Group.

«The size of our digital transformation is huge and Google Cloud’s capabilities will help drive this forward, increasing the pace of innovation, as well as bringing new services to our customers quickly and at scale.»

The Department of Defence (DoD) improperly evaluated a Microsoft storage price scenario, the US judge that halted work on the JEDI contract has said.

Federal Claims Judge Patricia Cambell-Smith added that because of this AWS is «likely to succeed» in its court challenge, according to Reuters.

The Joint Enterprise Defence Infrastructure project, awarded to Microsoft in October, is a $10 billion project to migrate the DoD’s computer systems. It has been marred in controversy from the start, with companies like IBM and Oracle attempting legal action over its single-vendor nature.

A large part of AWS’ complaint was that the bidding process was unfairly influenced by US President Donald Trump, but that was not mentioned in the Judge’s opinion.

Campbell-Smith wrote that Amazon «is likely to succeed on the merits of its argument that the DoD improperly evaluated» a Microsoft price scenario. She added that the scenario assessed by the Pentagon was not «technically feasible».

In response, Microsoft dismissed the decision as it focused on a «lone technical finding» of data storage under one price scenario out of six, according to Reuters. The company is confident it would eventually be able to move on with the project.

«The government makes clear that in their view Microsoft’s solution met the technical standards and performed as needed,» a Microsoft spokesperson said.

This could be seen as the first real tech-related argument of the JEDI saga, with large sections of the lawsuit focusing on the alleged interference of US President Donald Trump, who reportedly became concerned with the bidding when other cloud providers complained of being unfairly excluded.

Trump also has a long-running feud with Amazon CEO Jeff Bezos, which is seen as a possible reason for his alleged involvement. Former Pentagon Secretary James Mattis is thought to have said that Trump directed him to «screw Amazon» out of a chance to bid on the contract.