Archive for May, 2011

Interop Advice: Match A Cloud’s Capabilities With Your Needs

Last week when I spoke at Interop in Las Vegas, I pressed home the ideas that matching a Cloud’s capabilities to your company’s requirements is the single most important exercise for a successful move to the Cloud.  Why?  Because all clouds are NOT created equal.

All clouds are not created equal  

I know for a fact that different clouds have different stacks, staff  levels, customer support expertise, recovery capabilities, security, and service level agreements, along with different operational processes and procedures to support those customer-facing functions.  But, as I said last week, a number of pragmatic metrics can help you evaluate a cloud provider.  I focused on the following three:

 High Availability.  It’s a misconception that all cloud platforms are highly available; don’t assume high availability.  Ask what is the cloud’s record for available, and ask how that number was derived. Consider what additional services will you need from the vendor to achieve the level of availability your applications need.

Recovery.  Who is responsibility for recovery?  Is recovery completely your responsibility or the vendor’s?  Or is it one of the many options something in-between when you and the vendor shared responsibilities. 

Service Level Agreement.  The SLA does not protect you against a service failure; however, it spells out what the vendor will do for you if service fails, whether that is a monetary penalty or
a re-performance of a service, etc.  You must assess the impact of a failure on your business and determine whether the remedy is adequate.  

Beware of SLA touting 100% uptime.  The fine print usually says “with the follow exceptions.”  No one can promise 100% uptime.  SLAs with teeth, however, indicate that the vendor takes the requirements seriously and expects to take measures to guard against the loss of a customer’s workload,

Enterprise Cloud

Without exception, it is an enterprise cloud provider that offers serious SLA commitments.  Make sure your vendor has a good track record and shares your concern for all aspects of cloud performance.  Then you have the potential for a long and trusted relationship.  

Do you currently have service levels assigned to the applications in your data center?

 Learn more at SunGard’s Cloud Computing Center.

Five Considerations When Evaluating Cloud Computing Architectures

An excellent starting point for an organization looking at cloud computing platforms is to examine its IT architecture.  Only by aligning the architecture – compute, network, data center, power and storage resources – with applications can a company be on the path to achieve the reliability and performance it requires within a cloud environment.

In cloud computing, true protection is an outcome of the right architecture for the right application.  Organizations need to fully understand their individual application requirements and, if using a cloud platform, the corresponding cloud architecture.  With that knowledge, they can make informed decisions about what cloud platform best meets the reliability and performance requirements of their specific applications.”

Here are five considerations for companies looking at cloud computing architectures.  

Availability.  Not all applications are created equal, nor are all cloud platforms the same.  Organizations need to tier their applications, identifying which applications need to be highly available, which can accept downtime and how much downtime is acceptable.  They need to understand the business risk associated with a lack of availability of their data.  For those applications that need to be highly available, businesses should consider enterprise-class technologies that have been rigorously tested versus looking at building something internally. It’s also important to look at multi-site solutions and disaster recovery/business continuity planning.  For most businesses, this means working with a service provider or consultant because they usually have access to greater levels of expertise and provide these services as their core business.

Security.  Security is still the primary concern for businesses regarding the cloud.  Concerns include the loss of control of their sensitive data, the risks associated with a multi-tenant environment, and how to address standards and compliance.  Organizations need to know how a shared, multi-tenant environment is segmented to prevent customer overlap.  How is the solution architected and is the service provider’s cloud infrastructure – network, virtualization and storage platforms – secure?  

Manageability.  Businesses need to understand what they are accountable for versus what they expect from a service provider.  Most public cloud vendors do not provide administrative support.  Organizations need to either have the technical expertise in-house to design the right solution or seek the services of an outside provider.  There should be an understanding of what level of management their applications require and have an identified change management process.  

Performance.  As with a more traditional hosting model, it’s important to understand workload demands on the infrastructure.  Companies also need to understand what the bottlenecks are and how the cloud architecture they have or are evaluating can meet those needs. Organizations should perform their own testing to understand how a cloud environment affects compute, storage and network resources.

Compliance. Organizations need to understand where their data will reside as well as who will interact with it and how.  They need to understand which areas of compliance the service provider controls and how to audit against the standards and regulations to which they need to adhere.

Successfully Transform Enterprise IT With The Cloud: A Pragmatic Guide

I’m proud to be presenting at the Interop Conference next week in LasVegas.  The session is titled “Successfully Transform Enterprise IT With The Cloud: A Pragmatic Guide”.  Below is the abstract and details;  I hope to see some of you there – CM
 
Abstract - As cloud adoption has become the hot topic in the Enterprise, there are many lessons to be learned as to how the early adopters have found success (and some failure) with Cloud projects.  Most enterprises where not built on applications and systems designed to take advantage of the web technologies that are driving public cloud’s rapid adoption.  That doesn’t mean IT has to sit on the sidelines.  There are pragmatic means to use the cloud to accomplish your business goals without having to overhaul the critical applications that are the life-blood of your organization.  This session will examine how and where enterprise IT can lower costs and improve the efficiency and performance of their projects through the proper use of Cloud services.

Thursday, May 11
Rm: Mandalay Bay L, 11:00am – 11:45am

PaaS: The cost saving “middleware” for cloud infrastructures

Today we hear from  Sarabjeet Chugh on  – PaaS: The cost saving “middleware” for cloud infrastructures

Not long ago, a survey of Fortune 100 companies showed that 77% of IT budgets go to maintaining the status quo.  Only 23% of the budget actually drives new revenue.  In recent years, a few dents have been made in IT costs by better development tools and clouds.   Development frameworks like Spring, Ruby on Rails, PHP, Python, Django framework, etc., let programmers code websites, web applications and web services more quickly, and clouds spread the cost of infrastructure components across multiple companies.

Nevertheless, infrastructure maintenance—at 42% of the budget—remains the single biggest component of the maintenance burden.  One thing that could make another dent in maintenance costs is an easy way to on-ramp an application into production in a cloud.  Getting applications to the cloud more quickly and deploying them with less programming to link the application to the infrastructure resources would decrease both development and maintenance costs.

New Middleware for New Architectures 

To do that, new middleware that works with the new hardware architecture of the cloud is needed. Existing middleware is antiquated.  Programmers spend nearly 50% of their time coding non-application functions, such as database caching, billing, metering, messaging and authorization. Different framework and database combinations need different versions of the middleware, and every version has to be maintained as databases and servers move within  enterprise data centers, public clouds or both.

The new middleware should be a PaaS element that is open and supports multiple programming frameworks, from Java-based spring to PHP-based micro frameworks and Microsoft .Net, among others.  It also needs to be independent of the infrastructure, so it can support environments from public clouds built using different hypervisor technology to local laptops.  Similarly, it should be independent of the application business logic, so the application is not muddies with the logic for addressing databases and constructing messages and, thus, is more portable.  Finally, it needs to include a reusable library of services that can be easily assimilated into new and existing application code to simplify the programming of 3-tier applications.

Accelerate Time-to-Market

The major benefit of PaaS is improved developer productivity and, therefore, an accelerated time-to-market. Organizations using PaaS techniques typically report operational savings of 30% or higher. 2011 is being termed as the year of the PaaS and for good reasons. Enterprise-grade IaaS has gained mindshare and acceptance in small-medium enterprises.  By leveraging PaaS, developers avoid the many hassles of updating machines and configuring middleware and can focus their attention on delivering applications. Reducing these obstacles means faster delivery of applications and making cloud portability a reality for enterprise applications.

How much time does your staff spend maintaining applications for infrastructure changes?

Building Cloud-friendly Applications

Today we hear from  Sarabjeet Chugh, Director of Technology Business Development (Cloud Services and Infrastructure)

Cloud adoption is progressing rapidly.  Many companies are in the process of determining their migration strategy, and most vendors are refining their processes to provide a smoother on-ramp to the cloud. 

Now that cloud is a reality, we need to think about how application development has to evolve to fit the cloud.  The application life cycle is broken.  Programmers write code, run tests and “throw it over the wall” to Operations, where technicians then struggle to accommodate the resource requirements of the application.

Old Code is Often Slow Code

Applications heavy from poorly structured code that requires multi-gigabytes of memory and have a huge storage footprint can run in the cloud, but the expense will become obvious.  Further, such applications offer few options and little flexibility to mitigate expenses. 

New Technologies Improve

A cloud-friendly application is one that can be deployed on any platform, locally or in the cloud.  To achieve such an application, new application development frameworks, such as SpringSource from VMware, have emerged that help to tease out the application’s business logic from underlying resource requirements. They also improve developer productivity by providing supplementary web services, message routing, authentication and application-level services, such as memory caching and contingency handling.

By insulating the infrastructure-dependent components and permitting them to be resolved in the production environment, the application can be more portable, reusable and maintainable.   For example, a cloud-friendly application could run in your data center but failover to SunGard if an incident occurs.  Similarly, server images could be transmitted to SunGard and brought up with full affinity and metadata information.

Does your company have standards for writing portable code?

 Download SunGard’s white paper,All clouds are not created equal.”