Building a hybrid cloud service means not only understanding the hardware and software tools available but more importantly, how a hybrid cloud deployment will meet the specific needs your company is addressing.
Henrik Nilsson, VP EMEA at Apptio told explained to Silicon: “Hybrid isn’t a one-size fits all operating model as the level of investment in the private cloud vs public cloud vs datacentres will vary in each organization, and often, vary between different departments in the same business.”
Nilsson continued: “The right mix of deployments is a balancing act, and CIOs are responsible for finding what can be a very tricky line. For example, in some businesses where there is more legacy investment, it might be a more strategic business decision to migrate only certain workloads over to the cloud which is not business critical, to keep technical debt down but slowly embrace more modern IT systems.”
The popularity of hybrid cloud deployments is clearly on an upward trajectory. The State of Hybrid Cloud 2018 from Microsoft concluded that 67% of respondents are now using or planning to deploy a hybrid cloud. Many of those hybrid users have made a move recently, including 54% of users in the past two years. Data storage (71%), backup and disaster recovery (69%) and cloud analytics (65%) are the main reasons businesses are building hybrid cloud environments.
How enterprises are approaching the design, choice of vendors, and hosting service providers include several components which influence the buying decisions that need to be made. The hybrid enterprise requires a new set of tools that can deliver the agility and scalability that today’s businesses need to innovate.
In their survey Voice of the Enterprise: Digital Pulse, Budgets and Outlook 2019, 451 Research revealed 62% of the IT professionals quizzed said they now use a hybrid IT environment with integrated on-premises systems and off-premises cloud/hosted resources; 17% said their IT environment is entirely off-premises, distributed across various SaaS, IaaS and PaaS clouds. Only 8% claimed clouds are not an essential part of their IT strategy.
A transformation is taking place; CTOs and CIOs are pushing on the boundaries of what is possible with their current private networks. DevOps is a good example here. Moving to a hybrid cloud infrastructure enables a business to use on-demand consumption-lead services that can be tailored to specific needs and outputs. Here, the hybrid cloud delivers IT as a service, which fosters higher levels of collaboration.
Building a hybrid cloud is, therefore, a multifaceted exercise that requires a deep understanding of not only the services available, but more importantly, how these can be used to deliver on each business goal, process optimization and the strategic agility in IT that all enterprises must be developing. Research from Flexera concluded on average; survey respondents are using 4.9 clouds across both public and private. Respondents are already running applications in 3.4 clouds and experimenting with 1.5 more.
Says Lee James, CTO, EMEA at Rackspace: “The key hurdles CIOs face when it comes to deploying hybrid cloud arise from the pressure to deliver both changes and results straight away. One of the major barriers to successful hybrid cloud deployment is buy-in and support. It’s easier to ensure a successful deployment if the transformation is being presented as something key stakeholders will be actively involved in, rather than a process they will be subjected to.”
James continued: “Cybersecurity is on the agenda of every C-suite, so balancing the migration to hybrid cloud infrastructure and securing assets sitting across cloud environments is a major focus for CIOs. They can do this by proactively understanding the ever-evolving threat landscape, developing a cloud security strategy from the outset, and engaging partners to help implement or strengthen network security.”
How your business approaches creating the hybrid cloud services it needs should be approached carefully and slowly. Starting small will enable you to see how the various components of the hybrid cloud infrastructure work together and the impact (both positive and negative) on your enterprise’s processes. Don’t choose mission-critical areas of our business to perform these tests upon. Robust and secure hybrid cloud deployments always work more efficiently when an iterative approach is used.
One of the key considerations your business must take is how the new technologies you want to deploy will affect your staff. Hybrid cloud deployment isn’t just an exercise in technology. Don’t forget the people that will be using these services. Ensure you fully consider their views and address concerns across your workforce to reduce the inevitable friction that the deployment will have.
The choice of vendor or vendors should also be carefully assessed. Your business’s existing IT infrastructure will need to integrate with the new cloud services. Here, legacy systems can have a significant impact on a hybrid cloud and its overall performance. Interoperability if your enterprise intends to buy from several vendors, should be carefully considered to ensure high levels of compatibility are delivered.
“Container-based cloud platforms are going to become dominant, especially with the rise of Kubernetes as the software of choice for many of these platforms,” Cloud Foundry Foundation’s CTO, Chip Childers explained. “Not all Kubernetes services are the same today, but I expect that the market pressures will cause both the maturation and eventually more consistency between the services.”
In their Enterprise Cloud Index 2018, Nutanix concluded: “Technical barriers limit application mobility across clouds today, keeping enterprises from fully embracing the promised flexibility of hybrid cloud technologies. These challenges likely account for at least part of the disparity between 91% of respondents ranking hybrid cloud as the ideal IT model and today’s relatively low hybrid cloud penetration levels of just 18%.”
Also, don’t overestimate the time and resources it will take to create the hybrid cloud infrastructure your business envisages. Often CTOs will underestimate how long it will take to move their existing business, which is likely still based on siloed data sources and legacy systems to an environment which can take advantage of the hybrid cloud landscape they want to deliver.
Often, CTOs will be developing more than one hybrid cloud. Leveraging the unique services and features on offer from the leading cloud providers is a route many are taking. What this means in practice is that CTOs will need to ensure they have a full end-to-end view of these networks, what their particular needs are and where – if at all – they overlap.
“The most common mistake businesses make when moving to hybrid cloud is getting distracted by what we call ‘the magpie effect’,” Rackspace’s Lee James told Silicon. “If a business is running an environment with more than one cloud provider, the sheer volume of product releases can have it chasing the latest and greatest, losing sight of key requirements. A partner is also key here, with the ability to manage multiple platforms, guide a business through all the options, and help build a hybrid cloud model that works for it.”
This view was reiterated by Greg Hanson, CTO EMEA, Informatica who advises: “A common mistake is having numerous cloud integration tools. I’ve seen businesses with several different types of integration solutions from different vendors. Every time they purchased a new cloud solution for a specific use case, they bought a different integration solution to cater to that specific use case. The result is a messy integration team skilled in an ensemble of different technologies but lacking in real expertise in any of them. Again, a more comprehensive cloud strategy that addresses these issues upfront helps to avoid this issue. Without this, companies will find it extremely difficult to offer that differentiated experience in real time to customers.”
Often the top of the agenda when any hybrid cloud service is being built is security. Cloud Foundry Foundation’s Chip Childers said: “One mistake that I have seen CIOs make is assuming that their private cloud is more secure than a public cloud could be. While it may feel better to ‘own the servers’, the underlying public cloud platforms have a likelihood of being just as secure if not more so, given the number of people focused on ensuring that security.”
Childers concluded: “What’s required is a shift in thinking about security overall. Even in private cloud environments, there should be a strong security program for the platform itself led by a team focused on the platform. With that in place, it becomes possible to secure any of the cloud environments with similar levels of protection.”
For CIOs and their CTO colleagues, building a hybrid cloud infrastructure is a long and complicated process. Understanding your current infrastructure, process management, data storage and usage, is vital to ensure the hybrid cloud services that is built will be fit for purpose.
Build or buy is often one of the first conversations CTOs, CIOs and of course, CFOs will have. Here, the cost imperative can quickly become the overriding factor when choosing vendors.
As Henrik Nilsson, VP EMEA at Apptio explained to Silicon, the hybrid cloud is not a cost-cutting exercise: “Many organizations see increased agility and scalability with the cloud. However, many businesses move to the cloud to reduce costs and are shocked when they get a bill back from AWS that is way beyond what was expected.
Nilsson continued: “Monitoring resource usage and spend across a hybrid cloud environment is a complex and time-consuming task. Billing varies wildly between vendors and, combined with inherent differences in the way public and private cloud are used, CIOs have a hard time making a like-for-like comparison on how much value each cloud service is bringing to the business. This variability and volatility in cloud usage mean CIOs are under much pressure to make the right, data-driven investment decisions about how they embrace hybrid.”
The current hybrid cloud market place is plagued with multiple vendors offering their services and platforms. As Informatica’s Greg Hanson explained to Silicon, the future will look quite different. “The future always tends to reflect the learnings of the past, and we will see business and the tech industry take action to reduce the impact of cloud sprawl and data fragmentation.”
Also, Hanson concluded: “I expect that organisations will take analyst’s advice to standardise on a limited set of technologies, such as an integration and data platform to enable a productive multi-cloud approach. This approach will reduce the issue of lock-in to a single supplier and will enable businesses to adopt cloud technology that provides a competitive advantage. In parallel, this strategy will reduce the potential to introduce complexity through fragmentation of data and multiple skill sets, which is the enemy of agile businesses.”
Constructing a hybrid cloud is not a panacea for rising network costs, improved efficiency or agile digital transformation, but a core component of the strategic planning that is needed to achieve these goals.
The future of your business’s IT infrastructure is some form of hybrid cloud deployment. How comprehensive this evolution will be is determined by your enterprise’s current needs and how the cloud will become an enabler of your company’s future innovation.
John Bidgood has over 25 years of international sales and service delivery experience gained by selling, developing and running network solutions that support, for example, a client’s hybrid cloud, IoT, mobility and security objectives. He is currently using this experience in his role a CTO, steering operational excellence and strategic initiatives for Systal Technology Solutions.
The overall pressure for CIOs and CTOs when deploying hybrid cloud is actually to manage a perfect balance of helping to support the business agility demands, managing IT economics, (cost-effectiveness), application performance, service levels and security controls (business continuity).
If we break this down further, we see added emphasise for CIOs and CTOs to manage.
The perception within their business that the hybrid cloud is easy to implement, and that it is able to almost completely replace an enterprise IT operation (its IT people and its infrastructure) – the hybrid cloud allows a shift of focus for the IT department to emphasise and meet agile business demands, but it doesn’t necessarily replace an IT team or the infrastructure they support.
The cost of interoperability and migration of workloads to public/private cloud services, even though the hybrid cloud itself might be a cost saver in comparison to a classic IT environment (i.e. a datacentre). However, there can be a considerable cost implication of integrating and migrating hybrid cloud into your existing infrastructure. This would be the extra cost of, for example, network connectivity or merely the time required to build and test all of the supporting infrastructures before migration can begin.
The more complex technical and overall support challenges of hybrid cloud interoperability. To a certain extent, you have to tailor your existing environment to accept the addition of hybrid cloud, as you cannot just buy in the services and expect them to work. Investment is needed in in-house skills and/or partners to establish and maintain a hybrid cloud operation.
Managing the services integration between classic IT, cloud and Internet Service Providers is the biggest challenge in supporting the long hybrid cloud ambitions of companies.
To provide the optimum operational IT solution, companies have to find the ideal compromise between in-house and external skills. Consideration also needs to be regarded for when things go wrong.
There is a big difference between purchasing Infrastructure as a Service (IaaS) and Platform as a Service (PaaS), in terms of both levels of integration and multi-site resilience. Business continuity planning is key. Many Service Managers are surprised when things go wrong, and a single point of failure is identified only after the event has occurred. A traditional business recovery service might not be needed in terms of physical hosting, but the process and tools are still required to test and implement a recovery situation.
This means that companies should be on the lookout for potential integration partners as well as individual partners who would perform the roles traditionally filled by a Cloud Service Provider (CSP) and an Internet Service Provider (ISP). This may even extend to companies performing Security, Incident and Event Management (SIEMs) to secure the overall end-to-end service.
Ideally, these potential integration partners are technology and vendor agnostic, and able to work on the company’s behalf to secure the best long-term solutions for your business. Furthermore, they may offer an opportunity to avoid margin on margin pricing by helping companies to source and procure ISP and CSP services directly, and even help to manage these on your behalf.
When building a hybrid cloud model, common mistakes include:
Thinking that you can perform a simple ‘lift and shift’ of workloads to cloud from existing in-house classic IT services. Many people underestimate the effort and time required to migrate from a classic IT service to the cloud.
It is underestimating the skills that you will need in-house to make it work. To achieve a hybrid cloud deployment companies ideally need to find the optimum IT model by integrating cloud, internet service and in-house solutions, while at the same time, continuing to provide the same day-to-day service for the operation of the business. This means that companies may need to expand with a mixture of potential new training, new hires and potential partners to implement and support.
Overestimating what you might get from a cloud vendor. A company like Amazon Web Services will not help you integrate their service into your business. You will have to perform the integration yourself or work with a partner to do it.
Incorrect planning of cloud services resulting in mismatched sizing requirements and service agreements. You need to plan the target environment you need and plan how to scale, and that environment can meet the SLAs that the business demands. Often, people ask for too much capacity and end up paying for something they are not using or asking for too little and underestimating their eventual costs. Sometimes public cloud can often end up costing more than retaining an existing operation or investing in their own private cloud.
Hybrid cloud is causing CIO’s to change the way that security is implemented, but it doesn’t change the underlying methodology. They still need to own the overall security policy and to make sure that the hybrid cloud is compliant with that policy, much like running an IT operation without a hybrid cloud.
Therefore, the essential requirement is still to confirm an overall corporate security policy and use this to outline what is needed to meet business compliance requirements.
However, how do you build trust around something you don’t own? When your assets are split over the public and private cloud, this makes security more complicated as you cannot claim full control over assets that you do not own.
This means a focus on the workloads, how users connect, and which digital platforms they are using to determine the security segregation model. For example, if a large number of users connect via third-party platforms or internet connections, it may be better to treat all users as ‘untrusted’ to preserve security.
User policies focus for ID/password management might mean the support for a single active directory with resulting integration and administrative requirements.
However, these users will rarely connect to a single cloud and will still probably connect to some traditional IT services in, for example, existing co-location or data centre locations. Therefore, it makes better sense to build a ‘trusted zone’ of interconnectivity between an IT operation’s hybrid cloud and the classic IT services in those data centres. While treating remote connectivity for users and their connected devices on the ‘outside’ as untrusted.
This will require the correct physical placement of, for example, firewalls. This is important as you do not want to impact the required capacity and performance from the network to interconnect with users, multiple clouds and classic IT systems.
We see DevOps being used in two ways:
Using resource in the cloud gives businesses a virtualised environment on an ‘evergreen’ public cloud or bare metal private cloud infrastructure, which does not impact the service on the production environment.
And as a pay-as-you-go model, it comes at a fraction of the cost of investing in and maintaining hardware and software to achieve similar results.
It’s for these reasons that some companies may find that they use all the cloud for DevOps, especially if they want to retain their existing infrastructure due to economic or service reasons.
To meet the scalability of hybrid cloud requirements, CIO’s are first confirming what hybrid cloud requirements they need to support application hosting.
This ‘application first’ hosting policy is based on the economic considerations combined with the service and security constraints dictated by the business-critical applications. Using this model, we can see that:
To join these cloud and data centre environments together, an end-to-end architecture is formed based on a service catalogue of desired features and drawing upon all the IT features that will be required to host the applications. This catalogue will include all the resources in the legacy data centre, the various cloud options, the network, the security mechanisms and the digital platforms required to access the applications.
These are the future hybrid cloud trends that we see: