A 4-Step Plan to Cloudification
James Crawshaw, Senior Analyst – Service Provider IT and Automation, Heavy Reading
A recent blog by McKinsey poured scorn on the concept of “digital transformation” as a panacea for telecom operators struggling to remain relevant in the 21st century. As the article points out “labelling an initiative 'digital' is counterproductive since it often alienates more-experienced managers and results in a handful of pilots that are eventually too abstract to drive change”.
Part of the problem with the term digital transformation is that it is so vague it can mean whatever you want. As this article from telecoms.com notes, “Like many persistent buzzwords it’s sufficiently broadly and vaguely-defined that it can be applied to pretty much every aspect of corporate evolution.” Telecoms.com succinctly breaks digital transformation into two buckets: technological evolution (including virtualization, cloudification and automation of network management) and cultural evolution (from waterfall to DevOps, from asset-based to service-led, from standards-oriented to open source, etc.).
In this article I’d like to explore one element of that technological evolution for operators: cloudification. Telecom operators, like most enterprises, have largely embraced virtualization in their IT systems. However, putting legacy systems, such as CRM and billing, on a virtual machine does not bring the elastic scale-out capabilities of true cloudification. For example, an operator might end up with multiple virtualized resource pools from different suppliers some of which might be overwhelmed at peak times while others are almost idle. IT teams are left with a new challenge of orchestration given the lack of a unified management platform to run all the cross-domain IT applications.
This can mean service rollouts and adding new functionality is just as slow and painful for virtualized applications as when they ran on dedicated appliances. Even straightforward tasks like registering a DNS can take the IT team several days to complete. Needless to say, basic virtualization does little to help the operator reduce the opex associated with its IT systems as they may still need aging UNIX servers and storage devices to support them.
To address these problems operators might consider a four step plan:
1. Formulate a hybrid cloud transformation roadmap. Core applications such as BSS and OSS might need to remain in the traditional IT architecture indefinitely while others (such as CRM) might be migrated to the cloud over multiple years.
2. Build a Software-Defined Data Center where all infrastructure is virtualized and delivered as a service. This step requires a centralized management capability that spans multiple heterogeneous virtualization resource pools and physical machines. The unified portal for monitoring and management might also span a mixture of private and public cloud infrastructure. Common templates can also be created for application lifecycle management to free up administrator time.
3. Migrate core services to the cloud while ensuring zero data loss and minimal service interruption time. This is a multi-year process as legacy systems are often comprised of a “mash-up” of software and hardware. Many aging computers and storage systems need to be evaluated to determine which applications on the live network can be moved to the cloud. Application dependency can be very complex and upgrade downtime and migration windows can differ considerably between services where overlapping migrations lead to prolonged downtimes. Migration tool chains are essential to capture and restore application associations. Prioritizing the migration depends on the cost saving potential and risk factors for each specific application. Migrations should be thoroughly tested and verified in a sandbox environment before their planned and automated execution.
4. Continuous optimization of the cloud environment and improved operations and maintenance (O&M) efficiency. It is important to have a unified management and O&M system that covers all of the operator’s private and public clouds. Key to automation is the creation of a library of common service lifecycle use cases and scripts (e.g. for installation and configuration) that are orchestrated and managed by automated process tools.
Following a plan such as this can lead to significant benefits for operators. One Huawei case study of a European mobile operator refers to a reduction in new service roll out times from 6-8 months to 2-4 weeks as a results of data center cloudification. The same case study cites an anticipated reduction in total cost of ownership over 5 years of 42%.
Another Huawei case study involves the Nanyang government cloud project. Nanyang Mobile provides IaaS, such as cloud hosts, cloud storage, networks, and security services. In addition, Nanyang Mobile provides PaaS capabilities that support third-party application ecosystems such as government libraries, data sharing and exchanges and government big data analysis.
A third Huawei case study is that of Telefonica Peru. Telefónica Perú chose Huawei’s integration delivery services, which greatly improved the delivery efficiency of the Lince data center project. You can read more about this here. Digital transformations only have a starting point, no end. They are a journey of continuous improvement. Perhaps that is why they are so hard to define. Nevertheless, telecom operators cannot afford to let their IT estate ossify and whatever you want to call it, the move to the cloud seems like the inevitable next step.
To find our about Huawei’s data center integrations solutions please visit this page:
This blog is sponsored by Huawei.
— James Crawshaw, Senior Analyst, Heavy Reading