Tune into: Hype Hopping
About a year ago we tuned into “the need for speed” and how a concept like serverless computing” was increasingly catering to this. We are now a year further and the term “serverless” is taking on unexpected proportions. With some even seeing it as the successor to cloud in general or at least as a successor to the clouds’ poorer cousin in terms of revenue, hype and adoption: PaaS.
The question we need to ask is whether this constitutes an example of Hype Hopping: to effortlessly pivot to the next new thing once the previous one turns out to be just a bit less attractive and certainly a lot more complex then we all thought at first. The Gartner Hype Cycle has been calling this for years the trough of disillusionment, a valley that only the strongest of innovations manage to pass in order to reach the slope of enlightenment or even the plateau of productivity that lies beyond.
But even before the through there are pitfalls that hypes need to survive in order to thrive. Cloud computing itself once started its journey as “on demand” or “utility” computing. Terms that in retrospect were not sexy enough to survive. Whether Serverless will survive the markets sexiness test is something that remains to be seen. Also because – unlike there are no real clouds in cloud computing – there are plenty of real servers in what is called serverLESS computing.
Whether serverless indeed will be sufficiently different to be deemed as the next generation of “How to do IT” is not easy to answer. After all, we saw plenty of earlier generations of different approaches, such as Structured Programming, Object Orientation, Service Oriented Architectures and now Micro Services, all lay claim to such a change agent role. But deep down they all were just similar enough to allow a grey bearded mainframe type to claim: “been there, done that, on our sixties S/360”. And let’s face it, when it comes to serverless, did not virtualization, software appliances, containers and everything delivered “as a service” already take many steps to remove any physical servers from our direct field of vision.
The most visible incarnation of serverless is currently Amazon Web Services’ Lambda. Although this was by most accounts not the first implementation of the idea. Manta of IaaS provider Joyent – recently acquired by consumer electronics giant Samsung – and Iron.IO’s Iron Worker arguably were earlier. And neither is Lambda any longer one of a few. Due to rapid succession introductions of new offerings such Azure Functions, Google Cloud Functions and IBM OpenWhisk. Although many of these newbies are still in a beta or even alpha stage, the term functions is rapidly becoming a standard when it comes to naming serverless offerings. And Functions as a Service (FAAS) or Function Platform as a Service (fPaaS) is even used broadly as a more precise (but therefore more confined) alternative for the term Serverless altogether.
Most of today’s serverless implementations enable users to execute user-defined functions based on various event triggers. For example, one can have a thumbnail created every time someone saves a picture, or send a bill every time someone streams a song, or verify the identity of a user every time he triggers an event. Behind the scenes the invoked function is usually performed in a container (mainly because they are so fast to start-up). While the individual containers are often running in an isolated and secured user dedicated environment, in most cases a Virtual Machine (mainly because they provided proven insulation and safety). On some platforms you declare your functions by inserting a piece of code or a script, with others you can insert functions in the form of a ready to run (binary) container. The latter feels -with a container basically being a portable machine incarnation – somehow a lot less “serverless” than the script approach.
The essence of serverless in my view is however that in addition to no longer having to worry about WHERE (on which server or which virtual machine) your functionality will be running, you also don’t have to worry anymore about WHEN your function will be performed. This is taken care of by the trigger or event engine of the serverless platform. This may make classic programming constructs such as loops and infinite, nested and complex “If – then – else ” trees, a thing of the past. And we all know how much code it can take to handle the logistics, versus the core transformation, in any real world applications. Not to mention how hard it is to debug such logistical flow code. Not surprisingly one of the most frequent comments heard about serverless computing is how amazingly little code you have to write to get something done.
With serverless the provider/operator of the platform is responsible for the WHERE and the WHEN and the user/developer needs only to determine the WHAT. In some way this sounds as a familiar promise. Did not non-procedural and event-driven 4th generation languages lay similar claims? And if so, could serverless long-term turn out to have the same disadvantages as these predecessors. And I don’t mean just the increased lock-in that these platforms brought, but the fact that in case of performance issues tuning let alone refactoring was almost impossible, as the environment almost fully abstracted the user/developer from the HOW.
Whether we will all be shredding our just recently printed business cards and updated linked-in profiles claiming our new found role as “Cloud Something” to replace them with “Serverless Whatever” is therefore questionable. If only because “it runs in the cloud” still sounds so much better than it runs “at the Serverless”.
“At the Hop” by Danny & the Juniors rose directly to the top of the charts in 1958 and turned out to be the bands’ biggest but certainly not their only hit song. Others were the largely forgotten “Dottie” and the more persistent “Twistin ‘USA”. Although the dance moves of the Hop were clearly different from the Twist and from subsequent Rock & Roll and Hip Hop variants, for many older people it all seemed just more of the same pointless hopping around.
The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.
Comments are closed