1. The Dawn of Hybrid Apps
A hybrid app is an application that’s more or less an aggregate of various microservices. It’s a way to build best-of-breed applications and differentiated services that can help organizations stand out from competitors.
Imagine the ability to leverage data services from one cloud provider and machine learning or analytics from another. Or perhaps you have a business partner who has built services that want to run adjacent to your data. For organizations, innovation is no longer limited by what can be offered from a single cloud provider. A platform can bring all these heterogeneous services together and run these in ways you just can’t do anywhere else.
Meetings with IT leaders throughout 2019 showed me there’s an appetite for the flexibility to build applications that combine the best native cloud services and open source technologies.
2. Solutions at the Edge Come to Life
Organizations are looking for technology partners to help solve problems and accelerate investments at the edge. I foresee holistic edge solutions coming to market that will:
- Decrease or neutralize edge computing costs.
- Consolidate the amount of infrastructure needed at edge locations.
- Allow enterprises to deliver new technologies to remote locations entirely in software, improving business velocity and agility
An example is a single hardware appliance that provides software-defined WAN, while also running a few applications in isolation. These will reduce costs by consolidating infrastructure and improve network performance.
3. Specialized Hardware as a Shared Pool
When applications require specialized hardware such as an FPGA or GPU, enterprises often have to dedicate servers to those applications. In 2020, I expect to see the emergence of remotely connecting to specialized hardware as key design principle.
When you combine hyperconverged infrastructure with solutions, such as Bitfusion, that allow you to connect applications to remote GPUs or FPGAs over Ethernet, you can take a modular approach to IT infrastructure. You can also forgo the need to maintain hundreds of different server builds to satisfy myriad application requirements. Going forward, applications can remotely connect to that specialized hardware at the time they need it.
4. The First Steps Toward Intrinsic Security
Malware is highly sophisticated and constantly evolving. To combat it, an organization’s security posture must be more dynamic than the highly dynamic threats they face.
Security should be intrinsic to IT infrastructure. The challenge is security systems and processes are far too critical to disrupt all at once across an organization, and networking and security professionals are risk-averse by nature, as they should be.
Instead, enterprises will slowly evolve toward intrinsic security models, starting with a single application or new project. We are reaching a point where network and security policy, along with firewall rules, are simply attributes of an application. This means that rules and policy are dynamically created at the time an application is initialized and can be destroyed when the application is retired.
The ways in which we protect applications and data are evolving, and it’s hard to find IT professionals with these skillsets. The best approach is to start small. Pick a new Kubernetes project, for example, and go greenfield with modern, software-defined network and security solutions. Your team can build knowledge organically and scale these new methodologies across your enterprise over time.
“We are reaching a point where network and security policy, along with firewall rules, are simply attributes of an application.”
5. Big Ideas for Small Devices
I think we’ll see expanded use cases for smaller devices, such as the Raspberry Pi 4, in the enterprise. It’s less than $100, but it’s a powerful device with quite a bit of flexibility. Bringing virtualization and other enterprise technologies to these devices, for even more security and isolation at the edge, could introduce new opportunities to innovate that we’re not yet considering.
6. Machine Learning for the 99 Percent
Thus far, machine learning (ML) required a high amount of data science and sophistication that’s a bit out of reach for a mainstream organization. For example, a small business can’t hire its own team of data scientists.
Soon, we’ll see an increasing number of turnkey ML services from cloud providers and open source communities. It will become far more accessible, so businesses can apply ML models without a high degree of expertise. The providers that succeed in making ML assessible for organizations without data science expertise — the 99 percent — will be the dominant players going forward.
7. Further Cloud Disaggregation
In some cases, it’s impractical to move data to the public cloud. So, why not move the cloud service to where the data is?
We’ve already started to see cloud services run independently of cloud data centers, such as running Amazon Relational Database Service (RDS) on-premises. I expect to see far more examples of this in 2020 and beyond. The physical location of where a service runs ultimately becomes an implementation detail that is dictated by business requirements and enforced via policy.
8. Shared Services Platforms
In certain industries, bringing business partners on board is often contingent upon the business partner bringing their own hardware and software into the facility. A retailer’s business partners might have to bring their own desktop computers with their own software solution into the store, which the IT team would then have to connect through the firewall.
The problem with a business model where partners need their own equipment, is that it limits innovation to partners that have the capital to implement their solution. I expect to see the beginnings of platforms that can live in a single location and run services from multiple business partners, with isolation provided via virtualization software.
Multi-tenant shared services platforms at the edge can be a big deal, as it can democratize innovation. These can allow organizations to expand potential business partnerships to anyone with a great idea, regardless of whether they have the capital to implement their solution.
Platforms that allow enterprises to not only onboard new business partners, but also potentially create new revenue streams from leasing compute capacity and data access at the edge, will further differentiate winners and losers in industries like retail and manufacturing.