What is Happening?
Yesterday, Saugatuck attended the AWS Summit 2015 in San Francisco, where Amazon gave an update on their business, and released several new products to the nearly 10,000 attendees at the Moscone Center and the 7,000 who watched the livestream. Andy Jassy, SVP of Web Services at Amazon kicked off the keynote by highlighting some key statistics about the business: from 4Q13 to 4Q14 they experienced 103 percent year-over-year growth in the amount of Data transferred into and out of S3 (Simple Storage Service) and 93 percent growth in the use of their compute service EC2. They now have over 1 million active users who have used the service in the last month.
Jassy brought several companies out to discuss the value of the AWS Cloud infrastructure. Jason Kilar, the founding CEO of Hulu, and now CEO of Vessel, a startup focused on video sharing and consumption, highlighted the ability to keep his team small, and focus on the business without having to build infrastructure. Wilf Russel, VP of Digital Technology Development at Nike described how the Cloud has fundamentally changed their application architecture and described their shift toward DevOps and Microservices. Valentini Volonghi, CTO at AdRoll discussed how the Cloud gave their business the reach to reduce latency by distributing their app around the globe. And Colin Bodell, CTO & EVP at Time Inc. who is migrating all of Time’s datacenters to AWS, citing that in the UK, they took their datacenter monthly run rate from $70,000 to $17,000.
Finally, Jassy used the opportunity to make several product announcements:
- Amazon Elastic File System (EFS) – a fully managed filesystem that can span multiple EC2 instances to enable multiple concurrent and scalable connections to a single file directory.
- Amazon Machine Learning Service – A one-size-fits all service that enables non-experts to implement Machine Learning algorithms on their data sets, or within their applications. This offers a simple API for training and modeling calculations which can then be called to perform specific machine learning tasks.
- Amazon Workspaces and AWS Marketplace for Desktop Apps – an Amazon VDI product to enable companies to create virtual desktops, as well as purchase, manage and provision the software for those desktop applications.
- The GA release of the Amazon EC2 Container Service – The Amazon tool for deploying, managing, updating, and running Docker clusters.
- The GA release of AWS Lambda – a Service designed to perform trigger / event driven compute in small doses. Provides a way to perform small, scripted tasks in real-time when triggers are initiated. Amazon highlighted its use in sending notifications, indexing, IoT, and as a serverless mobile backend. At present, Lambda only supports Node.js, but is adding support for Java in the coming weeks.
What is Happening?
Earlier today, Saugatuck Technology released the findings from its just completed Cloud Infrastructure Survey. The research and analysis clearly shows that businesses are moving rapidly away from traditional On-premises systems toward a range of Cloud infrastructure alternatives – including Internal Private Cloud, Hosted Private Cloud, Public Cloud and Hybrid (On-premises + Public Cloud). While CRM, HCM and Marketing-based SaaS solutions have dominated early Cloud decisions deployments, and more recently Cloud-based Finance offerings have begun to gain traction – the growing migration of On-premises production workloads to the Cloud, as well as the creation and deployment of Cloud-native production workloads clearly shows that we are entering a new phase in the transition.
Across major infrastructure services, both in the Cloud and On-premises, companies indicate a broad desire to upgrade their capabilities over the next two years. Additionally, On-premises virtualization – the dominant platform for IT Infrastructure today – will be supplanted with a combination of both Internal and Hosted Private Clouds, often supported by next-gen containerization technologies. While very few companies expect to be running entirely on Public Cloud by the end of the decade, a combination of Private and Public Cloud infrastructures (supporting production workloads), along with publically-available SaaS solutions (across an array of functional domains) will become increasingly the norm.
These are just a few of the conclusions from Saugatuck’s 73-page Strategic Report, released earlier today (Next-gen IT – Cloud on the March, 1553SSR, 02Apr2015). The report leverages a global web survey of 327 senior IT executives, across major geographic regions and business sizes. Continue reading
Software Defined Networking (SDN) is emerging as a new architecture for infrastructure flexibility in enterprises and in providers of Cloud infrastructure offerings. SDN is based upon separating the control layer from the data transport layer in a communications network. A key advantage is the use of a virtualized architecture that enables the use of commodity servers rather than custom-designed network controllers.
Some enterprises have taken steps to adopt software defined networking in their data centers, where it provides cost advantages and more flexibility in network configuration, and thus can accelerate implementation of new applications through DevOps techniques. For example, using DevOps, an application can be readily tested by temporarily accessing cloud resources to run the application at scale.
While enterprises have learned that a key attraction of virtualization is that it provides a means to make changes to an infrastructure easily and quickly – SDN presents a dilemma for vendors of traditional networking controllers. SDN enables the use of commodity-priced servers rather than conventional proprietary controllers in network infrastructures. This is quite attractive to enterprises due to the potential for significant cost savings (see Figure 1), and due to the greater flexibility in equipment and vendor selection. Continue reading
What is Happening?
Over the last several months, Saugatuck has been looking into phenomena increasingly (and commonly) referred to as “Hyperscale”. Hyperscale appears to be entering common parlance just as Cloud became increasingly well-understood and then mainstream over the past five plus years. Common definitions of Hyperscale focus on two major factors: the speed of scaling resources allocated to a task; and the extent of the resources available in the infrastructure.
However, just as Cloud offerings today don’t look like the Cloud offerings of 3 years ago, the requirements and usage of infrastructures that can be termed Hyperscale are evolving rapidly. Not too long ago, a primary value proposition for Cloud infrastructure adoption was economic, due to improved hardware utilization through automation of resource provisioning / allocation to individual workloads. This same automated provisioning is now viewed as a significant boon to IT and Business agility, not just a means to save money.
Similarly, Hyperscale has been viewed as a requirement for some types of workloads and as a foundation for some Cloud providers. However, Saugatuck projects that Hyperscale will increasingly become recognized as a key enabler of new types of application design, and broadly deployed among a much larger pool of Cloud infrastructures. Continue reading
The on-demand Cloud model is beginning to displace traditional outsourcing. But outsourcing has developed over many years alongside industry, and it has had an important impact in shaping service customer expectations and supplier contractual obligations. Cloud services do not readily fit the mold. While simple or limited Cloud is seldom an issue, the widescale replacement of traditional in-house infrastructure is beginning to bring new concerns into focus. This is particularly important for the evolution of IaaS and PaaS.
Although Cloud is simpler in its agreements than its predecessors and envisions a utility-based model, it must still fit in with the auditing, technical integration, and governance needs of the firm. Traditional outsourcing models have tended to focus upon the Service Level Agreement (SLA) for contractually defining what is to be provided and how. However, the SLA has proven to be a thorny issue in the Cloud environment, particularly in public Cloud services. The issue has focused upon the definition of downtime, such as might be experienced when a Cloud provider goes offline for several hours—as has happened several times recently. What constitutes downtime, and what penalties (if any) should be paid to the client when this affects business? But this masks a range of related contractual concerns.
Cloud services need to be fitted into an organization to be used effectively. This means that they need to be integrated with other suppliers in the evolving multi-vendor environment. Management of vendors has centered upon the Vendor Management Office (VMO), which is increasingly being called upon to deal with services originating in the Cloud. But these services have different characteristics from their predecessors, different lifecycles, and different procedures for onboarding and releasing. Continue reading
Confronted with the requirement to cost effectively increase infrastructure capacity and/or flexibility, IT organizations strive to choose between traditional infrastructure providers, CoLo and hosting providers, and Cloud providers. In a recently published Strategic Perspective, Saugatuck offers guidance to any IT organization striving to plan an IT infrastructure.
Saugatuck urges that prior to initiating evaluations of technological infrastructure alternatives, an IT organization should first perform a thorough inventory and assessment of the existing physical infrastructure. This effort can be logically divided into two categories: Continue reading
The rapidly developing API Economy is made up of many parts that are infrequently examined. APIs provide access to software services on a network, and they have existed since the early days of computing. Now they are garnering new attention as Mobility and Cloud make them more important and more accessible. Releasing APIs has become an important marketing strategy for many companies, and their usage is encouraging integration of innumerable services through mashups on the Web. Yet we seldom look at specifics of what is being released. As APIs become more important, they signal directions that development is likely to take, and herald new opportunities from bringing together previously unavailable services. They also demonstrate the increasing importance of Mobility in driving software development. Continue reading
Mobility is having a large and diverse impact upon the growth of Digital Business. In a previous Perspective we looked at the Mobile/Cloud impact on Innovation (Enabling Innovation with Cloud Mobility, 1455STR, 10Oct2014). But the impact goes farther than that. Mobility is changing the way we think about IT and immediately enabling a wide variety of new possibilities – including entirely new concepts based on mobile sensors and their deployment. In the fast-moving environment of app development, evolution occurs swiftly in a continuous battle for survival of the fittest. Device desktop space is limited, and there are hundreds of apps performing similar functions; apps are quickly developed and deployed, and they can include powerful integration with Cloud-based services. And they are easily repurposed and moved to a new arena. Continue reading
Attraction of Cloud IT offerings remains mostly based on the perceived potential of reduced infrastructure costs. However, the trade press and IT analysts provide minimal guidance on sizing Cloud resource for a workload. Saugatuck’s ongoing conversations with users have shown that Cloud IT cost estimates are typically based on gross and erroneous approximations rather than on measurements of performance.
Saugatuck recently published a Strategic Perspective which focuses on factors which heavily influence performance of a Cloud IT infrastructure and, thus, the ongoing costs of running a workload on the Cloud infrastructure: Continue reading
During the week of 29 April through 3 May IBM hosted its annual Impact 2013 (subtitled “Business. In Motion”) conference in Las Vegas, Nevada. Befitting the location, Impact is a huge business conference (with 8,500 attendees this year) and has a definite Las Vegas flavor (the keynote opened with a car driven on stage and dubbed a “rolling data center”).
IBM’s market positioning and messaging continues to be a bellwether for large-enterprise IT trends overall. This year’s Impact emphasis on the combination of Cloud, Mobility, Social, and Analytics – plus the importance of native integrative capabilities – underscores the emergence and reality of the loosely-coupled, “elemental” IT and business architectures that are increasingly prevalent, that enable a variety of Boundary-free Enterprise™ configurations, and which are reshaping providers’ go-to-market positioning.
Note: Ongoing Saugatuck subscription clients can access this premium research piece (1216MKT) by clicking here, and inputting your ID and password. Non-clients can purchase and download this premium research piece by clicking here.