We know how to do services cheaply and in bulk, but that doesn’t mean there aren’t big companies with big projects

[gtranslate]

When designing services, we think first and foremost about automation, as this is the basis for rapid growth and selling a large number of services at the lowest possible price. From ordering and paying for the service, to setting up, upgrading, activating additional services and cancelling the service, everything is done automatically. When you move to a cloud infrastructure, your service can even automatically change physical servers based on the current load to avoid congestion. Automation does everything faster, safer and more efficiently than humans. This is our vision, which we have been successful in achieving so far.

Thanks to automation, we don’t need as many people

The vast majority of our customers have never needed customer support. They’re gonna do it all themselves. Many technical problems are detected by our diagnostics and solved by robots before they occur. As a result, we can technically service, in about 10 people, more than 642 thousand services, which is the service operation and customer support department. Another 30 or so colleagues are responsible for development, technology development and construction of datacentres, business affairs and advertising.

If we didn’t have automation, we would need 10 times more of us and it would mean much higher operating costs. Moreover, some things could probably not be done on such a scale and certainly not in such quality. Our robots keep an eye on many things and can correct many errors themselves before they become apparent. Of course, robots don’t make as many mistakes as humans. Robots work 24/7 and are more reliable than humans and never get tired.

We also have big customers

We have earned a reputation as a low-cost hosting provider. Our efforts to offer domains at purchase prices certainly play a part in this. However, this certainly does not mean that large companies do not have their projects here.

For example, from the CZECH TOP 100 ranking of the most important companies in the Czech Republic, at least 8 of them have a service in our country. These include T-Mobile, České Dráhy, Alza.cz, Kiwi.com, ČD Cargo and Notino. And that’s just the ones who are with us directly. Many of the projects of the companies in the ranking are with us through subsidiaries or the service is run by a solution provider and we do not even record such companies.

We know that there could be more, but the type of service also has an impact. We do not offer the most frequently requested serverhousing (yet). Dedicated servers are offered for our customers who need more than just VPS and currently the service with the most interest from large companies , WEDOS Cloud, is still in test operation.

We admit that’s a bit of a shame. Big companies take notice of us and like our datacenters, technologies, new ISO certifications and the fact that we are outside Prague. A lot of companies want to have backups at least 50 Km away, which you can’t do in Prague 🙂

We perceive all this. Last year we launched WEDOS VIP, which is a project where we can create complex, larger packages from our services (e.g. dozens of dedicated servers with their own switch).

However, we avoid individual server solutions. Within WEDOS VIP, we can, for example, replace HDDs with SSDs or larger ones for servers from our current offer.

We are considering the provision of large and high-quality storage

Where we currently see potential is in data backup. From our vendor Hewlett Packard Enterprise, we have successfully tested and are using HPE Cloudline CL5800 Gen9, which is storage for up to 100 drives, and HPE 3PAR StoreServ Storage. It is the so-called ALL FLASH storage. Plus, we have plenty of connectivity. Unless someone is attacking us right now, the 300 Gbps line will not exceed 10 Gbps peak. We could easily provide several 10 Gbps lines for regular or even continuous backups.

We would be able to offer this hardware as a dedicated service just for you. The question is, of course, the configuration. HPE Cloudline CL5800 Gen9 is equipped with 100 drives with a total capacity of 1 PB (we are testing 10 and 14 TB). We normally get requests for 100 – 300 TB of backup space, which is not enough. In addition, a contract would have to be signed for at least 1 year.

HPE 3PAR StoreServ Storage is something different from regular storage. We are gradually moving our services to it. It’s more reliable and faster. We currently have about 120 servers running on it with NoLimit web hosts and besides fewer problems we have also noticed a slight increase in performance.

We deployed HPE 3PAR StoreServ Storage on 49 servers with NoLimit. The average processing time of a PHP script has been reduced by several percent.

We will gradually deploy HPE 3PAR StoreServ Storage for other services. Either directly (WMS) or as an additional disk (WEDOS Cloud, VPS ON).

During the summer, we want to have a second synchronous copy of the data in the second datacenter, so that in case of a failure, the data will automatically come from the second datacenter. In this case, all data is automatically written to two different storage locations in two different datacenters, thus minimizing the risk of online data loss. At the same time, you will minimize the possible outage. In a matter of hundreds of milliseconds, you are able to start all services from a second datacenter where there is a synchronous copy of the data.

For the most valuable data, we are also preparing a version in case we have a third copy in a completely different location and with a different provider. We will have to test whether it will be a synchronous or asynchronous solution. We’ll see.

At the same time, there is room for some individual solutions. Whether backups or larger databases. We could probably dedicate one full 3PAR of storage to SSD or HDD.

More big companies with DC3?

We were planning to build a third datacenter. Unfortunately, the pandemic has arrived and the paperwork is dragging on. It’s a shame, because we have the money ready for DC3. After a very successful 2019, 2020 was even more successful.

DC3 as a building complex is also designed for serverhousing. There will be not only seagulls for rent, but also whole separate rooms. We also plan to lease part of the office space. So a few tens of meters from your servers, you will be able to have your technician.

We will offer serverhousing only with DC3. For now, we will mainly have WEDOS Cloud available for large companies. After the successful ISO certification, all that is left to do is to fine-tune a few last details for full operation. With the launch of DC2, we will also be able to offer geographically separated machines. 3PAR will provide large discs at your service.

This third DC3 datacentre will be adjacent to the future D4 motorway. So it looks weird and so we don’t have to rename it DC4 and build another – real DC3 – somewhere else in the meantime.

More services on the horizon?

We are preparing several services that will really be for everyone, including large companies.

The first one is WEDOS AnyCast. It’s not just anycast DNS, i.e. strategically placed DNS servers around the world that speed up communication between DNS servers and the visitor. As we put our technology into select datacenters around the world, we intend to make the most of it. We plan to always put at least one full HPE Moonshot there, which is 45 servers, 2 switches with up to 4×40 Gbps connection. Some of the servers will be dedicated to our needs, but most will be available to customers for other services.

WEDOS AnyCast will include, for example, proxy servers for faster query handling, CDNs for content placement, VPNs for secure on-site work, and of course our protections. This is just a small taste. We have more in the pipeline, but we need to get to the locations first. For the time being we have a test instance only at our place in Hluboká. But everything is going according to plan, so hopefully we’ll be sending out our HPE Moonshots to the world soon 🙂

By the way, since April, when we had the biggest attacks in the history of the Czech Republic, we have deployed anycast DNS around the world and some of our DNS servers are running around the world in different locations and are on anycast technology. We want to expand it gradually. We have now successfully tested GeoIP DNS, where we will be able to allow clients to set different IP addresses in DNS based on different source IP addresses. So, for example, if a visitor from the Czech Republic asks, DNS will return the IP address 46.28.105.7, and if a visitor from Central America asks, DNS will return a completely different IP. Of course, you can respond to different attacking IP addresses with something else entirely. This solution, combined with WEDOS Anycast, will be capable of incredible possibilities. It will be interesting as a reverse proxy with a lot of protection and antiDDoS solutions, it will be usable as a CDN… and many, many other things.

Another great service is our WEDOS OnLine, which we use mainly for our internal purposes. It collects and evaluates data about our services, servers and other devices. The results are then publicly available at status.wedos.hosting. Just as we test hundreds of thousands of services and post the output on our status page, it can be used by anyone to monitor virtually any device, service, API, etc. connected to the internet. The outputs can then be public or internal. It’s up to you. We have the capacity to monitor millions of anything, and everything is ready for virtually unlimited scaling if needed 🙂

So far, the service is intended more for personal use. IT-savvy people can set the controls themselves. If internet networks and technology are not your thing, you can use EWM (Early Warning and Domain Monitoring), where a robot will set up everything you need.

For companies that need to monitor their devices, services or websites on a large scale, we already have an API available. However, we don’t have a bulk import of controls ready yet, but this is also in the pipeline. The service is smart enough to group all the services it checks according to various criteria and assess, for example, if there are slow sites on the same IP address or if there is a network problem in your network range. And it informs you about all this in a timely manner and can automatically publish it on the website.

Conclusion

Sometimes we think it would be worth publishing a list of new interesting customers and their projects. Up until now, we have taken it to mean that we don’t want to advertise at the expense of our customers. On the other hand, we have already encountered opinions that many of them would welcome it. You know, all advertising is good and a backlink is always useful 😉

We have customers with billion-dollar turnovers – no exaggeration. They also have their main websites or email solutions with us. The funny thing is that they usually have our services to choose from and we don’t know about them. We have no negotiations with them and no tenders. They have the same services as you and pay the same money as you…

We’ll see…