The industry has, more or less, reached a broad consensus on what ‘cloud’ means. Broadly, it is a distributed service accessed over a network, run by an external company, and consumed as a service rather than an asset that we buy, own, and run ourselves.
The definition of hybrid IT, however, is still being worked out.
Much has been said about the benefits of adopting a hybrid approach to IT, but what exactly is it and why would you want to do it?
For this article, we examined the various components on a hybrid IT system and both the risk and rewards they pose. We asked how these elements should be combined, and what complexities this can introduce. We looked at the IT elements that make up a typical hybrid IT setup, and how this can inform your own designs.
Throughout our investigation, we discovered there is no one-size-fits-all solution. Making the right choice depends heavily on individual circumstances. Consultants everywhere rejoice: the correct answer is a resounding, ‘It depends’.
Hybrid IT is trickier to define because the meaning of ‘hybrid’ is quite broad. We understand that it’s a combination of multiple elements to create a new thing, but which elements are eligible, and what combinations count?
Let’s attempt to create some working definitions we can use.
On-site (often grammatically mangled as ‘on-premise’) refers to traditional IT systems that are owned and operated by the client, on their physical premises. They are responsible for just about everything to do with the ongoing operations of the IT systems, from power and cooling all the way to applications. The geek in the family understands this: they’re the one responsible for the TVs, internet, laptops, and phones in their house. These are all examples of on-site technology.
Co-location moves the IT systems to a professionally managed data centre, where the co-lo provider manages power, cooling, physical security and all the other data centre level issues for the customer, and they get paid to host the IT gear in their data centre. Examples include Equinix, NextDC, Global Switch and Digital Realty.
Managed services takes two main forms: running systems on the customer site as a service; or providing services to the customer from the supplier’s premises. Often combined with co-location capabilities, these providers offer a more vertically integrated set of services, usually based on some kind of virtual machine hosting environment plus a few more services. Examples include Interactive, Melbourne IT, ZettaGrid and many, many more.
Public cloud takes several forms, including infrastructure-as-a-service and software-as-a-service. The chief differences between cloud and a managed service are: the self-service consumption model; that the service is delivered over a network, often the public internet; and that it is a multi-tenant service offered to the general public. Examples include Amazon Web Services, Microsoft Azure and Google Cloud.
Private cloud is essentially a managed service with similar access and consumption methods as public cloud, but offered privately to a specific customer or small set of similar customers (sometimes called a ‘community cloud’). Examples include some managed service provider offerings and internal systems based on OpenStack or VMware’s vSphere suite, when sufficiently well automated.
Multi-cloud is the use of multiple cloud services, and generally means only public cloud. When a mix of public cloud and private cloud services are used, this is generally called hybrid cloud.
Finally, we come to hybrid IT, which is any combination of two or more of the above methods. With this, admittedly quite broad, definition we’d be hard pressed to find an organisation not doing hybrid IT.
Perhaps that is the point. There are many IT hybrids, each a function of the specific mix of services and their quantities chosen to support each business.
The other term growing in use is bimodal IT. Gartner has positioned this spin-off of hybrid as a way for enterprise IT teams, in particular, to manage a combination of traditional approaches and newer methods. For instance, this could mean moving to agile development methodology for front-end web design, while maintaining a more traditional waterfall approach for legacy software.
Bi-modal has become popular across major Australian IT shops, with the banks leading the way as they rapidly iterate new front-end user experiences using public cloud tools and open source technology while taking a more considered approach to their core systems, which will largely remain hosted on-site in monolithic systems built by SAP or Oracle.
The Department of Human Services is the latest Australian agency to tout the benefits of bimodal. DHS is hoping to revamp some of its web-based frontends with Angular, the open source framework developed by Google, while it also moves forward more cautiously with the $1.5 billion core systems upgrade project known as the welfare payments infrastructure transformation (WPIT).
Everyone is doing it
“There is no one true way,” says Dayle Wilson, chief operating officer of Brennan IT. “It is about using the right combination of IT that’s best suited to a particular business.”
Which all seems rather compli-ca-ted. Why would you do this? Because, it turns out, it is very difficult not to.
“Hybrid IT is out there more widely than people think,” says Paul Vinton, chief executive of Adelaide-based MSP Vintek. “It’s engaged in a much bigger way than we even realise – 90 percent of customers at least.”
Just running a simple business means performing tasks like invoicing customers, doing accounting, paying your taxes, and so on. Unless you do all of your banking by physically walking into a branch, you’re using hybrid IT by using the bank’s internet banking systems to pay suppliers and check customers have paid you. Dealing with the Australian Tax Office increasingly means doing things online via the Business Portal or using myGov.
Not having a hybrid approach means either doing everything in-house – almost impossible when one of the organisations you need to deal with is the government – or getting someone else to do everything, which means ceding all possible IT innovation and competitive advantage to someone else. Neither extreme seems reasonable on its face.
Given this situation, we find ourselves essentially forced into a hybrid IT approach where we make use of some internal systems, some hosted and run by others, and some a combination of co-delivery where we and another organisation share part of the burden.
“We can shift assets from being on-site, wholly owned systems to cloud-like services by slipping a contracting layer in-between,” says Joep Piscaer, CTO at Dutch managed service provider OGD ict-diensten. Many resellers are adding these kinds of cloud-like services to their offerings to offset the decline in customers purchasing systems outright.
Ideally, we want to choose how much of any given element in our hybrid IT mix we want to use, rather than having it dictated to us. Admittedly, this is difficult when it’s a government-mandated system, but for everything else, it would be wise to understand what the benefits and risks are for using the system, and to use it, or avoid it, as appropriate.
Risk and reward
Figuring out the right mix means understanding the benefits and risks of each style of IT.
“Having cloud in the mix helps to avoid significant up-front capital expense and lets businesses pay as you go for usage,” says Brennan IT’s Wilson. “But most companies have invested in IT before the cloud, and it often makes sense to retain parts of that traditional IT. They want to get the most out of those existing investments.”
The customer organisation may also have regulatory barriers to placing data in particular locations, such as offshore data centres or cloud regions, which may constrain their choices to more localised options. Organisations with offshore customers need to keep in mind the requirements of new legislation such as the EU’s General Data Protection Regulation (GDPR).
“Data sovereignty, compliance issues, and how data is replicated are key considerations,” says Matt Leib, field CTO, central region at Connection Enterprise Solutions, based in Florida. “Whether that’s a hybrid configuration or a true public environment, where your data is located and how secure it is are important aspects to evaluate.”
You should also consider what happens if the customers change its mind about a service and wants to move to a different one. “The issue is about getting your data out,” says Vinton. Data migrations are complex projects at the best of times, so it’s often wise to understand the barriers to exit before committing a lot of data to a particular choice.
Wilson agrees. “Adopting cloud services can be a costly and painful exercise due to the difficulty in getting your data in, and potentially out, of that cloud provider,” he says.
“Integrators can help you navigate this process,” Vinton adds. With the ever-increasing array of options, it’s difficult to maintain in-house expertise in everything. Judicious use of external advisors can be a wise choice, but ensure customers are advised to seek a diversity of opinions rather than risk having IT strategy dictated by a single voice.
No single right way
The way in which hybrid IT is adopted by an enterprise should be a function of the services their organisation needs, rather than some dictat from IT. While there is no one right way to adopt hybrid IT, there are plenty of incorrect ways.
Doing everything in-house is unlikely to be a wise choice for most organisations, given the interdependent nature of the modern business world. There are substantial costs, not least in ongoing maintenance, of businesses attempting to do everything themselves.
Getting rid of all IT capability would also be a mistake, and one that many organisations have made. By outsourcing all of their IT staff to vendors, this leaves no one who knows how to check the vendors are providing sufficient value.
“Hybrid IT helps you to expand your business faster than you could without it,” says Vinton.
The challenge for every organisation is to understand how IT contributes to value creation of the business as a whole, and how to maximise it. For some, this could mean outsourcing to a low-cost provider to achieve ongoing economies-of-scale as that provider relentlessly focuses on efficiency, leaving the client to concentrate on higher-value activities.
For others, it could mean keeping a substantial chunk of IT capability in-house so as to out-manoeuver competitors with their superior ability to use technology to fulfil the organisation’s purpose.
“Organisations should choose a partner that can deliver a consistent experience in managing the full spectrum: both on-premises and in the cloud,” says Brennan IT’s Wilson.
The organisations that make the best decisions about this mix are the ones that will thrive, and those who allocate resources poorly will suffer. This is not a profound or novel observation – or rather, it shouldn’t be – because this is the essence of good business.
The key is for customers to make a well informed choice of the hybrid IT mix that is right for their organisation.
A quick introduction to the ways in which traditional tier-one vendors are tackling hybrid IT
Dell EMC sells a range of converged and hyper-converged systems as part of a hybrid IT solution, with various software options layered on top. As the parent company of VMware, Dell EMC provides VMware as one option, but also has a strong partnership with Microsoft, including Azure and Azure Stack.
Composable infrastructure is the HPE term for combining its own range of traditional, converged, and hyper-converged systems together depending on the solution required, but aims to go one step further using software to create dynamically configurable pools of resources, much like the public cloud.
VMware on AWS is its latest foray into the cloud game, which means you can span a VMware environment from on-site or co-lo into AWS. It uses NSX software-based networking to provide connectivity between sites, and means you can run systems in the cloud the same as you would run them on-site in a vSphere cluster.
Cisco has a portfolio of products for different hybrid IT needs. There’s CloudCenter (formerly CliQr) for deploying to multiple cloud environments, a range of on-site infrastructure based on the flagship UCS and ACI platforms, and the OpenStack-based Metacloud for private cloud deployments.
Lenovo partners with various software providers while supplying the underlying infrastructure, usually as a reference architecture depending on the solution required. Partners include Red Hat, Suse, and VMware.