Last year, Capgemini’s VP of Packages, John Waymire challenged the SAP Solution Centre here to put SAP ‘in the Cloud’. A small team of us took up the challenge and started looking into how the public Amazon Cloud (Amazon Web Services or AWS Cloud) could be used.
SAP is deeply involved in the AWS Cloud and has been supporting the company’s internal systems since 2008 (see a video here and a wiki here). As a SAP Partner, Capgemini leveraged its strong relationship with the company to deploy a Proof of Concept (POC) landscape in the AWS Cloud in October 2009.
I’m pleased to say, it’s 2010 and the challenge has been met. Our SAP Solution Centre boasts a solid understanding of SAP in an Amazon Cloud environment. How? Simple. We put SAP in the Cloud to prove to ourselves it could be done, then built it for our clients to spread the word.
So what exactly did we build?
Currently the SAP Solution Centre has a Proof of Concept (POC) laboratory running in the AWS Cloud, primarily on Windows, but also on Linux-based servers, a key part of our next phase of development.
As the person responsible for the technical aspects of the landscape, I was part of the team that deployed, at various stages, the following systems from the SAP Business Suite 7 portfolio:
- Enterprise Core Component 6 (ECC 6) – on Windows 2003 & DB2 9.7
- Business Warehouse (BW 7.0) – on Windows 2003 & DB2 9.7
- Portal 7.0 on Windows 2008 & SQL Server 2008
- Customer Relationship Management (CRM 7.0) on Windows 2008 & DB2 9.7
- Solution Manager on Windows 2008 & SQL Server 2008
After signing a contract with a major British Utilities services company just before Christmas, the newly appointed Capgemini project team quickly found that there was no way to provision the required hardware for a blueprint system through normal channels. The project had to start within days of the contracts being signed, and the team was ready to move across multiple geographies.
In order to keep the contract on-time and on-budget, and ensure the climate was right to commence, it was decided to build or ’stand up’ the environment in the AWS Cloud. In four days we were able to build the above landscape in the AWS Cloud and have it ready for functional consultants to begin and complete work in 48 hours. In a traditional hosting environment this landscape would have taken weeks to create, simply because of the number of people it involves.
However, rapid deployment is not the only driver for the AWS Cloud as cost is also a key benefit. The solution outlined above has been deployed for approximately 40% of the cost of a standard hosting landscape. So successful was the deployment, it turned heads within our SAP Solution Centre. Within days, several consultants were exploring what part of their SAP landscapes could be put into an AWS Cloud.
This is a new and fresh challenge, as we seek to ensure that this solution is only used at the right time and in the right way. To that end, below is a framework that the Capgemini team has constructed to assess whether or not a solution can be deployed safely in the AWS Cloud.
Some of the high level criteria are:
constraints – How quickly do the systems really need to be available?
wants their systems immediately, but when does the system actually need
to be available?
requirements – What are the cost drivers for the customer?
course everyone wants to use the least costly option, but some drivers for
reducing costs are more compelling than others, and they should not be taken at
the expense of the project outcomes
is responsible for provision of the systems? – Is it Capgemini or the client?
has a huge bearing on whether a landscape can and should be deployed into the
– Is there an SLA for service provision?
AWS Cloud has no SLAs in place for a normal account, it is built upon the
concept of disposable servers and computing power
requirements – Who needs access and from where?
security – What sort of data will be held within the system?
AWS has passed several security certifications, there is still some fine tuning
required to the security model in relation to SAP
During the last 5 months, we have been able to move from a theoretical exercise of deploying SAP in the cloud to actually having a live SAP project deployed for a major British company as part of an overall programme that runs into the £100’s of millions. The learning curve has been substantial, especially moving from a POC environment to a live environment, as there are many things that are often forgotten in a POC. We have encountered several challenges during our landscape deployments, but we have resolved them and provided more functional and flexible environments in doing so. Technology always moves quickly, and all too often companies are guilty of claiming they can do something or that something is possible before it’s actually been done. We’re proud to say that we’ve not fallen into that trap.
In terms of next steps, Capgemini is working exclusively with SAP to provide better applications to help manage systems in the AWS Cloud and to expand our expertise into other areas, like SAP on a Windows Azure platform or IBM’s Compute Cloud. Following our success with Amazon, both companies are very interested in working with Capgemini to deploy SAP landscapes on their Cloud platforms.
The possibilities that the Cloud presents are immense, and on a personal level, I honestly can’t remember being so excited about being involved with something at such an early stage. However, in order to realise those immense possibilities, we as trusted advisors must guide our customers to use the right tools and methodologies at the right times. We must not allow ourselves to be prejudiced against traditional methods if they are right for the circumstances. This can spoil the party for everyone.
tweetmeme_url = ‘http://www.capgemini.com/technology-blog/2010/02/a_new_landscape_sap_in_…‘;
tweetmeme_style = ‘compact’;
This entry was written by Chris Kernaghan. Chris works in the SAP Solution Centre within Technology Services in Capgemini UK, he specialises in SAP upgrades and infrastructure migrations.