23 Sep Cloud Solutions: Azure vs AWS vs the rest
Cloud Solutions Fundamentals
Cloud computing brings on-demand computing closer to reality. It changes the way IT hardware and various applications are purchased, designed, deployed and used. With its capability to scale and a pay-as-you-go pricing model, the primary benefit that cloud services extends to organisations is greater business effectiveness at lower IT costs.
Typical Cloud Services:
- Business Process-as-a-Service
- IOT workloads
Dominant Cloud Providers
A 6-horse race:
Analyst firm Gartner’s 2019 Magic Quadrant for infrastructure as a Service (IaaS) has again found that Amazon Web Services and Microsoft Azure are the most mature clouds, but has omitted more than half of the vendors it covered last year on grounds that customers now demand more than just rented servers and storage.
“Customers now have high expectations from their cloud IaaS providers. They demand market-leading technical capabilities — depth and breadth of features, along with high availability, performance and security. They expect not only ‘hardware’ infrastructure features, but also management features, developer services and cloud software infrastructure services, including fully integrated PaaS capabilities“.
Given those expectations, Gartner dropped eight clouds from this year’s Quadrant, saying farewell to Virtustream, CenturyLink, Joyent, Rackspace, Interoute, Fujitsu, Skytap and NTT.
The big 3 are, of course, AWS, Azure and Google.
Alibaba is included only because of its dominance in the China region.
Oracle is optimised for its own workloads and is unattractive as a IaaS or PaaS solutions for non-Oracle deployments.
IBM has an utterly unconvincing strategy. It’s cloud largely supports legacy (mainframe) solutions.
How the Leaders Stack Up
AWS is the most mature cloud and has come to be seen as a safe choice, but be cautious: “Customers should be aware that while it’s easy to get started, optimal use – especially keeping up with new service innovations and best practices, and managing costs – may challenge even highly agile, expert IT organisations, including AWS partners. As new, less-experienced MSPs are added to AWS’s Audited MSP Partner program, this designation is becoming less of an assurance of MSP quality“.
Microsoft’s Azure has similar problems: Gartner says “Microsoft’s sales, field solutions architects and professional service teams did not have an adequate technical understanding of Azure“.
“Technical support personnel may also lack adequate expertise. The FastTrack program, which provides onboarding support from Azure engineering, results in successful pilots. Customers should use an MSP to execute a more successful implementation, but Microsoft has just begun the process of certifying MSPs; expert, experienced MSPs will be identified in partner directories starting in 3Q18“.
They also rate Azure as “optimized to deliver ease of use to novices with simple projects” which is great but “comes at the cost of sometimes making complex configurations difficult and frustrating to implement“. Overall, however, Azure is rated a fine choice and especially so for organisations already committed to Microsoft.
The Small Print
All the providers offer solutions that will meet common regulatory compliance needs, unless otherwise noted. All providers have undergone SOC 1, SOC 2 and SOC 3 audits, as well as SSAE 16, ISO/IEC 27001, ISO/IEC 27017, and ISO/IEC 27018 audits.
In general, monthly compute availability SLAs of 99.95% and higher are the norm, and they are typically higher than availability SLAs for managed hosting. Service credits for outages in a given month are typically capped at 100% of the monthly bill, but some providers have caps as low as 25%. This availability percentage is typically non-negotiable, as it is based on an engineering estimate of the underlying infrastructure reliability. Maintenance windows are normally excluded from the SLA.
Some providers have a compute availability SLA that requires the customer to use compute capabilities in at least two fault domains (sometimes known as “availability zones” or the like); an SLA violation requires both fault domains to fail. Providers with an SLA of this type are explicitly noted as having a multi-fault-domain SLA.
Very few of the providers have an SLA for compute or storage performance. None of these providers oversubscribe compute or RAM resources in standard compute instances, but some may have special, less expensive instance types that do, such as “burstable” instances.
Many providers have additional SLAs covering network availability and performance, customer service responsiveness and other service aspects.
Infrastructure resources are not normally automatically replicated into multiple data centres, unless otherwise noted; customers are responsible for their own business continuity. Some providers offer optional disaster recovery solutions.
All providers offer, at minimum, per-hour metering of virtual machines (VMs), and some can offer shorter metering increments, which can be more cost-effective for short-term batch jobs. Providers charge on a per-VM basis, unless otherwise noted.
Increasingly, providers are able to offer bare-metal physical servers on a dynamic basis, priced by the hour. Providers with a bare-metal option are noted as such.
All the providers partner with carrier-neutral colocation exchanges. This allows customers to obtain connectivity from a variety of carriers that are located in these facilities. In addition, many customers have needs that require a small amount of supplemental colocation in low-latency proximity with their cloud provider. For instance, they may have a large-scale database, specialized network equipment, or legacy equipment such as a mainframe.
Some providers offer a software marketplace where software vendors specially license and package their software to run on that provider’s cloud IaaS offering. Marketplace software can be automatically installed with a click, and can be billed through the provider. Some marketplaces also contain other third-party solutions and services.
All providers offer enterprise-class support with 24/7 customer service, via phone, email and chat, along with an account manager. Most providers include this with their offering. Some offer a lower level of support by default, but allow customers to pay extra for enterprise-class support.
All the providers will sign contracts with customers, can invoice and can consolidate bills from multiple accounts. All providers also offer online sign-up and credit card billing, as they recognize that enterprise buyers prefer contracts and invoices. Some will sign “zero dollar” contracts that do not commit a customer to a certain volume.
All providers offer a portal, documentation, technical support, customer support and contracts in English. Some can provide one or more of these in languages other than English. Most providers can conduct business in local languages, even if all aspects of service are English-only.
In Our Experience
As quoted by our Chairman:
“During my tenure at Gijima, Datacentrix, Telkom Cybernest, Liberty, Shoprite, and most recently EOH, we deployed PaaS and SaaS solutions to both AWS and Azure (as well as some SAP cloud hosted SaaS offerings).
SAP was always the easiest, because it is a contained cloud environment with little MSP driven management at an infrastructure deployment and day-to-day management level. Optimising matters was sometimes a painful process with a lot of to-and-froing with the Germans.
Azure was also relatively easy, as its administration portal, in my opinion, is superior to that of AWS. The pricing options are well defined, and it was easier to model costs and drive down pricing by using the option of long term reserved instances. Azure has 54 regions, including South Africa N and W. The Azure stack has also matured with the addition of Flow, PowerApps and PowerBI, making extensibility, workflow, analytics and dashboarding of both COTS (especially the Dynamics 365 family) and custom developed solutions a breeze.
The AWS EC2 environment is a tougher proposition. We found the choice of region (19 in total) can significantly change your monthly bill, and there are a number of parameters associated with an instance that increases the cost. The network latency based on the choice of region varied quite a bit. We had to experiment to get performance tweaked at both instance and throughput levels.
AWS QuickSight is simpler to use than PowerBI, but has far less features. QuickSight also hinges on a ‘pay-per-session’ pricing, making it more difficult to cost model than PowerBI“.
As quoted by our CTO:
“As Cloud infrastructure continues to mature and become more widely acceptable, it releases the need for on-premise hardware. This aids start-ups in making quick fast decisions on how to setup IT services and infrastructure in an every changing world.
Setting up virtual machines, virtual networks, IoT applications and back office systems can now take a few hours at most to set up, compared to days if not weeks for on-premise systems.
The ability to build hybrid based systems, on-premise and cloud system communicating with each assist in the migration to cloud based systems.
Per second billing in many of the cloud based systems allows for agile and start-up environments to flourish.
No expert skills are needed to get a handle the cloud based environments, where current IT skills can quickly learn the cloud concepts and with rich documentation, community forums, sample code and courses are readily available.
Backup and restore is easily configured to local or geo redundant environments, bringing the risk of data loss to a minimum“.
Choosing a cloud provider can be a rather emotive decision, as we are somehow instinctively attracted to one or the other.
Azure and AWS provide very similar offerings, with little to differentiate between the two. It holds true for security, workload scalability and SLA’s. However, costing the same workloads on the two platforms for comparison proves difficult, as their instances are not identical.
Microsoft has the advantage in cost efficiencies as they use their own enterprise software (at nexgen level) to manage their stack, and obviously Windows Server and SQL as fundamental building blocks are their own products. As a result Azure is 4-12% cheaper than AWS.
With our business being SA-based, and with Azure services being offered from the Teraco data centres in Cape Town and Midrand, it inherently makes sense to deliver solutions to SA/sub-Saharan Africa customers from Azure.
There is both pragmatism and altruism in this choice.
Pragmatically the service is delivered in ZAR (without the woes of exchange rate fluctuations) and is therefore much easier to cost and model in a business plan. And there is no network latency issues to deal with.
Altruistically, much of the money remains in the SA economy and provides jobs for South Africans.