Clutch spoke with Vittaly Tavor, the Co-Founder and Vice President of Product at Cloudyn, about the various cloud platform options for enterprise customers – an important decision for anyone looking for a robust Cloud Platform.
Learn more about Cloudyn at cloudyn.com.
Please describe your organization.
Our company is specialized in cost, performance, and utilization management. When considering the cloud, the most important operational metric is cost. If an application is built correctly, a client can assume that cloud resources are infinite and they can take advantage of more resources as needed. Most applications are built in order to scale horizontally for more performance. If something in the configuration is wrong, it will be seen through the bill at the end of the month. It's crucial for users to understand what they are paying for.
Before the cloud era, the typical utilization level of a server was in the range of 7%. When we started virtualization, it moved closer to 11%. With the addition of the cloud, this moved to a whopping 13%. Still, most servers are highly underutilized and most of the cloud space is highly inefficient.
Our operations involve monitoring utilization, performance, and cost. We highlight anomalies and inefficiencies, providing accurate breakdowns and root cause analysis of these dimensions, while offering additional services like cloud comparison, as well as recommendations about sizing and price.
As we worked with different customers, we discovered that virtually all of them are over-provisioning and over-paying. We used to work with companies like GoGrid and Rackspace Public Cloud. Today, Rackspace is mostly reselling Amazon services, but they do have some kind of offering. The main public cloud providers available today are Amazon, Microsoft Azure and Google. I've had some experience with each of them. There are other, smaller companies, but I would not consider them to be significant. There are private cloud offerings, or rather private virtualizations like VMware vCloud Air, which makes up the highest percentage of private virtualizations today. These are the cloud platforms with which I have experience, and the ones we're working with.
What is a typical business challenge a company may face that can be resolved by utilizing the cloud?
There are a few challenges that an enterprise would need to be aware of. This was the main challenge of 2015. Before, a lot of companies were talking about the cloud, but none were considering it seriously. Virtually all enterprises have some kind of cloud strategy at this time. This means that the cloud has reached a certain level of maturity. One of the perceived challenges was security. Today, it is a concern, but not really a challenge when moving into the cloud. Security is pretty much covered.
Challenges still exist around compliance, the main one being the mindset of IT managers. The younger ones embrace the cloud, while the older ones still have a difficult time getting used to having different types of control needs. Once this mindset has been overcome, applications used become the next issue. If a company takes its existing applications and moves them to the cloud, they'll probably pay much more for the service, without gaining a lot. The applications need to be elastic, growing and shrinking based on needs. Another issue will be management. When moving to the cloud, they need to get used to an entirely new set of tools. For example, the mindset by which cost is an operational metric and not a financial one is something they'll need to get used to. This is a question of maturity.
The third challenge is a more practical one: when moving to the cloud, there are no good tools which allow the user to address capacity management. A lot of applications try to give an idea of what they need within the cloud, but virtually all of them are inaccurate. This is a matter of trial and error: moving their application into the cloud and seeing how it behaves. It then needs to be sized according to needs.
Are there any key areas that you would highlight as an important method for implementation?
The main motto of moving to the cloud is the same as for the rest of IT: internal expertise has to be original, while whatever is outside of this, like specific cloud or management best practices, should not be original. Industry standard tools are ideal here, like Mirantis or others. Established tools should also be used for security, cost management, and so on. It's essential for a user to get a set of tools which make the cloud transparent for them because, as it is today, it's far from being transparent.
How does the cost of cloud computing compare to legacy IT platforms?
Looking at a dollar-to-dollar cost, when moving an application from legacy IT to the cloud, it is possible to pay even four times more on occasion. This is because legacy IT focused on worst case scenarios, and allocated as many resources as would be needed for those cases. In the cloud though, the allocation is done strictly based on the task at hand, handling the usage spikes as they come. If you adhere to this kind of strategy, the cloud will be much more effective and cheaper than legacy IT.
The second part of best practices or behavior has to do with the multitude of services offered by cloud vendors. They make the life of application developers much easier. Amazon offers close to 400 services, including queues, email services, databases, and so on. If these are used instead of custom-developed ones, there will be a much shorter time for bringing a solution to market. On the other hand, clients will lock themselves in with a specific provider and moving to a different cloud service will be much more difficult. There is always a balance between the services one needs to use, and how mobile they need to remain in order to switch providers.
Each cloud provider will be better than others in specific areas. Amazon has the highest breadth of services available. They offer a solution for virtually every need, including four or five types of databases, infrastructure services, built-in simple email, queues, and so on. They are by far the most advanced in this respect. This shows through their percentage in the market share. They have a higher market penetration than all other services combined.
Microsoft offers an interesting view by recognizing that, regardless of industry, private cloud does not have a future. As such, they've built their strategy around a seamless transition between private and public could systems. They've always been the best at integration technologies. Even if they don't offer a full set of solutions now, users can be sure that everything will be smoothly integrated together. Microsoft has to catch up with Amazon, but they've made huge progress during 2015, becoming a viable competitor. They have enough of the needed services, and make a good offer at an enterprise level. Unlike Amazon, who is still not perceived as an enterprise player, Microsoft is.
Google has positioned itself as the cheapest cloud service around. My opinion is that they should have sold their value, and not their price. Today, Google Cloud has the best network available, with its own transatlantic lines. They offer live migration, which allows them to seamlessly move a service (or virtual machine) from one area to another within normal maintenance windows, reducing the downtime they'd experienced in the past. In terms of resilience and networks, Google is the best. Additionally, they have a very unique service called BigQuery. It is a high volume and high scale database. If a client needs blindingly fast responses, this will be the fastest database within all cloud services. Companies like Spotify have announced moving to Google in order to make use of it.
I'm not sure what IBM's strategy is. We do not hear a lot about them, and it's difficult to call them a public cloud. They do not approach this sector, instead selling directly to enterprises, focusing on service and customization.
Is there one platform that you would overall recommend to businesses, or do you make recommendations on a case-by-case basis?
I believe it's a matter of case-by-case. Any enterprise of a certain size must have a presence within more than one cloud platform. You cannot rely on one vendor for any other type of service.
Are there any specific features or tools that set a platform aside from the others?
If a client needs a breadth of services, Amazon is the best choice. If a seamless transition between a public and private network and an enterprise approach are needed, then Microsoft is the best provider. Google offers the fastest network and cloud services, which mostly apply to developers who want to build solutions from the ground up.
Have you had any interaction with the support resources offered by these three providers?
Given a choice, I would rely on outside support, but I have used all of them, and each was very responsive in terms of understanding our issues. None of them actually provided resolutions.
As an example: Amazon offers billing files for customers. Within them, there are specific codes which show platform information and other details about the items provided. We've asked for documents which describe these codes. The answer was that this is an internal document and it cannot be shared.
Microsoft had just provided us a very good billing API. However, we had issues using it with large customers. It took them about a month and a half to fix this issue. Google has had similar types of drawbacks. They're all focused on new features and covering more ground than on support cases. None of these providers are very agile.
We have five additional questions. For each of these, we ask that you rate each platform discussed on a scale of one to five, with five being the best score.
How would you rate them for the availability of their features?
Amazon Web Services: Five.
Microsoft Azure: Four.
How would you rate them for ease of use?
Amazon Web Services: Three.
Microsoft Azure: Three.
The most basic and simple features would receive a five but, once you start digging in to more complex items, the rating goes down for all three.
How would you rate them for their support resources and the helpfulness of support staffs?
Amazon Web Services: Two.
Microsoft Azure: Two.
How would you rate your overall satisfaction with each platform?
Amazon Web Services: Four.
Microsoft Azure: Four.
This rating is based on personal satisfaction. For customers, the level of satisfaction varies between three and four for virtually all of them. If the rating drops below three, they will switch to a different provider. All of these platforms perform according to different expectations. They all do their job, and all of them offer the same level of durability and scalability. The differences are much more subtle than one-to-five ratings.
The cloud itself may have a lot of drawbacks, but it is the future. All companies need to have a strategy for it because, five years from now, half of all IT will be in the cloud.