Joe Curtis

Amazon Web Services (AWS) has announced the general availability of Amazon EKS, its managed Kubernetes service – putting it alongside rivals Microsoft and Google in this regard.

EKS, which was announced at AWS re:Invent back in November, is now operational in US East and US West regions, with further expansion happening ‘very soon’, according to the company.

The move puts AWS alongside Microsoft and Google in terms of managed Kubernetes. Naturally, given Google originally designed the container orchestration system, there are no prizes for guessing that the latter has various solutions in place. Microsoft announced AKS (Azure Container Service) in October last year as a managed Kubernetes service, building upon the technology’s move to standardisation.

“Prior to Amazon EKS, customers either had to do considerable work to architect a highly fault-tolerant way to run Kubernetes, or just accept a lack of resiliency,” said Deepak Singh, director of AWS Compute Services. “With the launch of Amazon EKS, customers no longer have to live with either of those trade-offs, and they get a highly available, fault-tolerant, managed Kubernetes service. It’s no wonder so many of our customers are excited.”

Moving EKS to general availability does not mean AWS is struggling for customers in the meantime, however. A total of 25 companies were noted as being adopters of EKS in the press materials. One, GoDaddy, will be familiar to readers of this publication; the company specifically cited an active interest in containerised apps when they went all-in on AWS back in March.

Other companies confirmed as customers include Verizon – another who last month revealed AWS was its preferred public cloud provider – Snap and Pearson. Chris Jackson, director of cloud platforms at Pearson, said in a statement that the move to Amazon EKS ensured his team ‘built like a startup even though [it was] part of a major enterprise business.’

In some ways, the move could not have come sooner. According to data published by the Cloud Native Computing Foundation (CNCF) in December, Amazon EC2 and ECS continues to be the leading container deployment environment, with 69% of survey respondents citing it ahead of Google GCE/GKE (39%) and Azure (23%). The CNCF, which acts as a guardian for Kubernetes, ‘graduated’ the technology in March. Sarah Conway, CNCF senior director of PR services, said at the time the move signalled Kubernetes was ‘mature and resilient enough to manage containers at scale across any industry in companies of all sizes.’

Cloud has firmly secured its place on IT professionals’ agenda, with 95% citing it as one of the top five most important IT strategies today, according to the SolarWinds IT Trends Report 2018: The Intersection of Hype and Performance. However, migrating a database to the cloud is not a straightforward process. There will be challenges along the way. So, what do today’s IT professionals need to consider before embarking on their journey to the cloud?

Prepare for a new era of work

While cloud has many benefits, it also has greater complexity—and that complexity needs to be managed. IT professionals who have been tasked with devising new and creative methods to monitor and manage this infrastructure not only need to put the right solutions in place, but also need to prepare themselves (and the organisations they work for) for continued advancements. Moving databases to the cloud creates a new era of work—one that is more global, interconnected, and flexible than ever.

Mind the IT skills gap

The same report found that fifty-eight percent of survey respondents said an IT staff skills gap was one of the five biggest challenges they faced, meaning that organisations need to focus on improving and cultivating fundamental skillsets that will carry them into the cloud. This suggests that the skillset needed across IT is becoming increasingly blended, requiring versatile tech professionals that can adapt and flex in accordance with a changing landscape. Once you’ve got the right team in place, it’s crucial to give the migration project the attention it deserves and, by planning meticulously in advance (including a rollback strategy), you’re setting yourself up for success. The project plan must account for the scope of applications and related objects involved—and the number of stakeholders. By engaging the right people in the process, the migration should run more smoothly.

Create a single view  

By looking to their peer communities, IT teams will be able to better understand and adopt various technology adaptations and abstractions like software-defined constructs, containers, microservices, and serverless architecture. Considering the exponential rate of change happening in the industry right now, IT professionals need a management and monitoring toolset that gives a single view across those platforms. They need the ability to consolidate and correlate data to deliver more breadth, depth, and visibility across the data centre, which will allow them to proactively identify issues and resolve problems quickly.

Take your time and prioritise tasks

Many migration projects fail because the team has tried to do too much at once. For example, one team will be focused on the migration, but another starts deploying code changes at the same time. This often ends in disaster, as it’s not possible to rewind any issues with the new code or hardware. It’s vital that the workflow makes sense from the beginning to avoid having to retrofit a new plan of attack later.

Perform validation tests

Once the migration is complete, it’s vital to perform validation tests before handing it over for final testing. These validation checks will vary by industry. For some, they will require weeks of parallel production testing and sign-off on reports matching to the penny. For others, it will be as simple as executing a few queries. The important piece here is that there is agreement and signoff that the migration is complete, and that’s the validation check.

For such a well-trodden path, migrations to the cloud are rarely simple. However, when planned properly, with a little foresight and flex along the way, you’ll reach the destination in as good a state as possible.

Bobby Hellard

Microsoft has leveraged the technologies of submarines and renewable energy to plunge a 40ft-long datacentre into the sea near Scotland’s Orkney Islands.

The experimental shipping container-sized prototype, called “Project Natick”, is currently in operation on the seafloor next to the European Marine Energy Centre, just off the Northern Isles.

This Davy Jones’ datacentre is the result of a year’s worth of research into environmentally sustainable data storage technology that Redmond hopes could one day be ordered to size, rapidly deployed and left to operate at the bottom of the sea for years.

“That is kind of a crazy set of demands to make,” admitted Peter Lee, corporate vice president of Microsoft AI and research, who leads the NEXT group, a Microsoft research program for pioneering new technologies, which is overseeing the experiment. “Natick is trying to get there.”

Microsoft is pursuing what it calls a “relevant moonshot”; projects the company believes have the potential to transform the core of its business and the wider cloud computing industry.

The cylindrical storage container is loaded with 12 racks containing a total of 864 servers and has a self-sustaining cooling system that Microsoft has adapted from the heat-exchange process used in submarines. It works by piping seawater through radiators on the back of each rack to cool them down, before the now-heated water is expelled back into the ocean where it mixes with the surrounding currents.

An underwater cable from the European Marine Energy Centre on the Orkney Islands will power the datacentre. The energy centre is a test site for experimental tidal turbines and wave energy converters that generate electricity from the movement of seawater.

By plunging datacentres in bodies of water near coastal cities, data would have short distances to travel to reach coastal communities. It’s estimated that half of the world’s population live within 120 miles of a coast, so off-shore datacentres could lead to faster web browsing, streaming and a boost for AI-driven technologies.

“For true delivery of AI, we are really cloud-dependent today,” added Lee. “If we can be within one internet hop of everyone, then it not only benefits our products but also the products our customers serve.”

Microsoft hopes to leave the datacentre in place for five years without having to intervene – once submerged, it’s impossible for engineers to gain access to the facility. It will monitor its performance for 12 months to see if such an idea is practical.

Pictures: Microsoft

Everyone’s talking a good game when it comes to DevOps strategies – but automation is not as widespread in organisations as one might think.

That’s the key verdict from software management provider Puppet in its latest report. The report, State of DevOps Market Segmentation, aims to ‘reveal additional insights’ from the company’s most recent State of DevOps report, issued last year.

The 2017 analysis polled almost 3,200 technical professionals from organisations worldwide, and found that high performers have 46 times more frequent code deployments, significantly lower change failure rate, and 440 times faster lead time from commit to deploy.

Yet while there is an evident difference between the highest and lowest performers, the difference in low performers via industry is startling. The media and entertainment industry only had 23% who were categorised in the ‘low performer’ category, while financial services (54%), industrial and manufacturing (53%) and insurance (53%) were significantly more bottom-heavy in comparison.

Leadership particularly affects performance, as this publication reported analysing last year’s report noted – but in the highest cases, a good leader does not necessarily mean good practices across the board. DevOps success also required ‘suitable architecture, good technical practices, [and] use of lean management principles’, the report said.

The majority of respondents said they reported high levels of manual work across configuration management, deployment, testing, and change approval processes. Many organisations start their DevOps journeys at the areas where the pain is most acute; version control, continuous integration and infrastructure automation among others. Puppet says that while automation is ‘a key enabler’ across organisations’ journeys, the process remains ‘inconsistent and spotty’.

Perhaps the most interesting part of the study was assessing how expectations have changed in DevOps initiatives over recent years. “As practices become more widespread, expectations are rising,” the report notes. “What many might have considered ‘great’ IT efforts just a few years ago might appear as fair to middling today.

“That’s an interesting twist that suggests that the gains provided by DevOps – getting departments and teams to work better across an organisation – is no longer just a ‘nice to have’ but a given,” the report adds. “DevOps is simultaneously raising the bar and expectations of what’s possible.”

“Today, every company around the world has the same priority – automation at scale – and they’re achieving this through DevOps,” said Nigel Kersten, report author and chief technical strategist at Puppet in a statement. “While the data shows that companies of all types and sizes are making progress, we still have a long way to go to eliminate manual work that prevents companies from scaling automation success.”

You can find out more about the report here (email required).

Esther Kezia Thorpe

Bobby Hellard

In 1747 Lord Chesterfield used the immortal line ‘Take care of the pence; for the pounds will take care of themselves’ in a letter to a friend.  It still holds true today. The cost of a cappuccino and avocado on sourdough every morning on the way into the office soon adds up.

But however lax we are with our pennies, none of us would continue to pay the barista for that early morning caffeine hit if we decide to give up coffee. Oddly, though, that is what many large businesses do every day – albeit not for undrunk coffee but for unnecessary and unwieldy technology contracts.

Ghost spending and a spider’s web of contracts

Getting a firm hold on this ghost spending is vital as enterprises look to migrate to the cloud to boost agility and cut costs. If you don’t have full visibility of all IT costs at the start, then any cost savings identified as part of a cloud migration business case are going to be ‘finger in the air’ estimates. Furthermore, it will be hard to determine whether it makes financial sense to move specific workloads to the cloud, or leave them where they are.

Costs are rarely the sole motivation for moving to the cloud with the operating model, agility and saleability also influential. Despite this, the financial arguments tend to sway the board.

There are various reasons for the lack of a single, holistic view of IT expenditure. Sometimes the people who set up long-term contracts with suppliers have moved to new positions and no one has since thought to check the details. In other cases, individual business units may be merrily spinning up their own marketing apps in the cloud without the knowledge of IT teams, leaving trails of multiple cloud service charges in their wake.

Unravelling this spider’s web of contracts and other unnecessary overheads is made more difficult by the headcount cuts that have lacerated many IT departments over the last 15 or so years.  

During this time, cost optimisation has often taken a back seat but now it is coming back in vogue as organisations realise they need a digital clean up to ensure the commercial structure and cost base are fully aligned ahead of a drive to the cloud.

Digital feather dusters

While it makes financial sense to deploy new applications in the cloud, the decision for existing workloads is often less clear-cut. If you have a Technology Business Management (TBM) function, then the first step in any digital spring clean is to engage with them; if not, consider setting up at TBM with the help of a specialist third party.

Begin by undertaking a cost transparency exercise to establish a baseline. By breaking down the total IT bill into chunks – for example payroll services or email services – you can identify the share of these costs across the relevant IT towers e.g. networks or compute – making it easier to sniff out areas where savings can be made. Doing this, one global company managed to cut its annual network run costs by more than a third.

Another area that would benefit from a pre-cloud spring clean is technology demand. As business priorities evolve over time you will find you have unwanted services that can be ditched and low priority services that can be scaled back. Once this demand optimisation has been achieved the cloud journey can commence in earnest.

All technology expenses need to be reviewed. We find that organisations typically overpay their telecoms provider by up to 15%, often because they are being charged for services that they no longer use or have been decommissioned. In the cloud environment, the billing tends to be accurate but does not include a single view of resource consumption. Without this, it’s hard to maximise the true economic benefits of the cloud and it’s easy to end up paying for capacity you no longer need.

Modelling the cost savings

Once these baseline costs have been reviewed and adjusted, cost transformation can begin. A ‘before and after’ cost analysis lets organisations model different cloud options (e.g. IaaS, PaaS and SaaS), as well as single versus multi-cloud, and determine the most cost-effective approach.

We typically find that large organisations can cut IT costs by about 15% by migrating to the cloud or outsourcing alongside the benefits of agility, flexibility and scalability. In one example, a client made a saving of over £10 million a year on their run costs by moving test environments to the cloud. Mindful of the move away from sunk to variable costs, the team were also far more aware of the cost of spinning up every test environment and worked smarter as a result.

Watertight governance

Despite the positive steps taken by this particular test team to regulate usage, it’s essential to put in place watertight governance procedures. Only then can you ensure that everyone in the organisation is aware of the rules surrounding the use of cloud services and prevent money leaks.

Fine-tuning costs

As well as keeping an up to date record of all the software/hardware owned, it’s good practice to have a single view of all cloud service contracts. This becomes even more important in a multi-cloud model. With intuitive dashboards you can see what cloud services are being used where, and by whom at any time.

From here it’s a natural step to fine-tune the costs based on usage. In an IaaS scenario, for example, swapping out a 4-processor server for the same configuration in the cloud will deliver substantial savings as you tend to move from a fixed cost model to a variable cost model. By monitoring the peaks and troughs in resource usage, further cost savings can be identified – and an optimal profile agreed with your cloud provider.

Turning on a dime

There’s no denying the many benefits of moving to the cloud. But it’s impossible to evaluate the true cost implications without first having a single, granular view of all current expenditure. To gain the agility from the cloud that allows your business to turn on a sixpence, you first need to count your pennies.

Read more: Why organisations need 'reality check' on cloud costs

Cloud Pro