What are the key trends that will change the way IT pros work over the next five years, and how will server and data centre technology evolve? Gartner vice-president David J Cappuccio’s highlighted 10 you social and technology trend you need to consider, and we captured them all live at the Gartner Infrastructure, Operations & Data Center Summit in Sydney as part of our World Of Servers coverage.
1. Software-Defined Networks
Software-defined networks (SDN) really kicked off 18 months ago, Cappuccio says. “Virtually all vendors now have some variation on software-defined networks.” The SDN shifts control to a central point from devices, which makes management easier. “As networks get more complex, the amount of work it takes continues to grow. With SDN, that changes the model.”
The time-saving benefits can be dramatic. “It takes sometimes days to get network configuraton done and it’s rife with errors. SDN eliminates that.” Changing traffic patterns in data centres also matter; data often flows between separate systems, so optimising the network for workflows is vital. It can also improve disaster recovery, letting workloads shift more easily if a potential disruption is anticipated.
There will often be resistance from network engineers. “Some hunker down and say ‘Don’t touch my network’. It can be a major issue.’
2. Bigger Data
Big data demand is “becoming pervasive in most companies”. By 2015, Gartner expects one million big data vacancies, only a third of which will be filled. The benefits are obvious (we discuss them a lot here at Lifehacker) but dealing with data remains challenging. “There are two sides to the equation — how do I analyse it and how do I manage it?” The latter can be difficult and disruptive.
“On the data centre side, we’re seeing a lot of focus on deduplication and automated tiering to extend the life of existing data. From the analytics side, we’re seeing a lot of work with analytics providers, cloud providers and storage providers. Big data is still driving a lot of companies to change their environments.”
3. Hybrid Cloud Services
“The word cloud computing is pervasive, but depending on who you talk to it means something completely different. We’re seeing a move away from talking about the cloud generally and thinking about specific cloud services.” Cappuccio draws a familiar parallel with payroll services. “It’s often about finding a single provider to do a single service, not trying to solve everything.”
The benefits can be obvious: making IT more productive and more creative, while potentially offloading unwanted workloads. “The downside of all this is the cascade effects. If I bring a secondary service into my hybrid cloud, it’s faster but the end user experience is still my responsibility.” Managing multiple cloud providers can become time-consuming. “The second problem for IT is my competition now becomes other cloud providers.” Balancing those issues is a major management challenge, but the benefits are compelling. “We’re seeing private clouds improving the ability of IT to bring in external services seamlessly.”
4. Integrated Systems
Integrated systems were initially driven by vendors to build a single ‘pizza box’ service. That’s easier to manage for both vendors and customers, and can create lock-in. “Most customers don’t like one vendor to rule them all. The idea of one throat to choke is really nice, but you only end up with one throat, so you have to massage it.” However, scalability and cost continue to drive those strategies in some data centres. “Everything becomes part of the stack, but it doesn’t have to be one vendor. The environment needs to be flexible enough to add nodes as you need them.” This approach is often used for single-purpose devivces (such as encryption). The concept has now “grown up” and is hitting the data centre, Cappuccio says. “I’m starting to hear ‘it’s not best-of-breed I want, it’s best of brand’. Ease of use and management is becoming a more dominant trend.”
This change happens over time, moving from a static world to a fabric-enabled dynamic environment. “Everything is designed around the workload rather than the technology.”
“Think about the server marketplace. If I’d asked you five years ago if you’d buy a white box server for critical systems, I’d be laughed out of the room.” Today, that’s often the case, especially if buying in services. “The specific device is no longer as important as the function it provides. We’re seeing a disaggregation of equipment from services.”
Gartner links the growth in mobile devices with the idea of a multi-channel world. “Users expect access to everything everywhere on any kind of device. The idea of IT dictating the kind of device people can use is gone. BYOD is driving IT to make some dramatic changes. Depending on your perspective, it will get even better and get even worse.”
“We’re starting to see a change in application development.” Applets which solve a single problem are becoming more common. This can add complexity, but it enables fast rollout. Users are increasingly demanding. “Apps delivery and apps refresh cycles will have to be sped up dramatically.” Delivering applets makes version control and updating easier, and avoids IT over-building outside of core functionality. “If you do it right, it becomes a much more controllable environment and a much more agile environment.”
6. The ‘Internet Of Things’
Mobility leads naturally into the ‘internet of things’: smart shoes, embedded devices. “This market has been growing outside the purview of IT. Everything gets a radio now.” Monitoring those systems is a challenge, but right now it often happens outside IT, in an area Cappuccio calls ‘operational technology’. This is already active in environments like hospitals, for tracking of equipment such as wheelchairs, or placing simple monitoring equipment on beds to monitor temperature and movement. “The internet of things is not a single technology, it’s a concept.” This lets apps be sensitive to context, which in turn can make them more relevant and effective. “It changes the way we think about applications.”
7. Open Compute
The Open Compute project, kicked off by Facebook and then developed in partnership with Rackspace, could have a major impact on data centre design, Cappuccio argues. Using a common open architecture has the potential to make equipment purchase easier. More than 60 vendors are now involved with the project.
Cappuccio singles out current rack design — originally designed for telecommunications gear — as potentially doing well from the update. The design is lighter, holds more trays, and integrates power and communications. “I’ll be able to plug in components as I need them, rather than physical servers. I can start building environments as I need them. And when a new processor comes out, I don’t have to get rid of my whole server: I just get rid of the processor. It makes the environment much easier to upgrade and to manage. We think as a concept, this is something to watch. Don’t jump on it today because you won’t find too many vendors selling product, but it’s definitely worth watching.”
8. Virtual Data Centres
The virtual data centre, with everything virtualised, doesn’t require the same change to equipment. But the benefits don’t always become apparent. Even in virtualised data centres, utilisation often peaks around 24 per cent. The secret is to focus on workloads, rather than just the number of virtual images, Cappuccio suggests. “Find the workloads you need to get that kind of peak. It might be 2; it might be 40. But you can get a lot more density and capability and function out of existing devices than we have in the past.” Virtualising the network makes moving workloads easier, whether that’s for maximising utilisation or business continuity. “Resources can be distributed across data centres and geographies, and workloads can go to ‘just good enough’ resources.” IT has often over-committed to avoid outages, but that becomes costly.
That creates staff challenges, since most environments still have vertical skill silos. In a virtualised environment, identifying the source of performance problems requires more collaboration. Good virtualisation staff often get bored and seek out a new challenge. Knowing how elements tie together will be increasingly vital. Using highly skilled people from one area in other roles can help meet that challenge, Cappuccio suggests. That makes staff more wedded to the organisation, and ensures everyone is examining IT horizontally. “Embrace the collective.” That can mean asking other staff for support rather than using a formal help desk. That can be costly with senior staff, “but in many cases they’re doing it anyway.”
9. IT Demand
Everything is growing: server workloads are up 10 per cent, bandwidth demands rise 35 per cent, power costs rise 20 per cent each year. However, annual IT budgets are growing about 3 per cent each year, and have declined in real terms. “Doing less with more will continue.”
That applies heavily in the data centre. CIOs increasingly are focusing on data centre efficiency with equipment, rather than people and processes. That happens for two reasons: the equipment has already been paid for, and it helps reduce the need to expand. “That focus on vertical density is realising some significant savings. But it takes a mindset change within IT to do that.”
10. Organisational Disruption
How does this disrupt the world? By 2014, Gartner predicts 30 per cent of organisations using SaaS will move back on premises. “Just because you’re doing it now doesn’t mean you’ll be doing it forever.” Support is demanded in real time. The change cycle is becoming faster. Being process-bound won’t help solve problems. “End users are driving IT. Get used to it.”
“The biggest impact is in the changes to the skill sets. All these things are going to impact us.”
With all these sources of potential disruption, a clear strategy is required. “We need to learn how to live with it and how to deal with it. If you don’t, you’re trying to put square pegs in round holes. You can’t fight it; you can’t lock down and say we won’t make any changes. If we try to do that, customers will find a way around it.”
So the five key pieces of advice? Prepare for organisational disruptions. Start investigating software defined networks. Evaluate your needs for hybrid cloud. Prepare to support an ever broader range of devices. And get rid of complexity where you can. It’s an ongoing battle, but that’s what it takes to be an IT pro.
Lifehacker’s World Of Servers sees me travelling to conferences around Australia and around the globe in search of fresh insights into how server and infrastructure deployment is changing in the cloud era. This week, I’m in Sydney for the Gartner Infrastructure, Operations & Data Center Summit, looking for practical guidance on developing and managing your IT infrastructure and using virtualisation effectively.