Copying and distributing are prohibited without permission of the publisher
Opex reduction and new revenue opportunities start in efficient data centre operations
16 April 2012
As operators continue transformation of their businesses, they are looking at how they can drive opex reduction and create operational efficiencies that reinvigorate the business model
As operators continue transformation of their businesses, they are looking both inwardly at how they can drive opex reduction and create operational efficiencies that reinvigorate the business model and outwardly at how they can adopt new approaches to managing infrastructure that enables new enterprise and consumer services to be offered. Data centre operations are a key aspect of either approach says Torsten Dinsing. Co-sponsored feature: Ericsson
Opex reduction and new service offerings
Operators' journey to the cloud
There’s a temptation to view data centres as a mirror of the traditional operator network business. After all, they’re composed of high volume, commodity priced hardware that’s reliant on electricity, security and cooling. Data centres are utility infrastructure in the traditional sense and, at first glance, it’s hard to differentiate in the data centre but, when the costs involved in operating a data centre estate are explored there are both business outcomes and real expenditure at stake. To run profitably, data centres must be optimised and to enable the high volume services of the future must be automated and flexible.
In their own businesses, operators rely on data centres to host their customer, operations and network data. Back office applications and the data they rely on also reside in data centres, many of which are operators’ own and are many years old. That breeds inefficiency, which can come in many different forms. Older equipment may not be efficient in terms of capacity, power consumption or cooling or even the floor or rackspace required. Older storage systems may not be suited to virtualisation and data centres can be inefficiently utilised because multiple locations are not integrated and available resource is not apportioned in the most efficient way. In addition, some operators that built their data centre estate in a less efficient era may have excess capacity or capacity available in the wrong places.
As a first step, such under-utilisation needs to be eliminated through consolidation and optimisation initiatives that take advantage of dense server technology, virtualisation and application migration. Such transformation isn’t a small step but the rewards can be impressive.
At one European operator, Ericsson has worked to address the operator’s challenges in terms of reducing cost in its IT and data centre operations. The operator wanted to optimise its hardware and licence costs and reduce all maintenance costs including those associated with personnel, energy consumption and software. It also wanted to achieve accelerated time to market and improved business agility from the data centre optimisation initiative.
The solution to these challenges involved data centres and networks consolidated into two locations with the data centre hardware renewed and consolidated onto blade systems that have a smaller form factor and lower cooling requirements that also result in radically lower power consumption. The operator’s new architecture encompasses full virtualisation for Windows and Linux based on VMWare and KVM and the benefits of that virtualisation are maximised through progressive application transformation. In this example, the operator is moving from a siloed architecture to an SOE (Service Oriented Enterprise) approach both for the operator’s services and data.
The engagement has seen a single platform architecture deployed across the operator’s entire IT environment, using standardised components. Approximately 90% of applications have been transformed and migrated to the virtualised environment. This enabled a 60% reduction in server numbers with the associated savings in licence, power, cooling and ADM (Application Development and Maintenance) costs that entails. Further benefits include virtualisation of development and test environments that have resulted in a 93% reduction in server numbers, resulting in a 60% saving in terms of power and cooling. In addition, the operator’s data centre capacity has been optimised with a 90% saving in floorspace, that has delivered a significant reduction in power and cooling costs.
Ericsson typically expects operators that use its Data Centre Optimisation services to achieve a reduction in opex of around 40% but there operators can gain more than simply cost savings from such programmes. By having physical infrastructure that is state of the art and encompasses efficient cabling, cooling, energy, bandwidth and processing power they are better equipped for business continuity and disaster recovery and have the foundation in place for delivering the operator services of the future.
Those services of the future are starting to be targeted by operators as their traditional revenue streams commoditise and the opportunity presented by delivering new types of service to enterprises begins to crystallise. That involves support for M2M, VDI (Virtual Desktop Infrastructure) and collaboration and mobility as well as SaaS aggregation and IaaS. That requires a robust, scalable and flexible cloud computing platform to be put in place to enable the operator to address these public xaaS offerings. Underpinning that is the network and also the data centre and operators are looking to build state of the art facilities designed for the purpose of supporting their cloud computing platforms and services and potentially those of their partners.
Operators’ journeys to the cloud will be long and are ongoing delivering business benefits on the way. The market today is currently focused on consolidation and standardisation in the data centre as part of their transformation initiatives. However, the pioneers are moving along the transformation path and optimising their data centre capacity and decoupling it from the services offered into a virtualised infrastructure. It is at this point in the journey that the idea of building cloud capability becomes important. If the goal of transformation is to deliver efficient internal IT, the goal of building cloud is to enable new enterprise services and new revenue. That necessitates cloud management and IT automation, cloud-specific infrastructure and ultimately, a global, federated network and corporate cloud infrastructure.
Ericsson has worked with one global operator on a project to enable efficient monetisation of the telecoms operator cloud. The challenge was to launch a cloud-based virtual data centre, virtual desktops and other cloud offerings into the enterprise market, as one global offering across all the local operators in the group and backed by a single global management solution with local execution in the operating business countries. The concept was that the operator would be able to differentiate through quality of service, service level agreements and local presence.
Ericsson was engaged as the main contractor to build, operate and support a network of global cloud data centres. That included the development of a complex and powerful automated management system comprising OSS, billing mediation, provisioning and SLA management. Throughout the engagement best of breed, commercial off the shelf components were used to create a fully virtualised data centre composed of computing, virtualisation, network and storage. That then integrates with the OSS and BSS of the operating businesses’ countries.
That building for the cloud concept is new to the industry and ties in with Ericsson’s consulting expertise which differentiates it from IT providers. In addition, Ericsson is vendor-agnostic in terms of hardware and software and does not compete in the operator market place. From its experience of more than 1,300 systems integration projects each year and numerous operator transformation projects, Ericsson thinks it is vital to have an understanding of the telecom technology and operations - from networks to BSS - to successfully build and integrate cloud platforms. GTB
Torsten Dinsing is domain lead cloud computing at Consulting and Systems Integration, Ericsson