The Lightcrest Blog

Let's talk about fluid computing, hyperconverged infrastructure, and hybrid cloud technology.

Why Are Cloud Environments Such A Headache?

Posted by on Nov. 18, 2016 .

Last update on Jan. 4, 2017.

The cloud ecosystem brings with it a huge number of benefits that can't be ignored, but it seems to bring a lot of unique troubles at the same time. For every bit of productivity or efficiency you get, the dreaded issue of the "learning curve" takes its place. Getting to the point where you're getting all of the benefits and none of the downsides can definitely seem like a big hill to climb, but that is just the cost of doing business in the digital age, right? 

Not necessarily. 

There are four main categories of issues that many companies run into when experimenting with cloud environments. They include: 

  • Performance

  • Administration

  • Clunky Interfaces

  • Security

These elements contain both platform and provider-specific issues, both of which are critical factors to consider when evaluating the cloud landscape. 
 

Performance

One of the most alluring features of the cloud is on-demand computing and storage. You eliminate long-term forecasting and “step-wise” infrastructure purchases in favor of “smooth-curve” infrastructure ramp-up and ramp-down. But one of the downsides of the public cloud model is that the infrastructure is shared, and often oversubscribed - meaning the performance is inconsistent, and there is no guarantee that any one virtual machine will continue to exist over a given time interval. 

What this means from a performance standpoint is that you have to be diligent about establishing expected latency and request rates for your applications (both at the compute and storage tiers) so that you can baseline the minimum viable infrastructure in your cloud. Since public cloud tends to be much more inconsistent, you also have to account for an increase in infrastructure to counteract drops in performance due to noisy neighbors, stolen CPU cycles, and I/O oversubscription. The end result usually means you need to invest much more in DevOps and automation to account for these variables and deal with a more expensive bill. 

This is fine for organizations with the in-house talent and budget, and not ideal for organizations who only want to hire developers and focus on their application. You need to be realistic when evaluating these technologies - do you want to be in the business of building your own DevOps toolsets for managing these wildcards? 

Private clouds, while more rigid at the hypervisor level, are much more consistent from a performance standpoint, and the right products tend to have simpler APIs for managing the infrastructure. You may want to trade off on-demand elasticity for simplicity and higher performance, as this can lower the amount of time and money you need to invest in benchmarking and accounting for public cloud performance degradation. 

Administration

The cloud promises lower time-to-deliver for computing and storage resources, but it doesn’t administrate itself. Sure auto-scaling and elastic load balancing can make your life easier, and certainly, on-demand computing reduces procurement time of virtual machines. But your guest operating systems still have to be managed, and you need infrastructure and network expertise to understand whether or not you’re over-provisioned - otherwise you just get “VM sprawl”, with your developers firing up VM’s all over the place and blowing out your budget. 

As you consider the public cloud and alternatives such as private and hybrid cloud computing, you need, to be honest with yourself. Is your team ready to support an infrastructure on a 24x7x365 basis? Do you have the in-house expertise necessary to manage the infrastructure you have on the frontlines of the internet? Do you have resident experts in caching, database optimization, and storage scalability? What about folks that can optimize for performance while keeping your cloud costs to a minimum? 

Many organizations find that specialization increases their bottom line. By focusing on their core competencies (i.e. application development, marketing, sales) and relying on a provider to manage and optimize their cloud infrastructure, they realize wider margins and can hire people that can differentiate their product and ultimately make their business more competitive. 

Clunky Interfaces

Many cloud providers offer user interfaces that are not only antiquated but slow, requiring multiple page loads and refreshes to get what should be real-time updates. Fast, simple user interfaces matter when you have developers who “just want to get work done”. As you make your first foray into the cloud, you should be ranking ease-of-use as highly as any other metric. 

If your cloud provider’s user interface requires documentation, something is wrong. If your developers can’t intuitively procure and manage infrastructure, they will lean more heavily on you to hire costly operations staff. 

This goes for APIs as well - do your cloud provider’s JSON results look like ancient Sanskrit, or are they intuitive? Do you need to scour documentation to get basic operations done through the API? How easy is it to allocate virtual machines, provision storage, snapshot, and clone? If you have API issues, can you call someone at your current provider to walk you through it?

The simplicity of the user interface and APIs is an indicator of code quality throughout the platform; so if you’re looking at over-engineered cloud products, it’s time to take stock of alternatives that will make your life easier. And if you have API issues that can’t be addressed by a human being, then you should consider a provider that combines excellent customer support with a superlative cloud product.

Security

Security is a huge concern for the modern enterprise, whether it be a burgeoning startup, healthcare organization, or Fortune 500 company. But not all cloud infrastructure is created equally when it comes to default network topologies, logical architectures, and security best practices. Furthermore, your cloud isn’t going to secure itself. In fact, the cloud can actually increase your attack surface and make you more vulnerable to compromise.

 Take a cloud portal for example. If your developers leak account credentials, either deliberately or inadvertently, the attacker not only commands your entire infrastructure build but can compromise every virtual machine in your environment. And once they gain access to your cloud APIs, they can explode your cloud bill to the tune of hundreds of thousands of dollars per month - something your public cloud provider may not be willing to credit.
 
And of course, most public cloud providers do not secure your guest operating systems for you. They provide you with the infrastructure, but they don’t provide you with the best practices required to keep your actual applications and runtime environments secure. This is why many enterprises rely on managed cloud providers to maintain and implement security best practices on their behalf.
 

Presenting Kahu: The Lightcrest Answer to Cloud Ecosystems

Lightcrest addresses the challenges of cloud platform optimization and on-going management with the Kahu Compute Fabric. By delivering ultra-fast software-defined computing infrastructure  on the very best hardware, Kahu provides the fastest UI in the industry, coupled with super-simple APIs and a self-healing control plane. When used in conjunction with Lightcrest FullStack Engineering Services, Kahu provides customers with hyper-scalable infrastructure, lower total-cost-of-ownership, and overnight increases in cloud expertise. 

Data sovereignty is a big issue for companies who may be hesitant about moving into the cloud for the first time, and Kahu addresses this concern head-on by providing leading-edge private cloud infrastructure. With Kahu, data sovereignty becomes a certainty without the traditional costs associated with a bespoke private cloud. For the first time, customers can own their own private cloud and eliminate cloud rent without their own datacenter or operations team. 

With heavily optimized and pre-integrated software and hardware stacks and software-defined storage and networking on every node, Lightcrest has positioned Kahu as the leader in software-defined infrastructure. To learn more about how Kahu can make it easier than ever to compete on a massive scale while saving money on your IT infrastructure, contact Lightcrest today.