The Multi-Cloud: Who Really Cares?

Feb 20 2014 | by Alex Wexler

0 Flares Twitter 0 Facebook 0 LinkedIn 0 Email -- Filament.io 0 Flares ×

Since we rolled out Cloudyn for Google Cloud, we’ve gotten lots of feedback, including some questions about why companies should be considering multi-cloud deployments, especially if they are content with their current provider. But while there is truth to the statement “if it ain’t broke, don’t fix it”, simply being content or satisfied in today’s business world, just won’t cut it with the C-level execs.

The mere fact that Google Cloud hit the ground running with a whopping 30,000 paid users when it launched the Google Compute Engine (GCE) is raising eyebrows. So when management asks what you are doing about “vendor lock-in” or if your current vendor is really the optimal choice for your deployment, you need to be prepared.

While there are some deployments that will need to stay put for various reasons, there always are workloads that can be ported from cloud to cloud and new projects can easily be started anywhere. All you need to do is make sure it makes sense from a cost and performance perspective.

To aid in this process, let’s take a comparative look at AWS and GCE.

AWS vs. GCE – 7 points to ponder

1. Instance Types– EC2 splits its instances into 6 families with a total of 29 instance types, whereas GCE splits its instances into 4 families with a total of 15 instance types.  These numbers may lead you to believe that GCE is trailing AWS, but that’s not 100% accurate.  In truth, Google is not missing much. Many of the compute-optimized EC2 instances go unutilized, and in actuality, these AWS consumers would benefit by migrating to a much smaller Google instance.

An additional advantage to EC2 is its GPU instances, which is beneficial for graphic applications such as game and 3-D application streaming.  It doesn’t look like GCE is looking to enter the gaming or 3-D industry yet.

2. Operating System Flexibility– GCE has a major limitation– it only runs on Linux. AWS, on the other hand, runs on multiple both Linux and Windows. Kris Bleisner of 2 Watch Inc, notes that the number one computing platform for enterprises is Windows with server shares in the 70 percent+ range (source: IDC). If Google wants to compete in this area, it will need to support Windows platform and do it soon.

Despite this limitation, Cloudyn data shows that GCE platform is compatible with around 53% of current AWS customers. Google is doing an amazing job in positioning itself as an AWS competitor.

3. Persistent Disks -(Best way to distribute data across the globe) -At Google I/O 2013,  Google Engineers explained that with GCE you could mount a persistent read-write disk to one VM and connect it to hundreds of VMs in the same zone as a read-only.  This capability will enable you to distribute data to large workforces in minutes!   If you want the data to be accessed in another zone, you can take a snapshot of the disk, recreate the disk, and mount it in another zone to an additional set of VMs.  The presenters warn to check the performance beforehand because it may vary based on workload.

On the AWS side of the cloud, customers still need to figure out a way to share and synchronize content among many instances.

4. Amazon’s Load Balancer – One thing that drives many DevOps folks crazy is Amazon’s Elastic Load Balancer (ELB).  AWS’ ELB cannot deal with an unexpected rise of traffic.  This requires the end-user to give a heads up on expected traffic spikes.  The only way to warn Amazon is to purchase AWS support.  Google capitalizes on this flaw, and writes on its site that its “native load-balancing technology helps you spread incoming network traffic across a pool of instances, so you can achieve maximum performance, throughput and availability at low cost.”

5. I/O Speed – GCE’s default I/O instance is very fast. In order for EC2 to match GCE’s default I/O speed, you’d need to use EBS-optimized EC2 instances with provisioned IOPS volumes.  In the end, it could cost you almost double.

6. Block Storage – AWS persistent disks only support one TB, whereas GCE supports a staggering 10 TB.   This disparity has caught the attention of many would-be and current AWS customers in the market, and might be THE reason for them to choose GCE over AWS.

7. Pricing – Google offers sub-hour billing, meaning that you pay in one minute increments with a ten minute minimum, whereas AWS bills you by hour and rounds up for partial hours.  This pricing policy can be costly as 65% of AWS instances run for less than two hours.

Additionally, GCE’s on-demand prices are cheaper than AWS. Although AWS offers Reserved and Spot instances which offers significantly lower pricing than On-Demand, rumor has it that Google sales can offer very discounted pricing for long-term commitment – akin to AWS RIs.

War of the Titansbattle-of-the-clouds
We are about to witness a war between two Titans -Google and Amazon. It will be exciting to watch the constant innovations and (hopefully) price cutting coming our way. Features and prices are always changing, so it’s essential to stay current with AWS and GCE comparisons to get the most bang for your buck.

See which cloud works best for you with Cloudyn’s new multi cloud monitoring tool!

 

Login

Forgot Password?

No account yet? Register.

0 Flares Twitter 0 Facebook 0 LinkedIn 0 Email -- Filament.io 0 Flares ×