Start United States USA — software Predictions 2018: Random Thoughts on AI, DevOps and Other Topics

Predictions 2018: Random Thoughts on AI, DevOps and Other Topics

280
0
TEILEN

In doing this series, we’re reflecting the knowledge and depth of field that our readers have in sectors all over the IT business. These are your stories, not ours.
If you’re a regular eWEEK reader, you’ll know that we’ve been running a long series of “Predictions 2018” articles all month, and we enjoy doing them. All of us like to find out what’s coming from the future via experts whose perspectives are undoubtedly better than our own.
In doing this series, we’re reflecting the knowledge and depth of field that our readers have in sectors all over the IT business. eWEEK’s been around a long time—at first as PC Week from 1983 to 2000—and, like things you gather for your home over the years, we’ve earned a lot of loyal readers along the way. To all of you, we say: “Thank you, and continue to help us stay ahead of the IT curve, so we can package your knowledge and publish this information for you every day.”
Frankly, it’s also nice that all we really need to do it edit these predictions, write a lead-in, whip them into a coherent style and publish them for you. These are your stories, not ours.
With that, today we’re publishing a potpourri–a series of random prognostications—about trendy and topical IT sectors. It’s sort of like shopping at Marshall’s, Home Goods, Tuesday Morning or a second-hand store; you never know what you’re going to find. We might come up with a few more before the end of the year, too.
Here are those random predictions. Enjoy, and have great holidays and a wonderful new year!
Nara Rajagopalan, CEO, Accelerite: AI isn’t just a hype; thinking machines WILL rule our lives (perhaps not literally). “My school-going daughter is playing with Tensorflow to see if she can generate harmonies for any given song for her choir. She doesn’t know the math behind the learning algorithms, but things are getting easy enough for most people who can write some code to explore machine learning. While the basic math behind AI has been well understood for years now, the cost of compute an its availability along with the tooling around it is moving at a fast pace to democratize AI.
„Several companies including us are exploring if we can convert a Q&A worksheet of what one wants to predict into learning programs. The job of assembling meaningful learning data sets, seeing the accuracy of predictions, constantly tuning the algorithms, even picking different ML models to improve your predictions – all of this is getting easier by the day. Not only are the machines learning, but they are picking and choosing among various learning algorithms to see which predicts the future better. Scary–and you thought Elon Musk is nuts to think machines will rule the world!”
Daniel Saks, CEO and co-founder of AppDirect: 2018 marks the beginning of the digital economy. “We’re connecting all devices online and in the cloud over the next 10 years. In 2018, we will see more connected machines than ever before. This will have a massive impact on manufacturers themselves as they now have the opportunity to remain competitive. They are essentially supporting the creation of a digital supply chain, which enables companies to run their business in the cloud and ultimately meet consumers’ demands. The combination of Infrastructure-as-a-Service, Software-as-a-Service, and distribution technologies is what makes this possible.
“For example, AWS reduced the cost of computing technology and SaaS companies have reduced costs for general business functions such as payroll, inventory, finances, etc., and now it’s cheaper than ever for distribution. Just look at how Amazon Prime drastically reduced the cost of shipping and delivery, which ultimately means new opportunities for logistics companies. The companies that embrace digital practices early on will be able to transform faster than others, and those that fuse traditional with digital practices will be the ones that ultimately thrive in this digital revolution.”
Cole Crawford, CEO and co-founder, Vapor IO:
Jason Hand, DevOps Evangelist at VictorOps:
Ted Dunning, Chief Application Architect, MapR Technologies: DataOps emerges as the key organizational approach to drive agility . “We see a trend toward embedding data scientists and data-focused developers into otherwise traditional DevOps teams to form a DataOps team. Better communication, focus and goal orientation by cross-skilled teams results (importantly) in faster time to value and better agility. Organizing work in a DataOps style improves the ability for timely and appropriate responses to changing conditions. Better flexibility and efficiency at the human level helps take full advantage of new technologies and architectures.”
LogicMonitor: “Our research found that in 2017,37 percent of all IT workloads continue to run predominately on-premises, with 31 percent in the public cloud. However, over the next two years, the percentage of premises-based workloads will drop to 27 and workloads running in the public cloud will grow to 41 percent with the balance running on private or hybrid clouds. Key drivers of this transition: 63 percent cited digital transformation followed closely by IT agility (62 percent) and DevOps (58 percent). This changes dramatically by 2020 when AI and machine learning takes the lead, followed by IoT.
Grant Wernick, CEO and Founder, Insight Engines:
Marc Holmes, VP of Marketing, Chef:
Chris Sharp, Chief Technology Officer, Digital Realty:
Gary Tyreman, CEO, Navops by Univa: HPC begins a major move to hybrid cloud. “For years, analysts have predicted the migration of high-performance computing to the cloud, but adoption has been slow. Aside from oft-cited concerns like bandwidth, latency, and compliance issues, there are other barriers including complexity, software licensing terms, and concerns about proprietary lock-in. Perhaps the biggest barrier is related to cost. Sustained utilization in well-run HPC centers can exceed 90 percent leading to rock bottom per core-hour costs that make cloud look expensive by comparison.
„Despite this, HPC users have an insatiable appetite for computing power. As we stand on the cusp of a new year, these barriers are now falling away. Cloud providers offer HPC optimized systems, applications are increasingly open-source or have cloud-friendly licenses, and technologies like containers ensure portability. It is now easy to tap cloud-based compute cycles opportunistically, economically, and seamlessly. The stars have aligned, and for HPC practitioners, hybrid cloud will become essential to ensuring competitiveness.”
Rob Lalonde, General Manager, Navops by Univa: Container orchestration getting to scale. “ While Docker and containers are all the rage, getting to scale is hard. Modern “born-in-the-cloud” applications benefit from container orchestration, but these applications are still the minority.

Continue reading...