Why Application Performance is the Cloud’s Achilles’ heel

Who should you trust for your Cloud Acceleration needs?

The cloud wasn’t built for application performance. A basic assumption that cloud providers lean on is that the Internet (and almost every one of the network of networks) will eventually catch up to the cloud’s evolving computing and distribution models.

Well, reality hasn’t been cooperating. Underperforming networks have long been a choke point for B2B communications. Now, they’re threatening the very viability of cloud-based applications.

Have you ever had this happen? You log onto Office 365 or Salesforce.com or any other SaaS/cloud service and wait and wait and wait? It’s a common problem.

Cloud providers have been doing a great job of disrupting outdated business models and workflows. Universality, ease of access, on-demand usage and pricing, and faster ROI are just few of the benefits that the cloud delivers to pretty much any business.

While transitioning to the cloud has clear advantages for the enterprise, cloud and SaaS vendors forgot to bake application performance into their service offerings. While compiling their lists of business productivity benefits, how did they overlook the fact that the network could bring the whole house of cards tumbling down?

The Emergence of Cloud Acceleration as a Service

Spend enough time in Silicon Valley, and you’ll soon realize that anyplace there’s a problem of this scale there is a corresponding opportunity of equal (or larger) size. Legacy vendors were the first to attempt to tackle the Achilles’ heel of cloud apps. They threw hardware at the problem, expensive private lines, and they layered on complicated, hardware-dependant solutions, such as WAN optimization.

Those solutions, however, were not designed for the cloud. They were developed to connect HQ to branch offices, or to roll in overseas offices, granting them access to in-house digital assets. The legacy solutions are expensive, slow to deploy, complicated, and weighted heavily to back-end costs (labor-intensive management and maintenance contracts), as well.

Wasn’t increasing agility, driving down costs, and removing labor burdens the whole point of moving all of those workloads to the cloud? The legacy cloud acceleration fixes undercut all of those gains.

Again, where there is a Silicon Valley-scale problem, some entrepreneur or another will rush in to solve it. We, at Aryaka (and many of the greener startups who have since rushed into the space) are leveraging the cloud to improve the cloud. We’re taking that Achilles’ heel and wrapping it in a loose but protective layer of kevlar.

We’re building a new network of networks, a business-scale network, or, simply a cloud network for the enterprise.

Cloud trends accelerate too

In a recent article titled “9 Cloud Trends for 2015,” Charles Babcock, Editor, InformationWeek mentioned that 2015 would “mark the start of the land rush of enterprise workloads moving into the cloud.” Mr. Babcock also highlighted the fact that AWS adoption was merely the beginning and that enterprises would gear up for cloud services provided by veterans such as Google Compute Engine, Microsoft Azure, CenturyLink Savvis, Verizon Terremark and Rackspace.

According to Rackspace, top enterprise workloads in the cloud or currently migrating to the cloud include email collaboration, web applications, application testing and development, CRM, custom business apps, business analytics and ERP. While private cloud is the preferred cloud deployment model for financial ERP, business analytics, application testing and development workloads, public clouds are in vogue when it comes to CRM, web applications, collaboration and content management platforms.

Given the magnitude and diversity of enterprise workloads migrating to the cloud, enterprise dependence on cloud-based services for their productivity needs is rising – at an accelerated rate. It is safe to say that there is now a direct correlation between how cloud applications perform and the enterprise’s bottom line.

Again, application performance was the one key ingredient forgotten when all of those cloud services were cooked up. The primary reason for this is the fact that the access mechanism for any cloud service is almost always the unreliable public Internet, which is plagued with latency, congestion, packet loss and jitter. Old solutions that used to accelerate applications to distant branch offices relied on expensive technologies (MPLS + WAN optimization) weren’t designed for the cloud, for transactional, TCP- and HTTP-based, interactive applications. Worse, those legacy fixes are just not affordable for cloud-age use cases. There’s no way your CIO (or CEO) will sign off on private lines to every important cloud datacenter, along with WAN optimization appliances at every end user location. That could easily eclipse your entire IT budget.

In fact, in our mobile age, the legacy approach isn’t even remotely feasible.

An ideal cloud acceleration solution needs two critical pieces of enabling technology:

  1. A reliable, cloud-scale network. B2B cloud services need an access mechanism that bypasses the unreliable public Internet and allows end users located in anywhere in the world, on any device, anytime (peak traffic times too), to communicate seamlessly with their cloud instances, so they can benefit from a LAN-like experience with predictable application performance.
  2. A cloud optimization stack.  As cloud services increase Internet bandwidth consumption by opening up several TCP connections at once, the firewall becomes a choke point, slowing application performance down to a crawl. WAN Optimization technologies, once removed from their hardware shackles, are able to optimize cloud traffic through compression, data de-duplication, and TCP optimization.

That’s what you need in your cloud acceleration solution. Anything less, and you’ll be stuck in the slow lane. Anything more (and by more I’m mainly speaking of unnecessary hardware constraining solutions better delivered as cloud services) and your ROI will evaporate.

The next time you log into Office 365 and then wait and wait; the next time you see the “still trying” message and then the dreaded “try again later” one; take a deep breath, and, perhaps, instead of blaming Microsoft for botching their migration to the cloud, pause and take in the big picture. There’s a good chance it’s not Microsoft, but the network between you and your instance in the Office 365 cloud.

Fortunately, it’s a solvable problem and cloud acceleration tools will help you do it.

At Aryaka, we have solved the cloud application performance problem by building the world’s only global and private cloud overlay network that is within 1-5ms of most cloud service providers. Our Cloud Acceleration solution provides enterprises with an optimized private network that ensures accelerated access to any and all cloud/SaaS applications.

All that enterprises need to do is use one (or more) Internet link(s) to connect to one of our 25 global points of presence. While the private core allows global enterprises to bypass the unreliable public Internet, Aryaka’s proprietary WAN optimization software stack enables them to leverage the benefits of TCP optimization, compression and de-duplication technologies to maximize throughput and performance of their resources in the cloud. Aryaka’s cloud acceleration offering supports all SaaS and cloud services – large players like Microsoft Office 365, Microsoft Azure, Amazon Web Services (AWS), Google Apps and Salesforce – as well as any smaller public, private or hybrid cloud deployments.

Our customers have reported up to 99% bandwidth scaling as well as 40X application acceleration benefits. But don’t take our word for it . . . try us out for yourself!

About the author