10 Steps to the Cloud

No doubt, adoption of cloud has yielded benefits for the companies, but at the same time it is not really necessary for all projects and may not be suitable for some at all.

For instance, keeping the RoI in perspective, if the existing services are meeting the business objectives and are cheaper than cloud then the migration is redundant. Cloud adoption makes more sense when it involves large scale of services that need flexibility or have variable/seasonal demand.

A business should opt for cloud only if the cost and efficiency post migration rules over the current services. Apart from choosing the right vendor and appropriate solution, it is equally important for organizations to analyze metrics for the pros of cloud computing, such as elasticity of computing power, leveraging already built-out, and established IT services, avoiding the costs of building and maintaining IT infrastructure. These are some of the factors that will ensure a successful migration.

In this relation, Dataquest sets the scene right for you by identifying the ways/means to deal with IT infrastructure while moving to the cloud.

Set Correct SLAs

One key requirement for implementing cloud is to develop mutually agreeable and clearly defined service level agreement (SLA) for demarcating general and high-level service offerings. Adopting cloud does not entail restructuring of the entire existing IT system/structure.

SLA contracts (internal or external) for cloud services are key compliance enforcement mechanisms and should be synchronized with an organization's unique requirements. The business will need to decide what features are critical for a cloud solution to provide, and what non-critical items will be there on the list.

SLAs of the cloud service providers need to be analyzed thoroughly by reading the fine print and all clauses (with regard to security, maintenance, and support services) should be clearly defined to avoid any confusion.

Performance of the cloud provider needs to be assessed through a small POC (proof of concept). For example, some providers provide 99.9% availability, while some provide 99.95% availability. The higher the availability assurance, the higher the price. You need to determine what the minimum level of assurance you need, perform a cost-benefit analysis, and make an appropriate choice.

Identify Criticality

Depending on the business processes, there are several cloud options to consider. For non-critical processes, a public cloud solution may be the fastest, most cost-effective solution. A private cloud might be necessary for a critical business process, especially for meeting the organization's security and regulatory compliance requirements.

For organizations which wish to move multiple business processes to the cloud, a hybrid solution may be the best way to leverage the efficiency of public clouds while maintaining corporate policy and regulatory compliance requirements with internal cloud solutions.

A semi-private cloud solution may be appropriate if like-minded organizations are willing to partner for an end-to-end cloud solution.

A SaaS solution may be the best fit if an organization wants to leverage an application provided by a cloud vendor.

Depending on the requirement of each organization vis-a-vis data management, storage, security, and compliance considerations, a right fit will have to be arrived at.

Pilot Framework

A cloud pilot project is a good way for an organization to evaluate cloud services to see how useful, reliable, and cost-efficient they can be. A pilot project should be a non-critical application that has limited scope, a short-time frame, and loosely defined estimates of its RoI. Building a pilot project has the added benefit of giving the organization a chance to learn how to use cloud services.

Scalability

Expanding businesses of companies of various sizes make the budget and IT requirement, critical points to consider. Expanding business makes enterprises realize that their current infrastructure no longer fully supports their evolving needs.

The need to consolidate data and optimize/cut costs leads to upgradation of their network infrastructure before implementing newer technologies like desktop or server virtualization. Establishing expectations upfront enables businesses to determine what they can actually afford and prioritize accordingly.

While opting for cloud, companies should provision for easy scalability and efficient management of resources keeping in view future projects. Scalability is not a default function when you deploy cloud, though PaaS deployments do come with some scalability from the existing environment. In order to scale up there is a need for applications to be placed behind auto-scalers within the cloud.

Back-up and Recovery

A data backup plan is a must for organizations and today, businesses of all sizes understand the repercussions of not having essential data always available. Issues like customers' unresolved problems, breach of SLAs, absence of suppliers, and degrading staff morale and productivity come to fore.

Senior management should consider the cost of repercussions, against the much smaller investment required to keep data and applications adequately protected on an ongoing basis.

Encryption and Compression of Data

Before putting data in the cloud, the company must use tools to compress and encrypt the files. This will not only save the storage space but also avoid redundancy of data. Additionally, tools available which ensure data is stored in the cloud only once to avoid duplication.

Plan Well

It is worthwhile to plan in detail the migration strategy and not underestimate the effort involved to thoroughly test critical enterprise application prior to embarking on this journey. Poor planning can cause disruptions to normal business operations that may prove very costly to recover. Besides, it would have severe impact on goodwill with customers, suppliers, and partners.

While deciding to let go off legacy applications, the transition needs to factor the technical, service, and commercial aspects of concluding the current hosting service. To this end, one needs to ensure that all customer data, software and operational requirements are safely stitched for the right user experience while adopting new cloud infrastructure.

Security

Prior to signing on with a cloud provider, an organization should conduct a thorough vetting of the procedures and processes followed by the provider. These include the fact that who manages/monitors the data; where exactly will the data reside; what are the protocols in case of emergency for data retrieval; in case of multiple tenants, what are the guidelines in place to separate 2 accounts; also the cloud provider should also provide details around security testing and audits, IDS, IPS, firewall, apart from clearly stating out the responsibilities and guarantees in the SLAs in situations such as Denial of Service (DoS) attacks that can greatly impair performance.

The applications might themselves contain several components, than would initially appear and some of the shared components cannot be moved to the cloud. So securing connectivity between the enterprise and the cloud needs to be established. Also, in the process, one must refrain from duplicating tools, processes or teams. Right to audit, ability to understand the impact of various regulations need to be well understood. Type of service undertaken according to the business need, equally needs to be mapped with respect to underpinning security controls.

Support System

It is very important for the provider to provide the relevant support services along with a clearer picture of the probable outages that may occur.

The vendor should have a qualified team of engineers who will be able to provide 24X7 support.

Support services are an important aspect while moving to cloud, and it would be wiser to choose a vendor who can provide the requisite bandwidth.

Cloud providers like Amazon Web Services and Microsoft Windows Azure offer a graded paid support service. Based on the criticality of the applications/data and the business needs, the right support option needs to be subscribed to.

Along with the IT infrastructure, it is equally important to design your support model. It is important to have various support layers from layer 1 to layer 3 or 4 as per the organization's requirements and capabilities.

You would need to plan for the geography spread and the support window. If you are looking at a 24x7 support, ensure availability of SME's (subject matter experts) based on the follow-the-sun-model and available on-demand based on elasticity factor. You would need to keep in mind the cross-section of security, operations, Infra, DBA, logistics, PMO, etc, and accordingly hire and schedule resources.

Administration and Maintenance

In-house IT administrations lack the expertise when it comes to managing cloud environments. Hence, it is imperative to streamline IT infrastructure management tools to manage more complex and multi-environment mix in association with the vendor. Additionally, standardization of processes go a long way in ensuring the cloud infrastructure growth is streamlined.

It is recommended that maintenance/upgrade windows are fixed before hand so that the necessary testing can be completed in advance ensuring stability in service. Furthermore, downtime for both maintenance cloud services and applications can be combined to cut down the outages for customers using the application.

On a Final Note

Decision makers need to carefully consider the benefits and pitfalls of utilizing cloud technology for managing IT infrastructure. Hence, it is highly important that a well implemented IT Infrastructure strategy should truly enable CIOs to get a centralized view of the entire business and cloud technology can play a critical role in driving this.

(This article is a compilation of views of various spokespersons and to view individual comments visit dqindia.com.)