Mode 1 is the existing enterprise world. The data center is the IT center of gravity. IT services pull in the users to the data center. The data center contains applications and infrastructure that are cannot fail. Mode 2 is the digital world, the next technology frontier. In the digital world, the center of gravity shifts from the DC to the users. Services are delivered to the users via mobile devices. Data is easily accessible to obtain insights. Systems are designed to fail, in that failures can be tolerated. Services still run.

Two very different IT DNAs. In South East Asia, most enterprises are faced with a situation where IT will have to make decisions between keeping Mode 1 running, and investing in Mode 2 to prepare for the future. And that decision is pivoted on a stretched IT budget.

Here is my take on some of the strategies to navigate a bi-modal IT landscape.

  1. Reduce Mode 1 infrastructure costs. Build a view of the IT landscape by workload characteristics. Identify application groups, required data services, platforms, SLAs. Have clear categories. Look at consolidating and replatforming aggressively. Leverage all-flash storage platforms to have an absolute reduction in infrastructure.
  2. Leverage software-defined infrastructure. Where possible, run Mode 1 workloads on all things software. Run VMs on SDN and SDS. Deploy software-defined storage platforms that runs both in storage-server mode, and in hyperconverged mode. Software-defined platforms are usually distributed in nature, supports various workload platforms and are designed for failure. Software rules.
  3. Initiate modular projects for predictive analytics. Demonstrate gradual value to the business. Look at modular data science lab engagements as a launch pad, and for skills transfer. Get help in building possible use cases for your business. Data science and big data analytics takes time to build up to maturity. Start early.
  4. Leverage bi-modal platforms. These are platforms that double up in function. For example, as an unstructured data platform that is analytics ready. Keep files and raw data into scale-out file platforms that supports Hadoop and MR in-place. This is to ensure that the “data is ready and available” for modeling and analysis, when the business demands it. Today, there are options on filers or object platforms.
  5. Deploy new gen applications on PaaS platforms. Like CloudFoundry. Let developers focus on development. PaaS takes away the tasks of infrastructure resource management, binding of services and application deployment. CloudFoundry in particular runs on any IaaS platforms. Let your developers begin building apps with microservices models. Again, skills take time to build up. Start early.
  6. Augment events-based security with intelligence-based security.In digital world, data no longer reside only in datacenters.Today, company and confidential data are ubiquitous in mobile devices, cloud services and sometimes, even social media. Not only are these new sources of data, but these are also new surfaces of attack. Deploy risk-based authentication engines. Correlate network packet data with logs, endpoint and cloud data for complete visibility and deep forensics.

As usual, please feel free to augment other strategies and experiences that work. Cheers.