Multi-Cloud Deployments – Cloud Integration (2 of 2)

on Mar 13th, 2014 in Cloud, Enterprise Integration & IT Strategy, | Comments Off on Multi-Cloud Deployments – Cloud Integration (2 of 2)

On the technical front, cloud integration remains a top concern for IT organizations as highlighted by IDG’s recent cloud computing survey . Having bought into the premise of cloud computing, enterprises start deploying cloud services, hastily and with no clear strategy, to later find themselves faced with a hodgepodge of fragmented cloud and information silos – ultimately hurting business operations.

A well thought out integration strategy  is a must if IT organizations want to avoid ending up with a mélange of cloud silos connected through a “spaghetti” architecture that is inflexible, difficult to manage, and which can lead to chronic business disruptions.  An effective integration strategy for the multi-cloud world needs to address all integration scenarios and models– be it cloud-to-cloud, cloud-to-on-premise, application integration, or data integration. Quick point-to-point integrations are not viable long-term solutions and should be avoided — they often require custom interfaces, are fraught with problems and can be difficult and costly to manage and scale. Instead, IT organizations should build one cohesive integration platform that is service-oriented, supports standard-based web services, and addresses all cloud integration scenarios.  Such a platform has to provide full service orchestration, intermediation, and tools for monitoring and managing integration flows – in addition to secure means of accessing and transporting data between the enterprise and the cloud.

When it comes to how to implement an integration platform, IT organizations have multiple options to choose from, including on-premise software-based solutions, on-premise appliance-based solutions, and cloud-based solutions (iPaaS):

On-premise software-based solutions: This is the traditional integration server software from providers such as IBM, Oracle, and Informatica. Although originally intended for enterprise systems integration, most of these solutions have been augmented to address cloud integration scenarios –whether cloud-to-cloud or cloud-to-on-premise. These solutions do cost more but they tend to be more advanced, more stable, and provide richer feature sets.

On-premise appliance-based solutions: These are usually point-to-point integration-in-a box solutions; they are preconfigured and optimized for a specific integration use case, typically cloud talking to enterprise and vice-versa (e.g. IBM Cast Iron Systems).

Cloud-based integration solutions: Aka integration Platform as a Service (iPaaS) from the likes of Boomi, SnapLogic, and CloudWork. iPaaS is a lower cost alternative to on-premise solutions; iPaaS provides a suite of integration services which enable the implementation and management of integration flows between on-premise and cloud-based services, applications, and data. Some of the key features of iPaaS solutions include:

  • Pre-built integration and connectivity to cloud providers
  • Wide selection of cloud providers and platforms
  • No software development or technical skills needed to use the service

All those solutions have their merits as valid integration approaches and IT organizations need to consider their unique requirements and situation in deciding which one(s) to choose. Most importantly IT organizations should:

  • Refine and document integration requirements by analyzing the data and application flows along with their sources — both On-Premise and in the Cloud.
  • Consider internal resources, integration skillsets, and cost constraints as well as speed of development, testing, and deployment requirements when evaluating different approaches
  • Ensure that the selected integration solution is able to grow with your business — the solution should have the ability to efficiently and effectively scale as data volumes increase and additional applications are deployed.
  • After selecting and deploying the most appropriate integration solution (On-Premise, iPaaS, or Hybrid), make sure you have proper resource and support structures in place for continued effective management and monitoring.

Finally, although enterprises have no choice but to embrace a multi-cloud world (see part 1) – along with its challenges and complexities, they do have a choice when it comes to whom they want to delegate that responsibility to – either the internal IT department or an external cloud broker.  Ultimately, the final decision should reflect the long term vision for IT’s role within the enterprise – and consequently, where the organization wants to commit its finite IT resources and budget: low-value technology integration and operations or high-value business enablement and innovation – hard to do both in a do-more-with-less world.

What has been your multi-cloud experience?  Have you had success with any particular integration approach (On-premise, iPaaS, Hybrid)? Did I miss any important points and/or best practices? Let us know if you’d like to meet to discuss further with a Cloud Readiness Assessment.

Multi-Cloud Deployments – Proceed with Caution (1 of 2)

on Jan 28th, 2014 in Cloud, Enterprise Integration & IT Strategy, Managed Services, | Comments Off on Multi-Cloud Deployments – Proceed with Caution (1 of 2)

Multi-Cloud Deployments – Proceed with Caution

For the enterprise, the future of the cloud is in multiplicity. In their latest “state of the cloud” report, Rackspace has identified hybrid and multi-cloud implementations as  “today’s strategy of choice for the enterprise” given that the majority of respondents either already have an existing multi-cloud infrastructure (77%)  or are planning for hybrid clouds (47%). Although I agree with the overall premise of that conclusion, I am reluctant to accept that enterprises are adopting a multi-cloud “strategy” by choice – rather, I believe that enterprise multi-cloud adoption is driven largely by necessity and /or by “accidents” (aka unsanctioned “Shadow IT” clouds).

Today, the majority of IT organizations are struggling for survival, they realize that in order to stay relevant they must transform from a mere technology integrator to a business innovator and enabler. As a result, they have very little desire to be sucked into complex and resource intensive cloud integration work and would happily settle for a single cloud – it is simpler to manage and quicker to provision. On the other hand, enterprises also recognize that the cloud is not a one-size-fits-all and that their requirements cannot all be met by a single cloud provider and/or cloud deployment model. Moreover, most of them already live the multi-cloud reality: a fragmented cloud applications environment born out of hasty cloud deployments – either by enterprise IT or lines of business. So, multi-cloud deployments are, and will continue to be, the norm and IT organizations have no other option but to embrace them.

As enterprises start wrestling with multi-cloud deployments they are bound to come to a few important conclusions:

Multi-Cloud is Worth the Effort

Adopting a multi-cloud deployment can help organizations attain an effective cloud strategy. A multi-cloud infrastructure can improve business continuity through geographical and platform diversity, can enhance a solution’s resiliency and cost effectiveness, and can help avoid vendor lock-in while leveraging vendor competition to achieve better features and cost options. In addition, the ability to distribute and move workloads across cloud providers and deployment models (private vs. public) enables organizations to quickly adapt their applications architectures based on risk, cost, and efficiency constraints.

Multi-Cloud Deployments are Complex

Multi-cloud deployments are complex and don’t always return value that is comparable to their level of complexity – So, achieving a positive ROI from multi-cloud deployments requires greater effort and substantial planning. Consuming cloud services from multiple providers presents a unique mixture of commercial and technical challenges.

On the commercial front, IT organizations have to administrate and integrate multiple relationships that come with disparate cloud providers along with their respective SLAs and pricing models. Multi-cloud deployments and the expanding value chain of cloud services further complicate SLA management – it is not uncommon for cloud providers to rely on each other to deliver a portion of their cloud services (e.g. a SaaS provider sourcing its own cloud infrastructure from a third-party vendor). When dealing with cloud SLAs It’s important to focus on end-to-end service delivery and strive to manage the SLAs across the cloud provider’s supply chain. In addition, since service delivery depends on multiple providers, standardization of terminologies in the SLAs is critical for success. Finally, it is important to point out that consuming services through a multiplicity of cloud providers (especially within one application stack) can render responsibility boundaries (client-provider and provider-provider) very blurry. As a result, IT organizations should focus on establishing greater clarity of ownership and accountability across cloud providers that are part of a service delivery chain.

Stay tuned for part 2 where we’ll look more closely at multi-cloud integration. In the meantime get in touch with us if you’d like to discuss cloud readiness.

Best Practices for Planning and Deploying Cloud Services

on Oct 8th, 2013 in Cloud, Enterprise Integration & IT Strategy, | Comments Off on Best Practices for Planning and Deploying Cloud Services

Planning and deploying cloud services can be a daunting and confusing undertaking with multiple options available at different decision points. To succeed in this task, IT organizations should strive to formulate the right approaches to address the following key areas: business user engagement, deployment strategy, application selection, cloud model selection, and performance metrics.  Without further ado, here is some guidance on how to accomplish that:

Partner with Your Service Consumers

A cardinal sin that IT organizations are guilty of, consistently, is that they tend to ignore the IT service consumer – at their own peril! IT organizations are notorious for failing to effectively capture end-users’ expectations and service requirements and as a result their IT services deployments are doomed before they even get off the ground.

Identify and engage with you consumers and stakeholders early in order to clearly define the need, refine the requirements, and to collaboratively build an acceptable solution.

Test Drive Before You Buy

Don’t expect to deploy cloud services and plumb them into your enterprise fabric overnight and without proper due diligence. A gradual and progressive approach is more effective as it gives you some leeway to adapt and evolve based on your users’ needs.

Start with a Proof-of-Concept/Value to gain experience, evaluate new cloud solutions, and mature organizational understanding of cloud computing. However, don’t turn your PoCs/Vs into marathon projects and risk losing momentum — Keep them light and short (30-60days) and make sure you evaluate key aspects of you cloud service like: cloud model fit, usability, scalability, manageability, security, and integration.

Tailor Cloud to Your Needs

Your cloud infrastructure architecture will ultimately be driven by the business and application requirements not the latest & greatest cloud craze. Most probably you have requirements that cannot all be met by a single cloud model and/or cloud provider. So, don’t try to fit all your IT in one cloud, focus instead on identifying the best-fit cloud model/provider for each service/application.

Consider private clouds for workloads that have high compliance/security, performance, and /or SLA requirements. On the other hand, use public clouds for your low-risk workloads, whenever you can – public clouds are cheaper, more scalable, and offer better automation. In addition, expect to invest in good cloud integration expertise to connect and manage your different cloud environments − If you can’t do it in-house; consider partnering with a cloud services broker for help.

Pick the right application

Be judicious with your application choices for cloud migration – the cloud is not a good fit for all applications. Invest time in conducting a cloud readiness assessment across your application portfolio to identify business and technical risks and constraints. Also, start with simple use cases and build momentum to tackle more complex use cases later on.  In general, Good candidates for cloud deployment are applications:

  • With low SLA and/or security/compliance requirements
  • That require rapid scalability
  • That are nearing a technology refresh and/or end of an ITO engagement
  • That have variable capacity needs (e.g. Compute, Storage)

Manage to KPIs,

Before you embark in your CISS journey make sure you have a good set of KPIs to track and manage your program’s performance – That is critical to ensuring long term success and business sponsorship. Some key KPI areas to consider are:

Business Agility — How quickly are you able to respond to business needs/opportunities?

E.g. New application/service launch; Application/service modification; Service scale up/down; M&A integration time

Cost Efficiency – How cost efficient is your cloud infrastructure vs. data center infrastructure?


Quality of Service: Did you improve the end-user experience?

E.g. Better support quality; better SLAs, improved application usability/performance/functionality

Those are some best practices that will hopefully help you design and execute your cloud infrastructure strategy. In my next blog I will talk about cloud integration and how you could effectively plumb together your cloud and non-cloud environments in a hybrid IT environment – the new norm!

IT Shared Services: Elusive But Not Impossible

on Sep 10th, 2013 in Cloud, Enterprise Integration & IT Strategy, Servers, Storage, and Software, | Comments Off on IT Shared Services: Elusive But Not Impossible

Implementations of IT shared services often reveal a tale of two conflicting camps. On one hand, an IT leadership team mainly interested in realizing the financial benefits of shared services irrespective of end-user concerns, and on the other hand a business user community deeply attached to its autonomy and diligently resistant to change regardless of the financial upshot.  That is the dilemma that IT organizations have to solve in order to succeed in their shared services initiatives.

The US federal government is learning that very same lesson as its data center consolidation program, which claims $2.5 billion in savings over a three years time span, struggles to get off the tarmac. It does not surprise me that the feds are having such a hard time herding hundreds of agencies to reduce 2,100 data centers (that number seems to be changing by the day) down by 38 percent and to consolidate a whopping 1,979 redundant silo applications. Over the years, the federal government has grown to become an “everglades” of custom-made application stacks — why do you think Washington DC is the 2nd best city for tech jobs, which further reinforced the agencies unique operational realities. The flagship challenge facing shared services implementations is how to reconcile the needs of entrenched silos with the one-size-fit-all approach. Addressing that challenge requires IT leaders to be more effective in managing change and at incentivizing users to embrace it.

Relying on financial benefits as a main incentive is a simple-minded approach to driving shared services adoption. Moreover, it can alienate business users and torpedo the shared services initiative. Shared services implementations that promise cost-effective but rigid blocks of services with long delivery timelines and little room for customization will fail to win the buy-in of business managers who are worried about losing control (aka: flexibility, speed, and agility) over their applications and being forced to accept less sophisticated alternatives. As a result, disgruntled business managers engage in covert efforts to deploy point applications outside of IT’s radar – welcome to the world of shadow IT! Today, such clandestine attempts have become increasingly pervasive, and harder to detect thanks to cheap and instantly available cloud services. The growing “consumerization” of IT and prolific adoption of cloud computing are driving an unsanctioned cloud applications sprawl within today’s enterprises hence challenging IT’s ability to pull a difficult balancing act: effectively govern enterprise technology use while remaining relevant. However, the cloud is not only about challenges but it is also about opportunities and enabling capabilities which, if properly leveraged, can help IT organizations bridge the gap between IT and business users.

Cloud centric shared services can offer businesses better economies-of-scale cost efficiencies and can effectively address users’ quality-of-service concerns through improved agility and scalability, a flexible consumption-based chargeback model, and a marketplace abundant with customizable offerings. Thus, cloud capabilities should be viewed as an indispensable ingredient of any IT shared services initiative. Nevertheless, leveraging cloud capabilities successfully can be a difficult task especially for understaffed and/or under skilled IT organizations.

The cloud computing paradigm shift raises unique and challenging questions which need to be addressed through a strategic plan. Today, the data center as a physical brick-and-mortar asset is morphing into a virtual fabric of loosely coupled services; that presents a new set of problems to be resolved: How to effectively aggregate, integrate, and deliver a hybrid set of cloud-based and on-premise IT services? How to select, prioritize, and migrate enterprise workloads to the cloud? How to manage and normalize SLAs, contracts, and billing structures across multiple cloud services providers? How to effectively govern and evolve a cloud centric shared services platform?

Resolving those critical questions would certainly be a daunting task for many IT organizations. Nonetheless, IT has no other option but to embrace shared services because its future does not hold many other viable alternatives. To successfully navigate this transformative journey, IT leaders need a strategic plan, mature best practices, and an integrated approach that accommodates the business appetite for risk and pace of change.

In the coming weeks, we will discuss how to plan and implement cloud infrastructure shared services. What have been your experiences and what would you advise others on the lessons learned?

Powered by WordPress