Today's data center is morphing away from hardware-bound systems to software-defined networks, storage and even batteries. One consequence from this shift is a change in the skill sets of data center techs.
Increasingly, data center teams spend less time manually configuring systems and more time working with application programming interfaces (APIs) to automate and connect a wide -- and ever growing -- range of software components.
Traditionally, IT professionals tinker with data center hardware, manually connecting the spaghetti web of cables found in system backplanes. The techs rely on obscure scripts and command line interfaces to allocate server, storage and network resources. Such approaches are time-consuming and error-prone.
The increasing number and variety of devices connected to enterprise networks is partly driving the move to APIs. For example, the University of Tennessee at Chattanooga's network supported 3,145 devices in 2011. That number rose to 14,906 in 2014 and jumped to approximately 48,000 in 2015.
"Users work with a wide array of end-point devices," noted Thomas Hoover, associate vice chancellor and CIO at the university.
Software applications have become larger and more complex. Microsoft Office 365 requires 4 GB of space to run on a Mac for example.
Historically, software ran as a closed, autonomous application. An accounts payable platform did everything from presenting the user interface to reconciling the month's bills. Integrating software was difficult and time-consuming without automation and API systems.
Constrained IT budgets mean that data centers do not possess the luxury of time for manual optimization and integration projects. IT budgets will increase by a meager 2.2% in 2016, according to market research firm Gartner Inc. Since 2010, CIOs have failed to convince management to invest in technology: The 2.2% rise represents the high watermark in the last five years, including a 1.1% decrease in 2011.
Automate rather than add staff
Automated tools enable IT departments to manage the influx of devices without significantly increasing staff. APIs provide the abstraction layer needed to ease software integration and automate more tasks. Without APIs, programming code is tied tightly to particular functions, such as the operating system or a user interface. If a firm changes a system management function, such as the amount of available storage, the IT team has to do the task once for devices running Microsoft Windows and a second time for those systems running Linux. A well-design API empowers data center admins to make the change without worrying about whether the server runs Windows or Linux OS.
Because of an increasing reliance on APIs, systems today are much smarter than they were in the past. Many devices come preconfigured and perform tasks such as automatically setting up their network connections.
Consequently, rather than physically touch every device, data center techs now make virtual connections. They work with API systems, software libraries that enable components -- such as software objects, routines and data structures -- to interact with one another. In effect, APIs transform tasks that humans once did manually into chores that machines complete automatically.
API systems = virtual everything
This newfound programmability is key to the software-defined everything theme now sweeping through the data center. Revenue for the software-defined data center market will rise from $21.78 billion in 2015 to $77.18 billion by 2020, a 28.8% compound annual growth rate, expects market research firm Research and Markets.
As a result, people use APIs in a variety of places. Rather than fine-tune physical servers, IT pros work with system infrastructure APIs to provision (create, move or delete) virtual machines, security services and network settings.
Microsoft Azure's API Management cloud API service is one example. It includes a gateway that accepts API calls and routes them to backend systems; verifies API keys, JSON Web tokens and certificates; enforces usage quotas and rate limits; and logs call metadata for analytics purposes. The product's portal enables a data center team member to define or import API schema; package APIs into applications; and set up policies, like quotas or transformations, on the APIs.
APIs are emerging for cloud systems. The need for cloud APIs is similar to the emergence of the TCP/IP standard and its effect on networking. Without common standards, cloud administrators spend a lot of time making system connections. With standard interfaces, the connections are automated, with OpenStack as a prime example.
API training options
With the growing emphasis on API systems, how can a data center tech gain experience or formal API training? Various groups offer certifications.
Arcitura Education Inc., with services in 60 countries, offers courses in service-oriented architecture (SOA). Its SOA consultant certification tests both the big picture concepts as well as the technical details needed to connect different elements using various APIs.
Computing Technology Industry Association Inc. (CompTIA), a non-profit trade association, has about 2,000 member companies, 55,000 registered users, 3,000 academic and training partners and more than two million IT certifications. The CompTIA Network+ certification is a vendor-neutral networking course designed to validate the skills needed to design, configure, manage and troubleshoot wired and wireless devices, and delves into API as an interface to access and automate network services.
A long-established certification and accreditation company, EXIN works with more than a thousand accredited partners to deliver certifications in more than 165 countries and 20 languages. Its OpenStack certification program enables techies to work with and understand cloud API systems.
In addition to these independent companies, the industry's biggest vendors -- EMC Corp., Hewlett Packard Enterprise, IBM and Microsoft Corp. -- have certification programs geared to API training for their particular products.
Improve IT capacity by cutting data center sprawl
How does IT test for partner API security?
NYSE outage offers automation lessons for IT pros
Bring back system programmers for cloud, automation
Take this quiz on API codes, RESTful APIs and more