What if the IT competencies of Google, Facebook and Amazon (aka hyperscale data centers) could be replicated by your own IT teams in your company’s data centers? It would then be called web-scale computing in an enterprise IT. According to a recent Gartner research paper, following six elements are required for a solid web-scale computing foundation:
- Software-driven data center
- Web-oriented architecture
- Industrial-strength data centers
- Agile development processes
- Collaborative organization style
- Learning culture
While larger enterprises can benefit from this new shift in Data center thinking, even smaller IT departments can benefit by improving the velocity of IT service delivery and reducing costs.
Hyperscale data centers (let’s call them Tier 1) like Amazon, Google, and Facebook have shown how to deliver scalable services faster. Others like Box.com, Ringcentral.com (let’s call them Tier 2) have been offering various storage and collaboration services in the cloud, albeit at a lower scale.
There are going to both vertically-integrated and horizontally-connected public clouds, some will be specialty clouds like Salesforce and Workday, while others will be plain vanilla clouds that customers can use as building blocks (e.g. Microsoft Azure) to construct their own interconnected services using APIs or other integration services.
There is also this massive shift happening towards hardware commoditization, in which software is sitting on top of an abstraction layer that drives this low cost hardware.
We all know that CRM, HRIS, B2B applications like Salesforce, Oracle HCM, Workday, Serus and many others have reached maturity levels. While companies feel more comfortable going for public cloud deployments for such applications, there are still some other applications that these companies will tend to keep in the private clouds for various reasons (security being one of the considerations).
In order to realize scalability, agility, security and lower cost of ownership, companies will use an appropriate mix of public and private clouds, called Hybrid clouds.
For private cloud deployments, companies need to evaluate potential platforms (e.g. Openstack) and the integration requirements between private and public cloud applications. IT architects need to get handle on end-to-end architectures for smooth business operations in their software-driven data-center initiatives.
It is in these private clouds, where your IT teams will have to bring in best-practices of hyperscale IT data centers and that is where web-scale computing will come into picture.
As enterprise IT departments start running more and more of cloud workloads in their data centers, and enhance usage of BIG Data analytics, they will need to transform their data centers to emulate the successful web-scale computing models. To do this, they will need to consider the following –
- A self-service infrastructure which can
- Rapidly provision services
- Enable movement of workloads between private and public clouds
- Allow for multi-tenancy and
- Is simple to manage.
- A high performance and scalable network which provides the right balance of
- Throughput and
- Data processing (ERP like Transactional systems) and Analytics capabilities (Data warehouse, Dashboards for drilldown, slicing/dicing) which support the new data repositories and applications.
While Web-scale computing is a new concept for enterprise IT departments, it will become a requirement in private cloud deployments and we are going to hear more about it in coming months and years.