6 DevOps Predictions for 2018
This past year was good for DevOps: we saw more enterprise adoption of containers, microservices and cloud native applications. Organizations are increasing their reliance on modern software development and delivery as part of their business model. Interest and adoption of DevOps is happening on a much broader scale, especially within the enterprise. As we head into 2018 here are a few things we think the future has in store for DevOps.
Artificial Intelligence is the next big thing in DevOps
Artificial intelligence (AI) holds great promise for DevOps. As humans, we learn from trial-and-error and we share our tribal lore with less experienced members of our tribe. That is exactly the promise of AI and machine learning. We prize our database administrators (DBAs) with 20 years of experience because they have vast experience in what has (not) worked in the past and because they can see patterns in the issues they deal with daily.
However, humans are limited in amount of data they can consume. Enter machine learning: if we are able to collect vast amounts of data on application change and its corresponding impact to our customers and systems, then it’s known problem to identify patterns in that data. In turn, we can prevent bad behavior and encourage good behavior, all without having to wake up at 2 a.m. to respond to an on-call issue.
Platform as a service will deliver on cloud promises
The classic three-tier architecture is not working for the cloud because it relies on a monolithic application that is brittle and difficult to deploy. Large companies will continue to refactor existing legacy applications to use microservices and transform them into cloud native applications. But that’s hard. Enter platform as a service (PaaS). With PaaS, companies can focus only on their application and leave the parts that do not generate revenue to the framework provided by PaaS. This will allow users to deploy to any cloud at anytime without having to roll out the entire application.
Pivotal (Dell Technologies), IBM and Oracle all offer PaaS via Cloud Foundry, an open source, multi cloud application platform as a service, to their customers. With Cloud Foundry, users can concentrate on their application and leave the heavy lifting of provisioning, security, logging and scalability to the framework. And it works. Large companies are adopting tools like Pivotal Cloud Foundry amazingly fast. Along with Pivotal Labs’ ability to inject cloud native application DNA into their customers’ development and operations teams with their onsite labs, Cloud Foundry will grow its adoption curve even more in 2018. The simple reason why is that it allows Cloud Foundry users to move their application to any cloud provider allowing for companies to deliver great customer experiences at lower prices.
RDBMS will continue to see growth and NoSQL will not replace it
Currently, the top relational database management system (RDBMS) vendors are as follows: Oracle, Microsoft, IBM, SAP and Amazon. However, Amazon’s growth rate in 2016 was 107.9 percent, according to Gartner. No other RDBMS vendor matched that growth rate. The second place in growth rate was Alibaba at 99 percent, which is easy to do when you go from $4.2 million to $8.4 million.
This is simply a result of RDBMS working in the cloud. In prior years, we heard that NoSQL will replace RDBMS as it is better suited for the cloud. That is simply not happening, nor will it. RDBMS solve a real business problem and the promises of NoSQL, outside of niche applications, will not deliver.
Simply put, SQL is the number one language used by programmers for a reason. It works and will continue to do so for transactions which make up the vast majority of database needs.
The race for second place in cloud services will be fun
Alibaba will continue its growth pace and leave Google in the dust. Both are growing fast, but Alibaba has the lead. Google had $500 million in 2016 while Alibaba had $675 million. Corresponding growth was 100 percent and 126 percent. It doesn’t take a math major to figure out who is going to win that race. The question is how Microsoft Azure will fare. Is Alibaba’s position with foreign suppliers greater than Microsoft’s tie with Office and Visual Studio users? That’s where the fun stuff will happen.
I’m reminded of what I overheard from an IBM’er in 2009 about Amazon: “Are you going to trust a book seller to run your production systems?” Since Amazon has already broken the stranglehold of the Big Four in systems software, do not discount Alibaba and its position with non-Western based developers.
Security will shift-left and development will claim it was all their idea
The portmanteau of DevOps continues to be “portmanteau’d” with DevSecOps. Of course, security has always been in operations and tasked with enforcement in dev, so the idea of calling it out is little more than marketing air cover. Nevertheless, the continued enforcement of security in development will increase and become more prevalent as development realizes the security team is serious…this time.
With cloud adoption continuing, security teams must enforce compliance at all levels. Previously, dev and test were wide open and allowed for no supervision for deployments. This was to enable speed. Security was only enforcing in production or “Prod-minus-one” environments. Now with DevOps, that system breaks down. DevOps teaches us to have the same deployment mechanism for all endpoints. That means operations deploy the way development does…it also means development has the same scrutiny and standards as production.
Once that requirement is made clear to Development, the Dev team will immediately begin to look for ways to automate that security enforcement, but in a Development way. In turn, those tools and processes will be then adopted by Operations. And the circle of life DevOps continues.
Fewer and fewer companies will forget the database with DevOps
The database is the hardest part in the application stack to manage, so it just doesn’t make sense that it’s always the forgotten piece of the puzzle. IT teams have been so focused on time-to-market and getting development to push out applications at the speed of light, but still manually manage the change process of databases that contain massive amounts of information. The good news is that as more enterprises continue to modernize and adopt DevOps processes, it’ll become harder to ignore the database. This is because DevOps is a process, an algorithm. It’s not static and it can’t be done some of the time. The whole purpose is to change and evolve over time. DevOps is about identifying friction that is slowing down software releases. Sometimes, it’s the testing team setting up environments manually. It’s time to automate environment creation to solve not just this one problem, but all problems across the IT department. It’s time to stop having DBAs perform manual SQL script review prior to a release and start automating the review so that they can continue to innovate and bring strategic value to their organization.
In 2018, more and more IT teams will start to see the benefit of bringing DevOps to other areas such as security and the database, and many will start to treat the database as a first-class citizen. Unfortunately, as this realization sets in, companies will regret that they didn’t tackle the problem sooner.
The Astros will repeat as World Champions
And Chris Devenski will win a Cy Young as only the fifth relief pitcher to do so.
Ok, so the last prediction is not related to DevOps but I wanted to throw it out there anyway. While no one knows what truly lies ahead for the future of DevOps it’s probably safe to say 2018 will be filled with more DevOps innovation as organizations continue to shift left to keep data safer, produce high quality applications and deliver a great customer experience.