Byte's Prognostication Bits
About this time every year I like to make a few prognostications for business technology. These are based on what I see on the horizon and the events of the previous year or so.
Containers
They're everywhere and everyone is getting into the game. This will only continue and probably accelerate. What I am watching now is the further maturation of management and performance related technology with containers. At the SuperCompute show, I briefly took notice of Singularity (http://singularity.lbl.gov/). This brings RDMA and other features important to high performance to the container space. I expect this type of functionality makes it into the major containers platforms in the near future.
We continue to see innovation around container management and orchestration in major projects like Mesos and Kubernetes and new startup efforts. This is a sign of a very healthy and still early stage ecosystem. Watch this segment for continued innovation and perhaps even some breakout moves.
Enterprise Storage
I think we all know that Enterprise Software Defined Storage is here to stay and that the traditional vendors are now fighting a losing battle to maintain their gross margins. What I see going forward is containerization of these environments. Containers are here to stay and are invading every nook and cranny of the technology space. The key is to make them unnoticeable while enhancing the flexibility of the storage deployment. This is the win that is coming here. The vendors that can capitalize on it with enterprise consumable deployment and management with real gains in management simplicity should gain significant mindshare.
Ceph is also making a major play for the data center. Its roots are in HPC, but it has not really been suited for anything but archival, up until now. This year, Ceph will have production BlueStore, bringing with it significant performance improvements. CephFS is being supported by one major supplier (and my employer, SUSE) already and is getting its own significant improvements. This project continues to gain momentum and being actively contributed to by both major enterprise Linux vendors and players in the storage industry makes it a force to be reckoned with.
64-bit ARM
This is a technology whose time is finally arriving. The problem it faces is consolidation in the SoC manufacturing sector. What makes me believe that this is a technology that is now arriving is the maturation of the ecosystem. We now have multiple SoCs in-market, ODMs producing platforms, and a growing software ecosystem. With the memory bandwidth and core density, some of these platforms could be very interesting for in-memory databases, analytics, etc.
Given the grenade that Qualcomm threw over the wall with the announcement of its 10nm ARM64 SoCs, the energy efficiency story is really taking shape. There was already a story to be made with the existing SoCs from other vendors like Cavium and AMD, but that is rapidly improving. Look for more entrants to start shipping silicon this year that rivals the traditional X86 space around specific use cases.
Tie 64-bit ARM with the open source SDS or OpenStack stories and there is definitely something to talk about, and I'm not even addressing the embedded side of the world where ARM is dominant already and is enabling smart devices and advanced functionality everywhere. This is an area where there is significant opportunity for businesses to take advantage of the technology at large scale and experience savings. Watch for more SoCs targetted at storage to emerge over the next 6+ months.
AMD
For too long, AMD has been in the shadows of Intel & NVIDIA. When they launched the AMD-64 platforms years ago, they stunned the world and shook the foundations of Intel's empire. Unfortunately, they slid towards insignificance in the data center, instead opting to focus on price-conscious consumers.
That all changed when they started letting out details around the Zen processor family now known as Ryken. This could be AMD's year. They have a chance to get the mindshare moving their way in the data center again. Intel is going to probably start feeling heat on multiple fronts thanks to the 64-bit ARM movement and the AMD processor releases. This may be enough to allow AMD to gain back momentum and foothold. My prediction, AMD shocks the data center segment with their new tech and moves the ball forward.
Storage & Memory Technology
In recent years we have seen many new introductions including NVME and NVDIMM technology. We have seen some thought on new ways to think about memory with The Machine from HPE. Now is the time to really think about revolutionizing the way we address and employ and utilize memory. I'm not an expert on how we do this programmatically or how we approach the challenge, but I know we are ready for a revolution here.
The shift towards microservices and API based cloud applications mean it is possible to do things with different systems underpinning them. As long as the right inputs and outputs happen, what significance does the underlying technology make? It's time for a clean slate and new ways of thinking. This will change processor technology, require new operating systems, etc, but it is worth doing. This is a longer term prediction, but one I've held onto for a few years and will continue to expect.
Looking at my team's product and roadmaps, it looks like we've got a strong play in each of these areas? ;-)