By Manoj Chandiramani, Sr. VP and APAC head of IT operations, MF Global
|Manoj Chandiramani, MF Global|
Today, the business need is for IT teams to move beyond their traditional role as a pure-play internal service provider. The CIO and his colleagues should move beyond this traditional view and provide services to internal business units — eventually making the transition to a full-fledged service provider with end-to-end offerings. This is where we can transform our infrastructure optimally to provide service bouquets for customers.
When we talk about a business as a service provider, the idea is to leverage whatever services that we offer, in the best possible manner — especially in the financial services sector. For example, in our case, MF Global offers broking services. As the first Indian organization to become an approved provider of Direct Market Access to various exchanges in India, customers expected us to provide them with fully resilient architecture and a high-speed network, thereby enabling them to get the fastest possible access to said markets. So how do you transform your business into a similar service provider?
Define service catalogs
A service catalog is the first step toward assessing your data center capabilities as a service provider. A clearly detailed service catalog is a crucial aspect that companies overlook when trying to present their data center offerings to a prospective customer. While many organizations have enough capabilities on the data center front, the business may never have defined a service offering as such (except for maybe retail or consumer-oriented companies).
For example, how many data centers do you know that can actually offer a proper data center services guide? If I walk into their data center, can these data centers actually present a document that details their service offerings? Will they detail aspects such as 1U billing rates, AC cooling charges, desk space and SSL connects? The answer is no. At the most, they'll talk about data transfers and hosting charges.
Most data center providers will never talk about security services or service offerings (like backup services, installation and management services). It's unusual to find them discussing a solution, which is what your application desires -- or if they offer to design your outbound and inbound connectivity links, and design a solution for you that is worth your while.
So if you have to design a proper service catalog talking about the services that you can offer, start by making it generic in nature. At the same time, ensure that services are defined in such a way that they can be interconnected, or that a couple of services combined together can give a solution to a customer. Very rarely has this aspect been thought about, since people try to sell services individually. This is another important concept in how you can transform your data center into a service provider.
Your biggest ally at this point is that you have defined data center costs. You know the incurred costs, what your customers expect from you, what your customer is expected to pay for and the threshold levels that you can go up to. Hence, it becomes easier to define the services that your data center can offer.
Don't forget the customer
The basic idea is to study the customer's needs, as we discovered in MF Global. In our business environment, as a customer you expect your infrastructure to talk to our infrastructure in a seamless manner, with a guarantee of the best possible service offering — may it be connectivity, power, cooling or software application programming interfaces.
In a typical scenario, investors have servers in their offices, run thin, low-bandwidth links to the brokers' offices and push orders to the exchange. The disadvantages of this approach are huge link costs and high-latency-ridden connectivity. To circumvent these issues, customers started putting their own servers into local data centers in brokers' locations. Here again, the problems revolved around how high you could go. If you want high-speed switched networks, you have to look at direct connections.
Investment firms were wary of hosting directly with the broker's data centers due to other reasons as well. Security of their equipment, applications, data, etc., are paramount in such situations. Thus, we realized that circumventing these concerns with a unique hosting service will make sense, as well as make it easier for them to replicate the service.
We decided to design the entire deployment from a customer's perspective. The core requirement was that the customer needs to place an order to the exchange -- at the minimum possible time, with the best possible connectivity and hosting solution.
So, can I keep my customer's server alongside my server in my data center? The answer is yes.
With this approach, customers get a high-speed network connected to the same switch. This means reduced switching, fewer hops and better service from us as a broker.
Another advantage is the security aspect. We have state-of-the-art solutions available today (such as access cards) that can be deployed in such a fashion that they alert the customer in case of certain events. Furthermore, provide the client with surveillance systems that actually show what's happening in the data center from anywhere in the world, thereby enabling him to be virtually close to his server.
The manageability aspect can be addressed through multiple modes of low-cost connectivity options at the customer's end. Today, you have many options that allow you to access your infrastructure remotely from anywhere. These provide the flexibility and cost-saving aspects we are talking about.
So to sum up, we need to focus on designing optimal solutions with the final goal of increasing efficiency and reducing costs. Since you are doing all these things, you need defined service catalogs that detail the list of services that will suit every customer. At the same time, you should also be flexible enough to provide sets of services that are completely customized for the customer.
(As told to Anil Patrick R)