Equipping And Automating Data Centres for The Cloud Pt. 2

Governance and automation

Rezza: Previously, every unit in Telekom Malaysia (TM) ran their own servers, but it’s centralised now. The main challenge for TM now is to be a cloud provider. The challenges however are the lack of maturity of clouds in organisations and trust in the security and reliability of the cloud.

Hon: Of course there are consequences to taking on the cloud – like security. But this is a good opportunity for ISPs to roll out clean-pipe services, to help organisations make them always available and see the value of cloud. The data centre is ready, the apps are ready, but the infrastructure is not. It has to be there; otherwise, we can’t move.

Alain: There’s a lot of work that has to be done to demystify the cloud. We are regulated by the central bank, and we wanted to implement cloud. But at the regulator level, the understanding of cloud was quite clouded. Eventually, if we hadn’t gone in with the word ‘cloud’ and used ‘virtualisation’ instead, it probably would have gone through. So when vendors come to talk to me about the cloud, I always tell them it’s better for them to educate the regulators first, because that would make things easier for us.

Johari: Batch timing for banks is very critical, and we have to reduce the batch window. We need to complete our batch as fast as possible and as efficiently as possible. The software guys will just throw new software or applications to us and expect us to run without any scrutiny. We try to optimise as much as possible, not forgetting it’s not only one application we are running but multiple concurrent applications that may need to share files/resources with another, so we have to map it and create a critical path. One of the needs is automation. Once you get to a substantial size, you need to have a batch schedule. More often than not, the application guys will run their SIT and UAT and then pass it on to us, and the batch window can be 3 to 4 hours. I think we need to have a certain community or lobby group to look at breaking the silos and putting it all together.

Sandeep: I think we need to question what is happening during the 8 to 10 hours of batch processing, demystify and find out how we can reduce it because availability is the biggest challenge for insurance, which is very batch-driven. We need to ensure that the batch window can be reduced, to be as small as possible while still taking into consideration your regulators, your governance, and your reporting requirements. We must try to strike a balance. So even if you don’t have direct control of the applications, because they reside in separate business units, you can still add value and be an influencer and catalyst for users to question and discover a better way of doing things.

Nick Lim: When it comes to equipping the data center, generating inventory reports is very fundamental. Before you can think about the cloud, you have to know what you have – how many Windows licenses? How many Oracle licenses? How many of those are Websphere licenses? There are also different units of measurements; some are enterprise licenses and some are per CPU. The way you create inventory cannot be manual. It has to be automated, and this requires the discovery solution. Organisations need to know what they have before they can effectively manage it and ensure resources are fully utilised. Other questions to ask are: What are the top 10 applications or what is the hardware that consumes the most power? This will allow you to identify candidates for replacement or refresh, and ultimately reduce cost, but this all goes back to how you keep inventory up-to-date, especially in terms of hardware and software licenses.

Suhas: A trend I’ve observed is that fewer and fewer people are required to run IT in organizations. The costs involved aren’t always tangible. One example is virtualisation. One hundred physical servers become 2,000 virtual servers, and there are cost savings because there isn’t a need to buy more physical hardware, but who is maintaining those 2,000 servers? Also, necessary operations like patching and so on can be difficult without capable automation tools.

Nick: The cloud makes the IT infrastructure more complex, setting a higher service level expectation from end users. However IT resources remain the same, if not less. In reality, the IT department is always tied up with daily tasks and functions which could be automated instead, allowing them to focus more on creating innovation and other important value-generating work. IT departments always get the flak whenever something goes down, but no one praises them when things are working well. And CIOs are always asked to cut costs more than other C-levels because IT is perceived as a cost centre. In view of these challenges, CIOs will need to have the right people, processes and technology in place to ride on the “cloud” and be successful. – http://www.networksasia.net 

 Part one of this article can be found here.