Eight steps to best practice in end-to-end network optimization (pt 2)
In the first part of this post I talked about the first four steps for successful end-to-end network optimization. These steps were focused on identifying the applications that meet business priorities and the KPIs that will measure the success of the network optimization project in terms of the business. The next four steps deal with the actual work that needs to be done to optimize your applications and infrastructure.
Step 5: Determine the infrastructure constraints
Now that all the high-level decisions have been made, our process digs down into the IT infrastructure layers that support these business priorities. As outlined in the last blog post, you need to take an holistic approach and look at all the components of the network and application infrastructure, for example:
- Is application is "WAN ready"?
- What is the network transport and bandwidth usage efficiency?
- Are there any performance bottlenecks for the server locations and remote users?
- What is the application level security?
- Is application visibility and monitoring in place?
- How do application reports and SLAs relate to KPIs?
Once the infrastructure constraints have been established, we can then get down to looking at how to solve them. There is no one-size-fits-all answer and solutions can include: application code enhancement; server configuration tuning; WAN optimization or acceleration or proactive service management. The best approach for you business will depend on your particular constraints and goals.
Step 6: Determine what IT activities are core to your business
The previous step will give you visibility of what your current baseline is and how your infrastructure is performing. It will also sketch out what you need to do to meet your business objectives. The decision now is how much of this to do yourself. You need to ask yourself whether infrastructure management is a core or non-core area of your business. Network optimization is a rapidly developing technology area and to use it effectively, it is essential to keep on top of the market.
Because of the complexity of the IT infrastructure, many businesses decide to out-task elements of network optimization. To do this successfully, you will need to decide which parts of infrastructure management you need to retain control over, such as:
- Change management;
- Control over the actual devices;
- Application performance planning and management;
- WAN optimization and application delivery technology;
- Network & IT infrastructure management;
- Service management.
Step 7: Determine service model needs
Because all companies will make different choices as to the level of involvement they want in infrastructure management, it's vital that any service provider has a variety of service models. At Orange, we have developed three different service models to address these different needs. They are:
- The integrator model: This model uses a range of different optimization devices on the customer premise, customized to your needs and can be delivered on any network. This model is extremely flexible as you can decide to have a little or much control over the devices as you require and equipment financing options;
- The network-centric model: this approach is much more focused on delivering network optimization as a service, and uses functionality and features inherent in the Orange MPLS VPN based on a operational expenditure (OPEX) cost model;
- The data center model, and finally this approach is focused on IT services in the data centre, such as application optimization, systems integration and application-level SLAs.
The integrator model is probably most flexible as it can use any equipment and any transport network, while the network centric model is most cost-effective and is delivered using the same delivery and support model as the Orange MPLS VPN. The data center model is most useful for companies that want to out-task some elements of their application management and can be usefully combined with either of the first two. Of course, this is just one way to deliver WAN optimization, but gives you an idea of how a service provider can help you meet your requirements determined in steps 5 and 6.
Step 8: Develop the business case
The final step in our performance management lifecycle approach is key for any project looking to get sign off from senior management: the business case. All of the previous steps feed into this business case, so it will include the KPIs, service management models and out-tasking choices. You need to weight performance and management costs and key issues include:
- Validating technology against business and technical requirements;
- Reactive or proactive monitoring;
- Reporting - technical and business-view;
- Network, IT infrastructure and application management;
- Required SLAs - from it infrastructure, network and up to the application level;
- Service management and on-going consulting.
Performance lifecycle management summary
So to summarize, to be successful in WAN and application optimization, you need to take the following eight lifecycle steps:
1. Map business needs to IT initiatives
2. Identify organizational needs
3. Identify applications
4. Identify key performance indicators (KPI)
5. Identify infrastructure constraints
6. Determine what IT activities are core to your business
7. Determine service model needs
8. Develop the business case
You can read more about my thoughts in WAN optimization and the service models outlined in Step 7 in this white paper, and please feel free to let me know what you think of the lifecycle approach in the comments section.