This article is part of our Essential Guide: Essential guide to optimising hybrid IT infrastructure

Hybrid cloud strategy offers best of both worlds

More and more businesses are adopting a hybrid cloud strategy for their datacentre needs. Computer Weekly looks at what’s on offer

According to an IDC report commissioned by Cisco, published in September 2016, cloud adoption continues to grow, but few companies have mature cloud strategies in place.

Hybrid cloud is a significant part of the mix, with 73% of organisations surveyed saying they are pursuing a hybrid cloud strategy, subscribing to multiple external cloud providers and using a mix of cloud and dedicated (on-premise) IT resources.

The study also showed that 78% of organisations are using or planning to implement some form of cloud – a 61% increase from the previous year. Yet only 31% of IT decision makers said they had repeatable, optimised or managed strategies. Although this is an increase from 2015, there is still room for improvement.

Automated management

Automated datacentre management, along with machine learning to understand the events that lead to degradations in system performance, are set to drive datacentre strategies in the coming years.

Analyst Gartner expects that, by 2020, the operating costs of enterprise datacentres will have been reduced by 30%. As a result, the analyst firm suggests that to attract buyers, technology product management leaders at datacentre infrastructure providers should differentiate themselves by providing targeted automation and interoperability software to reduce the cost of operating their products.

According to Gartner’s IT Key Metrics Data 2016 report, system costs make up between 18% and 24% (in the case of Linux and Windows servers, respectively), whereas the cost of software and administration ranges between 68% (Windows) and 72% (Linux).

Reducing operating costs applies equally to in-house datacentre operations, and the latest thinking from the large server, virtualisation and cloud providers is to offer businesses a way to bridge their own datacentres with public infrastructure as a service (IaaS) using a hybrid cloud that can move workloads automatically across public and private cloud infrastructures.

Federated cloud management

An organisation is unlikely to choose just one public and one private cloud. Often, several public and private cloud services need to work together to provide the IT environment the business needs. In many instances, the IT department does not have overall responsibility for the cloud services that business users procure and deploy.

And, as is often the case in IT, getting the constituent parts to work together is a complex and costly matter.

Buying off-the-shelf products can go some way to simplifying this integration. A report from Principal Technologies commissioned by EMC said organisations could gain the savings and advantages of a cloud environment sooner. “By taking 1.3 years longer to build your own solution, your organisation runs the risk of not realising the benefits and savings offered by a hybrid cloud consumption model during that period. These include $1.76m savings based on our hypothetical organisation,” the report says.

Dell EMC offers Virtustream and Pivotal Cloud Foundry to support both traditional and cloud-native workloads.

In October 2016, VMware said it was working with Amazon Web Services (AWS) to bridge private and public cloud infrastructures. At the time, Mark Lohmeyer, vice-president for products at its cloud platform business unit, said in a company blog: “VMware Cloud on AWS will be powered by VMware Cloud Foundation, a unified SDDC [software-defined datacentre] platform that integrates VMware vSphere, VMware Virtual SAN and NSX virtualisation technologies, and will provide access to the broad range of AWS services, together with the functionality, elasticity, and security customers expect from the AWS cloud.”

Technology tie-up

At the VMworld Europe conference in Barcelona last October, the company showed how the technology tie-up could be used. Lohmeyer demonstrated how VMware could be used with AWS to move workloads dynamically between the private and public cloud. “We leverage AWS Direct Connect to use vMotion to migrate private cloud VMware workloads onto AWS,” he says.

Read more about datacentre management

Demand for higher levels of availability will lead to increased use of automation. Over time, many IT roles will be taken over by intelligent algorithms.

HPE has developed a concept computing architecture which it says will power future generations of applications. We find out how it will change IT.

Also at the event, CEO Pat Gelsinger announced VMware Cross Cloud, which he said solves the challenge of using any cloud for any application. One VMware customer facing this challenge is betting company Ladbrokes, which has built automated provisioning on its VMware system and created an abstraction layer to enable it to run on-premise or in the cloud.

Payment processing company Worldpay, another VMware customer, is looking to use software-defined IT to enable it to jump-start its strategy for software-defined automation. Gelsinger added: “Users want freedom to choose any cloud, and you are responsible for control. This is the world of a hybrid environment for decades to come – this is the software-defined datacentre.”

The traditional server companies appear very much on the on-premise side of hybrid clouds. Nevertheless, automation has an important role to play in making on-premise hardware as easy to manage as cloud-based infrastructure.

HPE’s The Machine

Hewlett Packard Enterprise (HPE), for instance, has a long-term research project dubbed The Machine, which is a server architecture where storage and memory combine to form a so-called memory fabric. Kirk Bresniker, chief architect at HPE, says: “We want to change the basic economics, going from relatively scarce amounts of memory.”

In the past, says Bresniker, hardware architects worked out elaborate schemes to speed up access to relatively slow storage devices using cache memory. “The only time caches work is when we hit the cache,” he says. In HPE’s vision for The Machine, applications would have direct access to vast amounts of memory.

Some technology from this project is finding its way into next-generation Hewlett Packard Enterprise products. For instance, the latest HPE SAN networking components support embedded intelligence called HPE Smart SAN. This technology is embedded into HPE StoreFabric switches and directors and enables fully automated SAN orchestration directly from within HPE 3PAR StoreServ Storage arrays.

According to HPE, Smart SAN cuts provisioning and change management time by 90% and reduces the risk of human error, assuring resilience and even enabling proactive risk mitigation when network changes are made. HPE also provides Helion, a secure hybrid cloud with OneView management, which, it claims, can help organisations speed up networking provisioning times by 10 times while reducing configuration errors and operational costs by 50%.

On Microsoft operations management suite Azure, the Microsoft Operations Manager is used for hybrid cloud management. This was recently extended with Service Map, a tool to provide real-time dependency discovery and mapping.

Nick Burling, principal programme manager on Microsoft’s Enterprise Cloud management team, wrote in a blog post that the tool helps IT administrators to eliminate the guesswork of problem isolation, to identify surprise connections and broken links in the corporate IT environment, and to perform Azure migrations knowing that critical systems and endpoints will not be left behind. As part of the cloud management capabilities in Operations Management Suite, wrote Burling, Service Map helps IT administrators track dependencies in their hybrid cloud environment, making it easier to manage the complexity of multiple clouds.

Businesses are demanding more from IT, and among the areas CIOs need to consider is what happens when back-office IT becomes customer-facing. This is one of the challenges that global ticket distribution company Amadeus is facing. Previously, its systems were geared towards supporting travel agents’ business-to-business operations. Now that people book flights and hotels directly, the company’s IT operations need to meet higher user expectations. If the service is slow or unresponsive, a consumer will find an alternative site through which to book their travel arrangements.

During a roundtable discussion at VMworld in October, Wolfgang Krips, executive vice-president for global operations at Amadeus, said: “We are automating to take people out of the system. In the future, my people will become automation engineers.”

This is a big transformation. IT has to think about how it delivers services, and applications must adhere to newer paradigms. “You have to find ways to operate in a hybrid world,” says Krips. “Today, if we have a problem, the supplier fixes it. In a cloud-based world, your applications have to deal with the problem.”

One of the areas Amadeus is exploring is using artificial intelligence and best practices derived from fraud detection techniques in financial services to flag up potential problem areas based on deep analysis of telemetry data from running IT systems.

Read more on Virtualisation management strategy