Complacency at AWS Summit

Sitting there last week at the AWS Summit in London, I was struck by how similar Werner Vogel's message was to Googles at the GCS Next conferenece. It was just that AWS's CTO was far less punchy in his delivery and much more obsessed with technical detail. At the same time there was no recognition that AWS might actually have competitors. 

It is true that Amazon, with its AWS service, is far in front of Microsoft and Google in terms of richness of its technical offering and market penetration, but at the same time this gap is closing. Recently, I saw a great conversation on LinkedIn started by a recruiter for cloud specialists who had noticed a growing demand for Azure specialists, having spent the previous year only recruiting AWS staff. This shows, that in large corporates at least, a dual market is naturally emerging as Microsoft leverages the installed base of AD and Office products to move them to O365 and Azure. Also, I did not see the sort of financial innovation that Google is bringing to the market with its highly flexible pricing options.

In fact it took me a while to identify what differentiates AWS from its closest competitors. As much of Werner's pitch was spent identifying improvements to current offerings within its services. The things which stood out were:

• AWS has more virtual server type offerings than its competitors, with some services now into their 5th release or so of maturity. Although this comes at the downside of increased complexity and need for an intepreter to translate for the uninitiated as many of the services are labelled in the format AN (A = a character and N = a version number) in the manner of a technical labelling convention rather than a meaningful name.

• AWS is now offering "serverless computing" via its Lambda service using functions instead of virtual servers. Whilst this obviously offers a great deal of agileness for rapid delivery, it may be risky in terms of encouraging sloppy development resulting in monolithic resource hungry applications which are expensive to run and difficult to maintain. It also may present a new form of vendor lock in risk, as migration to other cloud platforms in the future would be less easy.

• AWS is now offering a niche service for FPGAs. Which allows a mature and tuned application to be "burnt into hardware" with corresponding double digit improvements in performance. This may be important for people with well proven machine learning applications

• The suite of security services available is also impressive with extensive DDoS capability embedded as standard.

• AWS now has a UK point of presence and others are being established throughout Europe to ensure that data privacy and export sensitive security concerns can be addressed. Although global coverage is still a little patchy and the Middle East in particular seems poorly served.

• AWS's IoT offering also appears to be maintaining its edge. It's still the only one to have a rules engine built in and additionally, Amazon is now launching a standalone environment which can be run on the Things (i.e. the intelligent devices which are networked in IoT networks) which is consistent with the rest of the service.

For me the most interesting part of the day was visiting the start-up partner stands and a presentation on Amazon Launchpad. Launchpad is its service for helping "ready to go to market" startups to promote themselves and gain extra visibility in the online market place, coupled with help in delivering their on-line presence. This is currently focused on "tangible product" companies, but may be extended to services in the future. Noticeable amngst them were Cocoon (a company which provides remotely manageable security devices based on infrasound), Beeline (a company which provides simple digital navigation devices for bicycles) and Roli (a company which makes electronic music generation simple and accessible).

So, to recap and summarise, AWS maintains its technical edge, but appears over comfortable and complacent. It needs to hone its marketing and value proposition to continue to remain relevant in the future.

Post Digital Outsourcing - Going for Value

One of the things which always has frustrated me when dealing with traditional outsourcing companies has been the huge gulf between the Sales Proposition and the Reality of Day-to-Day Service.

Companies usually think they are going to buy the best skills in the market and get access to superlative service from a transformational partner who delivers agility and control, so they can forget about the complexities of running IT. In reality they get bland, slow moving and unimaginative bottom up services which are slow and expensive to change. Where people have looked for partnership, this has usually failed to materialise as relationships descend into "Robust User-Supplier Conflict" (or RUSC). In fact the traditional model has been an anti pattern to progressiveness and this has only been made worse by offshoring to Asia, where extreme cultural differences and expectations have only made things worse.

As I mentioned in a previous blog, The Death of Outsourcing, this model is no longer sustainable in the light of all things digital. So what can a Post Digital Service Provider offer now that XaaS threatens to steal outsourcing's breakfast?

The reality is that, whilst XaaS does a lot to free an enterprise from the Tyrrany of Infrastructure, XaaS introduces new complications as the overall technical environment is much more complex; Digital also means that enterprises need to learn to move quicker with disciplined lean practices that enable continuous change. New IoT based models also bring new challenges of scale.

So the new breed of Digital Service Partner needs to position itself to avoid RUSC and focus on Assured Value, Integration and Speed. Where: 

• Assurance covers secure and consistent delivery of change and operations;
• Value focuses on user and customer experience, insight and the right quality at an affordable price;
• Integration deals with the complexity of identity and integrating multiple sources of XaaS services as well as working with multiple SI partners and in-house development teams to deliver joined up services;
• Speed addresses agility, continuous change, innovation and responsiveness to changing business and technical opportunities and risks.

Such services are likely to include 3 key elements:

1. a Foundational Use of IT Access Service - everything needed to deliver a user device centric service, e.g. smart phone, pad, laptop and supporting network, gateway, office automation software, storage, print, peripherals and anti malware based services.
2. a Service Integration And Application Management Service (SIAM) - which integrates and delivers services on a Hybrid Cloud basis. This is likely to include Architecture Management (AMO), Programme Management (PMO), Service Delivery (SMO), and Security Operations (SOC and Security Operational Processes), as well as Activity Based Costing (along TBM or OBASHI lines).
3. a Lean System Integration Service - which can provide specialist development and implementation skills, but also embraces partnering with internal and 3rd party partners and provides support for the full range of Agile and DevOps processes needed to deliver continuous change.

Naturally there will be other more specialist services which may come too, e.g. computer forensics and advanced threat intelligence, Fleet Management for IoT devices, or managing innovation communities as innovation goes social. But these will be value adds building on a core foundation.

Google The Rainmaker

On a day when CIO Magazine reported on an Uptime Institute study indicating that only 13% of enterprise loads are operating in the Cloud, Google is holding its Google Cloud Next conference in London.

The key message is that Google is in the cloud hosting game for the longterm and is offering to partner enterprises through their Digital Strategies and to help them unlock the value of their data.

One of their most impresive case studies is helping LUSH the bathroom smellies company to move its global operations to Google Cloud Platform in just 22 days.

Otherwise Google is trying to make a compelling case for its offerings in terms of flexibility of scaling an pricing, the richness of G suite in the collaboration and app devleopment space, AI and bot support as well as data management and Big Data agility. All backed up by rock solid security and sharp performance. 

However, for small companies and start-ups the affordability of its office apps, ubiquity of platforms and flexible pricing makes GCS a "no brainer" for consumer facing brands.

So rather than just trading on its born in the clouds heritage, google is fully positioning itself as a Cloud Builder and a Digital Rain Maker.

IoT Platforms

Companies going down the cloud bases PaaS route for hosting their applications have some interesting choices. There are 2 main leaders: Amazon, Microsoft and Google. There are also a lot of other platforms built around major applications, e.g. Salesforce & SAP, or technology stacks, e.g. IBM and Oracle.

Most of these platforms have quite features around provision of virtualised servers and storage as well as load balancing, with extensive options for scalability, as well as pricing models. All come with various database management system services as well as data analytic services for BI/Big Data usage.

Whilst Gartner makes a great show of assessing them against its own set of  Enterprise requirement criteria, this is unlikely to be meaningful for long as the leaders are engaged in an arms race to introduce an increasing number of features and capabilities which means that any 3rd party analyst's assessment is bound to be out of data almost as soon as it is published.

Enterprises have a fair guessing game about which platforms are going to be dominant in the future. This is almost impossible to get right. So more pragmatic approaches are needed. If an enterprise intends to move almost everything onto cloud platforms, then some analysis of what services its main SaaS applications uses may be appropriate. For companies tied into .Net, Microsoft Office and AD, then Azure may be a no brainer.

However, when it comes to IoT based applications, this may not be so simple. At present Amazon appears to have the leading  IoT support framework of the big 3 platform providers. Google and Microsoft appear to be trying to get in on the end device with specialist operating system offerings, so that they can own the whole stack. Likewise, Oracle is aiming to lever its Java specialism with its technology stack to provide specialist SaaS applications which facilitate rapid development in the IoT area.

My take is simple. For now anyway, most applications are going to have to deal with at least 2 PaaS platforms. One for internal applications and a second for externally facing applications and IoT. In reality, most corporations may need even more, especially if they want to exploit big application services such as SAP's and Salesforce's.

The End of Digital Adolescence

Are we growing from just talking about it to doing it?

Over the last 18 months I have attended a number of events with CxOs and other senior stakeholders from many different companies.

A key theme has been that we are all being pressed to do something, as we all work in organisations where customers, employees, business partners and senior managers expect us to be doing something and most of us have.

A key concern has been that we are all scared that we have missed something. Is there an "Unknown Unknown" that will emerge to destroy the new value that we are trying to create. We have all been thinking a lot about the subject and I think that collectively we have come to the following conclusions:

The 3 technologies that we have to get to grips with are:

• Identity Management (and subscription)
• Encryption
• Integration

The things that we should worry less about are:

• Security of the various PaaS and IaaS offerings, as the vendors who supply them spend a lot more time and money securing them than most user enterprises can dedicate or afford;
• Traditional technology selection approaches and worries about vendor lockin - the richeness, utility and value of the continuously evolving offerings obviates the need.

The things that we need to get good at are:

• DevOps - so we can move at Digital Clock Speed
• Service Integration (or SIAM) - so we can run this seamlessly from end-to-end
• (agile) Enterprise Architecture - so we don't lose track of what we've got (where we are spending money) and what we want to achieve in the future
• Security Governance - again so that we

The conversations that we have with other stakeholders in our businesses should focus more extensively on Business Value, rather than infrastructure maintenance and "keeping the lights on". But we also need to establish a different approach to projects, applications and investments, as the traditional ROI based Capital Appraisal, Invest and Forget model does not fit the continuous evergreening needed to sustain Digital Assets and keep them relevant in the face of customer demands.

There are plenty of other things as well, each worthy of a blog of its own, but this is the gist of all these discussions and power breakfasts.