New demands on DevOps environments—from soaring data volumes to increasingly unpredictable workloads—have IT managers searching for new ways to extract more utility and cost efficiency from their data centers and clouds.
Increasingly, they’re finding that the answer to delivering applications and services at high velocity is to dynamically size data center assets to fit the workload at hand. In other words, the answer is Software-Defined Servers.
Why are Software-Defined Servers popular for DevOps environments? We can think of eight reasons:
1. They’re easy. With TidalScale, you can build a single virtual system from ordinary sweet spot cloud servers using a standard 10GbE switch, with no changes to operating systems or applications. TidalScale’s WaveRunner point-and-click orchestration software simplifies management and lets you build a new virtual server in minutes.
2. They’re flexible. Deploy TidalScale to gain the flexibility to change the size of your system on a moment’s notice.
3. They deliver results faster. By sizing the computer to the problem, there’s no longer a need to parse data sets and rewrite code to run across clusters. Meanwhile, you can improve performance by up to 300X.
4. They drive down CapEx. Software-Defined Servers utilize existing low-cost, commodity, two-socket servers and standard interconnects, minimizing the need for additional capital investment, lowering TCO, and maximizing your ROI from your current data center assets. Gain maximum and efficient use from hardware resources, and particularly servers, to deliver on difficult schedules.
5. They drive down OpEx. Users who run memory-intensive applications historically worked at a cost disadvantage, particularly when it comes to software licenses. Software vendors and hosted service providers typically price subscriptions on core counts, with a specific amount of memory tied to each core. (In other words, a hosting service might allocate 2GB of memory per core, so to run software on a 512GB system, the service bases the subscription on a 256-core server, even if the application doesn’t need them. That amounts to an unnecessary surcharge on users needing a computer with a lot of memory but relatively few cores.) By right-sizing the resources in your Software-Defined Server, you pay for only the cores you really need.
6. They’re built for high-velocity DevOps environments. TidalScale’s clean REST API enables an automated data center infrastructure to create, modify and disassemble Software-Defined Servers on demand. This equips DevOps environments to respond instantly to changing user requirements.
8. They derive more utilization from your cloud. With Software-Defined Servers, you can increase hardware utilization in your cloud by an order of magnitude. An excellent example of this is TidalScale’s own testing environment, which is the focus of this paper. At TidalScale, we use Jenkins to run all testing. The combination of Jenkins, Docker and TidalScale has increased our throughput by 21X.