Adoption of HPC on cloud will see growth: IBM

IBM has been a leader in cloud. IBM also recognized the value of containers early on and invested several years of R&D resources to transform its software portfolio to be completely cloud-native. Here, Subram Natarajan, Chief Technology Officer, IBM India/South Asia, tells us how businesses can simplify cloud adoption. Excerpts from an interview:

DQ: How will businesses look to simplify cloud adoption?

Subram Natarajan: Simplification is sought in two ways essentially. One that focuses on simplification of IT operations and deployment: this includes the transformation of infrastructure elements into programmable entities and therefore provisioning, orchestration, management, monitoring can all be software-driven. The second is to abstract many of the common functions of the enterprise and deploy them as a service provisioned on the cloud. This can be in the form of the entire application function (SaaS) or specific functions within an application, such as Functions-as-a-service.

DQ: How will cloud-native innovations and a renewed focus on security reshape cloud management?

Subram Natarajan: We’re currently at a tipping point in the security industry, driven by two major market forces that are converging – security fragmentation, and the march towards widespread cloud adoption in the enterprise. In order to avoid becoming the next security headline, companies have raced to adopt the latest and greatest security tools, causing the market to expand rapidly – catapulting worldwide security spend to $124 billion in 2019 according to Gartner (nearly double the spend from just 5 years ago). We’ve reached a point where large companies are often using 50 - 100 different security tools from more than a dozen different vendors, with each point-product addressing a small piece of the ever-growing cybersecurity puzzle.

At the same time, this ad-hoc adoption of cloud technologies has created a more complex IT landscape to secure, with gaps in coverage and visibility, and data being spread across multiple tools, cloud and on-prem infrastructure.

More than half of security teams say they struggle to integrate data with disparate analytic tools and to combine that data across their cloud environments to spot threats in advance. As we head into 2020 – there is a clear need for simplified, connected security platforms as well as better visibility into threats across the hybrid, multi-cloud IT environment.

If we were to look closely at how we address the cloud management requirement especially in light of changing security landscape, it will be apparent that what we need is a platform-based approach which allows for quick integration of security tools that exists in an enterprise, in order to drive deeper intelligence into threats across hybrid, multi-cloud environments.

Therefore, in a cloud-native environment which accelerates enterprise-wide innovation, the following traits become supercritical:

* Whether it is on-premise, private cloud or public cloud, the security solution should be installable in any environment, and it should run anywhere. Connect to security tools openly and operate with a single interface. Operating in a cloud-native environment, the software tools must be deployed in an open containerized format. IBM Cloud Pak for Security installs easily in any environment.

* Gain security insights without moving data. Transferring data in order to analyze it creates additional complexity. The platform should connect to all of the existing data sources in an enterprise to uncover hidden threats and make better risk-based decisions while leaving the data where it resides. Preferably, this should have an integrated search application, which can access all sources including the third party, for threat indicators.

* Respond faster to security incidents with automation. Speed and agility is of great importance when it comes to threat detection and resolution. Therefore, the platform should allow the automation of workflows with a unified interface so that one can respond faster to security incidents. In an enterprise, Security teams are having to manage an average of 200,000 potential security events per day, and coordination of responses across dozens of tools.

Vendors of security solutions are driving their strategy more and more towards such design principles and their offerings are increasingly aligning with this thought process. A great example is IBM's Cloud Pak for Security, which is a platform designed to meet the security requirements in a cloud environment.

DQ: How will IBM focus on helping enterprises use the OpenShift development platform, on any cloud, to modernize core business apps?

Subram Natarajan: IBM recognized the value of containers early on and invested several years of IBM R&D resources to transform its software portfolio to be completely cloud-native. We are now delivering our cloud-native capabilities as pre-integrated solutions called IBM Cloud Paks. The IBM-certified and containerized software will provide a common operating model and common set of services – including identity management, security, monitoring and logging – and are designed to improve visibility and control across clouds together with a unified and intuitive dashboard.

Additionally, IBM has optimized more than 100 products from across IBM’s expansive software portfolio to run on Red Hat OpenShift. With new Cloud Paks and IBM software optimized for Red Hat OpenShift, enterprises can now build mission-critical applications once and run them anywhere -- on all private and public clouds, including AWS, Microsoft Azure, Google Cloud Platform, Alibaba and IBM Cloud.

DQ: Why and how will cloud interoperability become paramount?

Subram Natarajan: In any organization, heterogeneity in the IT landscape is a given. There will be a mix of bare metal, on-prem cloud, public cloud and that too from multiple service providers - the choice of deployment mainly driven by the application.

Interoperability brings the flexibility to deploy workloads and migrate to any of the platforms at will (mainly driven by ease of use and cost considerations). The underlying requirement for this interoperable environment is, of course, the need to embrace open standards that brings the commonality and removes the vendor lock in.

DQ: How will the use of high-performance computing grow in the public cloud?

Subram Natarajan: What has traditionally been an on-prem setup predominantly, HPC is gaining a foothold and expanding on cloud steadily. For those SMEs who continue to experience high-cost entry barrier for HPC setup, now cloud-based high-performance computing provides an avenue for running their workloads and test their simulations quickly.

As the HPC ISV ecosystem continues to expand its support on the cloud, the adoption of HPC on the cloud will see growth. Cloud-based HPC will provide better cost equation with respect to pay-per-use of the setup, protection from technology obsolescence and above all allow the companies to focus on solving their scientific problems rather than having to worry about the maintenance of IT.

DQ: Do you see open source cloud-native development battles target service meshes/serverless?

Subram Natarajan: Both service mesh and serverless play distinct roles in code deployment patterns. Cloud-native development targets service mesh and serverless as one of the key architectural deployment patterns, among many others. Clearly, serverless architectures have limitations to the type of use cases: for example, for HPC workloads, it may be cost-effective to pre-deploy a specified set of server resources.