The key to mainframe DevOps success is in quickly identifying and removing major bottlenecks in the application delivery lifecycle. Major challenges include collaboration between mainframe and distributed teams, lack of visibility into the impact of software changes, and limited resource flexibility with scaling out necessary testing initiatives. Now let's take a closer look at some of these key challenges and how IT departments can address them ...
Back in the day when the word “cloud” was only used for those things in the sky, software typically ran on local machines sitting in local datacenters. Often, this software was licensed from a vendor on a per-seat basis. Vendors tried their best to make customers renew their license agreements, and vendor lock-in was a common phenomenon.
Such lock-in could be caused by a variety of factors: sometimes purposely caused by the vendors by making migration to competing products unnecessarily hard, sometimes caused by customers being naïve and betting too much of their software stack or infrastructure on the products of one single vendor.
Over the past decade, two factors have made it significantly easier to escape that legacy lock-in model. Free/open-source software has reached a level of maturity and adoption in almost every area of the industry, offering more and more valid alternatives to commercial products.
Secondly, the commoditization of cloud computing removed the need for companies to run their own datacenter.
A Storms A-Brewin'
Cloud providers often claim to offer the perfect escape path from legacy vendor lock-in by providing modular solutions in which customers only pay for what they use, instead of being tied into complex, long-term, per-seat licensing agreements.
However, most cloud providers (in particular the big players) have extended their portfolio over the past few years, offering a broad variety of services that go beyond computing and storage. They provide everything from load balancing and DNS, to messaging, monitoring, log management, analytics, databases, and much more.
It's very tempting for customers to replace more and more components running in their own colos with these in-cloud solutions. Why? Because they typically are easy to set up and they remove the costs of hardware ownership and datacenter footprint.
It's also an easy purchase because a contract with the cloud provider is already in place, and adding (or removing) a service is not much more than a mouse click — no need to negotiate with a sales rep over complicated long-term license agreements like it used to be with many legacy, on-premise solutions (and no need for the provider to deal with selling through fear attempts and related nuisances).
I Can See Clearly Now …
With so many advantages, it is easy to be blind to (or willingly ignore) that these services have huge lock-in potential, and many providers probably exist for exactly that reason. The more of these services a customer uses, the more difficult it is to move their application stack from one cloud provider to another, or to their own datacenter or a hybrid solution.
Whenever companies make decisions to move services to the cloud, it is worth it to thoroughly review the architecture specifically from the perspective of potential vendor lock-in. It might also be worth it to go with solutions that live and function outside of a specific vendor's cloud, even if it is less convenient and maybe more expensive. It's not unlikely that today's convenience and savings will turn into a nightmare and cost explosion tomorrow. There is no such thing as a harmless, one-vendor dependency, not even in the most comfortable and fluffiest cloud.
Sven Dummer is Senior Director of Product Marketing at Loggly.