We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.
Start a Discussion

Are Servers Bloating in Pre-Production, but Shrinking in Production due to SOA?

Vote 0 Votes
John Michelsen: We seem to be noticing that with more distributed components and teams working on service-based "units" of functionality, that the number of actual servers used in Dev & Test is skyrocketing vs. the number actually needed in Production.  Is anyone else seeing this in practice and what are your experiences?
While as practitioners, we tend to live in the integration and virtualization side of SOA & Cloud, I have a theory (and I've bet a pizza with a colleague on this) that we have not less, but MORE servers popping up throughout the software lifecycle (dev, test, integration), all the way until Production, where there are actually LESS machines used -- enterprises are leveraging Cloud just for those "over-capacity" times in production and maintaining leaner environments with VMs for under-utilized times. Non-production servers can outnumber production servers by a factor of 20:1. Have others noticed this trend in the field?
Best real-world example answer gets a $25 Amazon gift card from me, they can use it to buy a book or pizza.

Editor's Note: Winner will be announced next Friday, March 5th.  All decisions are final.

7 Replies

| Add a Reply
  • The number of servers in pre-production are definitely outnumbering those in production, but this has to do with more with virtualization than it does with SOA. Virtualization has revolutionized QA. In the old days--like, last century--servers were expensive and QA budgets were tight (some things change; some things don't). People carefully scheduled testing around server availability and reused builds, populating and repopulating applications with scenario data.

    Today, QA simply creates a new virtual image and loads this on commodity hardware. Got a new collection of test suites that exercise a particular integration scenario? Just create a dedicated image. Load and execute as needed on available commodity hardware.

    This inevitably leads to some VM-sprawl, but if the team is well organized this new approach can be extremely effective. No group in the development and operational life cycle has been so transformed by virtualization as has QA.

  • Gotta do something with those excess servers left over from consolidating the production environment. I like giving my expired credit cards to my son to play with... :)

  • Agreed, there is potential for lots of server consolidation with cloud, virtualization and SOA on the production side. Vendors such as IBM with its big-box System z machines that can virtualize and run thousands of instances of various OSes are pushing this hard.

    There are several forces at work that may be fueling the rise of non-production servers. For example, with open source, folks can hang on to their hardware assets a bit longer, since they can throw Linux and other open-source solutions on older boxes for development, testing, and basic noodling around -- with relatively little additional cost.

  • My experience supports your view: there are a lot more servers in non-production environments than in production.
    It is reasonable that the number of servers is related to the number of environments. Since there are more non-production environments: Development, Test, Integration, QA etc. expect more servers.
    This argument was also true in the old Mainframe days (1960s and 1970s). For example, One Mainframe for production, One for Test and one for QA. The production Mainframe capacity was a lot larger then all the other Mainframes combined together. In a distributed Service Based, Virtualized and Cloud based environments the number of other environments servers is skyrocketing vs. the number of Production servers. This is true as well for virtualized servers instances. The reason for that phenomenon is that Production environments are more controlled because no one can allow uncontrolled Production. The risks of non controlled Development or Test environments are less obvious. As the number of instances grows the number of servers in the non-controlled environments grows more rapidly than the number of servers included in the controlled environment. However, logically the number could be reduced at least to a ratio of 1:3 or 1:4 and not 1:20 if the non-Production environments will be managed and virtualized similar to the Production environment. My prediction is that sooner or later costs reduction will drive proper management of the other environments and the ratio will be 1:5 or 1:10 and not 1:20.

  • Good points all - more opinions? Pizza bet day for me is Friday and we are wondering if the twain shall ever meet (production vs. non-production accountability for acquiring, setting up and maintaining servers - whether hardware or virtual in this regard).

  • John asks if the twain will ever meet -- production -vs- non-production accountability for acquireing, setting up and maintaining servers.

    In a majority of customers I speak with, their organization, culture and funding models are no where near having this happen... different silos are managing hardware and software resources and different SLAa are associated with them. In fact, in some cases, SOX compliance keeps organizations from being able to mix and mingle their networks and data between pre-production and operations. That said, I see one place where this can happen.. in the cloud...

    when customers move to outsource infrastructure in the cloud, it's conceivable that the cloud service provider can provide infrastructure on demand regardless of whether it's pre-production or production as long as they maintain data privacy, security and SLA assurance...

    interesting question by the way!

  • K-Scott, Joe, Tarak, Avi, Kelly, really appreciated the insights. I probably owe my colleague a pizza, but I'm going to call Avi the "winner" if there is one on this question. Kelly you are onto something with the organizational reasons.

    We are doing such a good job trimming budgets in production hardware, that for preprod, I think somewhere in the organizational and SLA process we'll find the answer. I blogged about this at http://bit.ly/c1YO7E. Thanks everyone!

Add a Reply

Recently Commented On

Monthly Archives