Kathy Gibson reports from Gartner Symposium in Cape Town – The disappearing data centre is one of the main strategic imperatives that CIOs need to keep in mind as they move forward.
“The physical data is still there, we’ve still got them, but the boundary of the data centre has blurred,” says Gartner analyst Philip Dawson “You have to interconnect those data centres and make sure they are working together.”
This interconnectedness is the second strategic imperative for CIOs, followed by containers and application streams.
“We have bimodal, and the biggest aspect of this is changing the application to be small, not monolithic spheres that scale up. These little applications or bubbles all talk to each other, work together – but they last for a much shorter time scale, and they are difficult to measure.”
On a tactical front, CIOs have to adapt to business-driven IT while ensuring that they deliver IT as a service. “IT is about delivering services, not infrastructure,” says Dawson.
Stranded capacity – capacity you cannot use – is something that could be used to give IT the turbo boost it needs when it needs it.
The Internet of Things is another tactical issue that CIOs have to consider, and plan for.
On the organisational front, remote device (and thing) management is going to be important.
CIOs also need to manage the micro and edge computing environments – having first determined where the edge is.
“Finally, what about skills?” asks Dawson. “Skills has always been an issue; having the right skills and people, and getting them to work together, is always a challenge.”
Strategic issue – disappearing data centres
The amount of computing taking place in the on-premise data centre is declining while the amount that happens off-premise or in the cloud is increasing, and they should cross over in about 2020.
“You have to start planning for this disruption,” says Dawson. CIOs also need to differentiate between off-premise, hosted services and cloud.
“But the work of IT will continue,” he says. Hybrid environments mean that there is more complexity, and staff still need to be skilled up.
If companies want to move from one cloud to another, they have to consider that migration will not be easy or cheap. The same goes for moving to a cloud provider in the first place.
Dawson warns that unused space should be shut down or repurposed, or it will remain as a cost.
In addition CIOs need to make sure they modernise their processes before moving them to the cloud.
Strategic issue – Interconnect fabrics
CIOs will be dependent on different parties for these interconnect fabrics, says Dawson. So organisations have to invest in software-defined networking (SDN), switch fabrics and security services to make sure that their traffic is secure and reliable.
CIOs will also have to take care of integrated orchestration to make sure that services are delivering from the right place, for the right price, from the right platform. They have to centrally operate, orchestrate and automate to allow configuration from one place.
Strategic issue – containers and application streams
Containers are not new, but they are now available with or without hypervisors, Dawson says.
Containers enable new forms of application and agile development, and can be deployed on virtual machines, especially in cloud IaaS.
Containers are set to become the primary tool for application developers, Dawson says, and they are particularly fluent in the DevOps environment.
“This blends with the Internet of Things (IoT) and with the consumerisation of IT,” he says.
“So you need to support containers on top of existing systems, they will not only be operating systems but hypervisors. And this will give us more portability – this will become mainstream within the next couple of years.”
Virtual machines decreased deployment times to minutes or hours; containers reduce this further to milliseconds or seconds – and the applications themselves could have a lifespan of just minutes.
Streams take this even further; they are deployed in milliseconds, for a lifespan of milliseconds, which will exacerbate the challenges.
CIOs will have to find new ways of measuring the use of containers.
Tactical issue – business-driven IT
CIOs will need to do IT in new ways, extending it to the business, says Dawson.
The issue is that business is not willing to wait for IT, while cloud services are proliferating. This means that implementing standards is almost impossible for IT – and costs escalate.
The choices facing CIOs are whether IT is setting the standards – it isn’t; is IT focusing on Mode 1 only – the answer is also no; and can IT reset the playing field – the answer here is yes.
Tactical issue – data centre as a service
This is about the application rather than the infrastructure. The decision on which infrastructure to use – on-premise, cloud or hybrid – is all around latency, reputation security, performance, service continuity, mode 1 versus mode 2, data protection, compliance and disaster recovery.
Some of these functions will always remain an internal function, so CIOs need to be aware of compliance and regulatory issues.
Service delivery, however, can come from anywhere, Dawson says. “It’s not about ‘my data centre’ anymore you are dealing with different aspects, with everything as a service. It’s about delivering compute resources to the business, in the best way possible.
“So there is a primary data centre, with the data services being just about anything. Service delivery could come from anywhere, but the CIO needs to be aware of the costs associated with them.”
Tactical issue – stranded capacity
Stranded capacity refers to resources that are unutilised or underutilised.
Currently, companies have about 28% of ghost servers, 40% under-provisioned racks, 32% additional server performance and 40% storage over capacity. “That’s an over-investment of about one-third.”
Cost optimisation is about getting the best computer per kW and increased asset densities, Dawson explains.
Tactical issue – IoT platforms
IT generally happens close to the user – but IoT goes out into the wider world.
“What you have to look at with IoT is how quickly the microservices are using these functions; and how scalable these resources are,” Dawson says.
“IoT tends to carry a high cost of integration, while the sheer scale means you have to plan for major investment.”
There is no end-to-end solution between the different IoT providers, Dawson adds, so there will always be multiple points of integration for these systems.
IoT will impact networking standards; computing platforms and solutions; consortiums; protocol standards; and the ecosystem.
“This all adds to the complexity of managing the IoT strategy,” Dawson says.
Organisational issue – remote device (thing) management
To monitor and manage this mesh of devices involves the installation, maintenance and operation, and retirement of devices.
“These things are very physical, in the field, devices. Do you get the user to manage them, or can you do it remotely?” Dawson asks.
Physical maintenance of the hardware is an issue that needs to be considered, and it will be an ongoing step process.
IoT is going to be on a massive scale, with massive numbers of things coming into the system. And, because there is no dominant IoT ecosystem, multi-vendor collaboration is going to be necessary.
Organisational issue – micro and edge computing environments
This needs to be defined, first of all, and this will involve intercolocation networking.
“You have to look at what is your IT edge compared to what is your IoT or device edge,” Dawson says. “There will be a difference in site-specific or service-specific computing.”
Delivering services to consumer wherever they are will probably be more important than looking after resources in a site-specific manner. “The expansion of the personal identity will be an explosion for edge computing.”
Organisational issue – new roles for IT
“Skills becomes a massive issue,” Dawson point out. “Skills drive your core IT – but you need new skills to drive new IT directions like edge, IoT, containers and others.”
The issues to deal with include having to consider new ways of doing IT. These include IoT architecture; cloud sprawl; strategy architecture; capacity and resource optimisation; broker and provider management; and performance and end-to-end management.
“How do you do that?” asks Dawson. “You need to focus on not just the horizontal, but vertical skills sets.”
This means people need to work together rather than in silos. “To embrace change in the skills model is vital – and it’s the key to success,” Dawson says.