post-img

AWS Re:Invent 2018: Learnings and Promises

Posted by Pascal Joly December 13, 2018
AWS Re:Invent 2018: Learnings and Promises

It's been just over a week since the end of the 2018 edition of AWS Re:Invent in Las Vegas. Barely enough to recover from various viruses and bugs accumulated from rubbing shoulders with 50,000 other attendees.  To Amazon's credit, they managed the complicated logistics of this huge event very professionally. We also discovered they have a thriving cupcake business :)

From my prospective, this trade show was an opportunity to meet with a variety of our alliance partners, including AWS themselves. It also gave me a chance to attend many sessions and learn about state of the art use cases presented by various AWS customers (Airbnb, Sony, Expedia...). This led me to reflect on this fascinating question: how can you keep growing at a fast pace when you're already a giant  - without falling under your own weight?

Keeping up with the Pace of Innovation

AWS ReInvent Innovation

Given the size of the AWS business ($26b revenue in Q3 2018 alone, with 46% growth rate), Andy Jassy must be quite a magician to come up with ways to keep the pace of growth at this stage.  Amazon retail started the AWS idea, back in 2006, based on unused infrastructure capacity. It has since provided AWS a playground for introducing or refining their new products. including the latest Machine Learning predictive and forecast services. Thankfully since these early days, AWS customer base has grown into the thousands. These contributors and builders are providing constant feedback to fuel the growth engine of new features.

The accelerated pace of technology has also kept the AWS on its edge. to remain the undisputed leader in the cloud space, AWS takes no chance and introduces new services often shortly after or at the top of the hype curve of a given trend. Blockchain is a good example among others. Staying at the top spot is a lot harder than reaching it.

New Services Announced in Every Corner of IT

Andy Jassy announced dozens of new services on top of the existing 90 services already in place, during his keynote address. These services cover every corner of IT. Among others:

-Hybrid cloud: in sort of a reversal course of action, and after dismissing hybrid cloud as a "passing trend", AWS announced AWS Outpost, to provide racks of hardware to customers in their own data center. This goes beyond the AWS-VMware partnership that extended vCenter in the AWS public cloud.

-Networking: AWS transit gateway : this service is closer to the nuts and bolts of creating a production grade distributed application in the cloud. Instead of painfully creating peer to peer connections between each network domain, transit gateways provide a hub that centrally manages network interconnections, making it easier for builders to interconnect VPCs, VPNs, and external connections.

- Databases: Not the sexiest of services, databases are still at the core of any workload deployed in AWS. In an on-going rebuke to Oracle, Andy Jassy re-emphasize this is all about choices. Beyond SQL and NoSQL,  he announced several new specialized flavors of databases, such as a graph database (Netptune). These types of databases are optimized to compute highly interconnected datasets.

-Machine Learning/AI: major services such as SageMaker were introduced last year in the context of fierce competition among all cloud providers to gain the upper ground in this red hot field, and this year was no exception to this continuing trend. All the sessions offered during the event in the AI track had a waiting list. Leveraging from their experience with Amazon retail (including the success of Alexa), AWS again showed their leadership and made announcements  covering all layers of the AI technology stack. That included services aimed at non-data scientist experts such as a forecasting service and personalizing service (preferences). Recognizing the need for talent in this field, Amazon also launched their own ML University. Anyone can sign up for free...as long as you use AWS for all your practice exercises. :)

Sticking with Core Principles

source: theschoolrun.com

Considering the breadth of these rich services, how can AWS afford to keep innovating?

Turns out there are 2 business-driven tenants that Amazon always keeps at the core of its principles:

  • Always encourage more infrastructure consumption. Behind each service lies infrastructure. Using the pay as you go model, in theory there are only benefits for the cloud users. However, while Amazone makes it easy and sometimes fully transparent to create and consume new resources in various core domains (networking, storage, compute, IP addressing), it can be quite burdensome for the AWS "builder" or developer to clean up all the components involved in existing environments when no longer needed. Anything left out running will eventually be charged back based on the amount of uptime, so the onus remains on the AWS customer shoulders to do due diligence on that front. Fortunately, there are third party solutions like Quali's CloudShell that will do that for you automatically.
  • Never run out of capacity: By definition, the cloud should have infinite capacity and an AWS customers, in theory, should never have to worry about it. At the core of AWS (and other public cloud providers) is the concept of elasticity. Elastic Load Balancing service has long been a standard for any application deployed on AWS EC2. More recently, AWS Lambda and the success of Serverless computing is a good case in point for application functions that only need short lived transactions.

By remaining solid on these core principles, AWS can keep investing in new cutting edge services while remaining very profitable. The coming year looks exciting as well for all Cloud Service Providers. Amazon has set the bar pretty high, and the runner ups (Microsoft, Google Cloud, Oracle and IBM) will want to continue chipping away at its market share, which in turn will also fuel more creativity and innovation. That would not change if Amazon decided to spin-off AWS as an independent company, although for now that topic will be best left to speculators, or even another blog.

CTA Banner

Learn more about Quali

Watch our solutions overview video
post-img

How to Accelerate and De-risk Hybrid Cloud Deployments

Posted by Shashi Kiran October 31, 2016
How to Accelerate and De-risk Hybrid Cloud Deployments

Pubic, Private and Hybrid Clouds

The last 10-15 years have seen a tremendous amount of investment take place in building data centers. Enterprises worldwide had targeted data center consolidation as one of the top CIO initiatives in a bid to optimize costs and increase efficiency. This investment also evolved to be the foundation for private cloud as virtualization, as-a-service offerings and software defined data centers grew.

The last 5-7 years or so have seen an increased adoption of the public cloud. With the agility, simplicity and ubiquity of public cloud, the need for large enterprise-owned data centers has therefore somewhat diminished. At the same time for various reasons – legacy workloads, compliance, need for control as well as in some cases – cost, traditional data centers and private clouds continue to remain relevant.

Promise of Hybrid Clouds

It is therefore no surprise to see the increased want of hybrid clouds. Several analyst firms have estimated between 50-75% of Enterprises marching down the path of hybrid cloud deployments over the next 2-3 years.

On one hand, they allow the status quo to prevail offering better control, visibility and manageability with familiar operational models. On the other hand, they promise the speed and simplicity that public clouds bring. Hybrid clouds represent the “best of both worlds”. It is therefore no surprise that both private and public cloud vendors are forging partnerships to ensure a better experience for customers. Case in point - the recently announced integration between VMware and AWS.

Challenges

While the promise of hybrid clouds is alluring the pathway is somewhat challenging. Some refer to the hybrid cloud as “the wild west”, in part due to the non-standardization of toolsets. Hybrid clouds bring their own challenges in terms of varied operational models, architectural mismatches and learning curve. All these can add risk to hybrid cloud deployments and dampen the velocity of cloud deployments in general.

Hybrid Cloud Sandboxes

This is where Hybrid cloud sandboxes can help. These sandboxes replicate environments that are representative of production environments in public and private cloud deployments. These can be brought together in the context of the same sandbox offering simplicity of tooling, standardization of operational procedures and increasing affordability. With such sandboxes – it is easier to build out environments of common use-cases (say Dev/Test, Compliance, Capacity Augmentation etc.) and test these scenarios out in advance combining them with DevOps centric practices. VMblog covered some of these issues and Quali's take in their article here.

Embracing cloud-centric DevOps with hybrid cloud sandboxes can help increase cloud adoption while decreasing risk and increasing quality.

Join this Webinar!

To illustrate these concepts, we are conducting a webinar on November 2nd at 9 AM PST for audiences in the United States and a follow-on webinar on November 9th for audiences in Europe and Asia Pacific. I’m hosting this with our CTO Joan Wrabetz and our demo guru Hans Ashlock.

We’ll go through the top use-cases, showcase a demo and answer questions!

 

CTA Banner

Learn more about Quali

Watch our solutions overview video