What you need to know about the IaaS-provider even before the start of cooperation

/ photo Dennis van Zuijlekom CC

“Let's face it - technical aspects of the service, such as uptime, scalability and security, are not unambiguously key factors in the decision to choose a provider. Representatives of businesses that are not directly related to technology are not eager to take these factors into account.
')
They pay attention only to the promises (often in words) of the IaaS provider to remove from the company the burden of responsibility for the performance and stability of its IT infrastructure. Despite the prevalence of this approach, we want to caution you - if you expect to delegate some of the responsibility on such critical issues to the cloud service, think twice, and preferably three times. ”

Don Sullivan (VMware) and Michael Corey (Ntirety) gave such advice at the last Data Summit in New York, advocating a cautious attitude to the promises of IaaS providers to take care of customers' technical needs.

Corporate cloud service users should be proactive and find out if the provider will really pay attention to emerging issues. Corey and Sullivan well defined the basic idea of ​​the division of responsibility: “When a service is unavailable or does not work correctly, it should concern the provider to the same extent as you as a client.”

In addition to these theses, representatives of the virtualization industry have compiled a list of questions that should be answered reasonably before signing a contract with an IaaS provider. This list is designed to help you in separating the “chaff” seeds. We, the team of the virtual infrastructure provider 1cloud.ru , decided to immediately answer these questions in the framework of the self-test.

1. Can you provide three real feedbacks from your customers (not those whose legs grow from your PR activity)?

Below are three reviews from customers left on our site:

“Thank you for your hosting! We came to you on the recommendation and decided to stay! Adequate flexible prices. Competent responsive technical support. Convenient payment options, at least in an hour. Reliable servers. Highest uptime. For the Russian market - 1cloud.ru - just a godsend! Highly recommend"

- Aaron Urkin, Co-founder of the Real-Control.Pro Real Estate Sales Management Service

“Terrific simple and professional service. I began to meet him with a convenient wizard for choosing a configuration. Everything else was just as pleasant and clear.

We in our web studio adhere to exactly the same approach in working with clients! The speed of work, professionalism and courtesy of support - all you have at the highest level. For confirmation of this review or any other consultations in the field of web development or SEO you can contact me personally at i@bizzcon.ru ”

- Denis Ivanov, PM bizzcon.ru

“I have been using the 1Cloud service for a long time. I used to use Azure. Moved for the following reasons:

1. Substantially cheaper.

2. Servers are much faster deployed.

3. There are many options for systems ... for tests there is everything that is needed.

4. The cloud itself is Russian, which is critical for our company (ping is stable and minimal).

5. Excellent support, everything is decided almost immediately.

6. With payment, too, everything is convenient and transparent. There are no hard restrictions and cuts (for example, in azure, if the credit limit has been spent, then everything stops working until the month ends.

In general - thanks for the work, the service is very like "

- Ivan Koroblin, Consulting LLC System Administrator

All feedback from our customers, both positive and negative, is posted on our website and on other sites, for example, by hosting101 . We do not immediately request feedback from customers, but after 3 months of using our service, when their opinion of us has already been formed.

2. Can you formulate the technical specifics of our company's business?

In terms of automating the management of the IaaS environment, it is difficult to isolate the specifics of each client’s business. We analyze only the information received from the client: contacting technical support and online chat, indicators of demand for specific categories of services, reviews at various sites (by the way, you can also share a review or suggestion about the work of our service, for example, here ).

If the representatives of the client company wish to share information about the specifics of their work, we take this information into account during the formation of individual proposals and in the process of customer service.

3. Can I add / remove services as needed?

Yes, all services offered by us can be connected / disconnected at any time through the control panel. Charging is carried out every 10 minutes, so you do not need to pay even one extra hour of using the service, in which there is no need. If you turn off the virtual server, the funds will be charged only for the leased disk space.

Processing tasks to add or remove options occurs automatically in a few minutes. For example, in a couple of clicks, custom server templates are created, and new virtual machines from these templates are also deployed. Similarly, a private network is created (knowledge of VPN is not required).

In a couple of clicks, our customers can also change the hardware configurations of the servers themselves for current needs: increase CPU, RAM, HDD “on the fly”. Below are the order calculator for the new virtual server (top), and the resource control panel configurator (below). We developed these tools precisely to ensure that we can use only those capacities that are needed at the moment.





4. Do you get bonuses from manufacturers for lobbying any specific hardware or software brands? If so, under what conditions?

All services offered by us are based on proven enterprise-solutions from well-known global manufacturers. Today our main equipment suppliers are Cisco (UCS, ASR, Nexus), Dell (PowerEdge), NetApp (FAS8000). For virtualization used vSphere from VMware. On this page you will find a detailed overview of our equipment.

Since we provide our customers with a ready-made service at a fixed price, then in principle there can be no interest in lobbying any particular supplier for us. A vendor either satisfies our requirements for stability and flexibility - then we use its solutions in our infrastructure, or it does not satisfy - then we simply don’t work with it.

In our opinion, choosing a different approach, the hosting provider risks, at a minimum, incurring additional costs associated with servicing imposed from the outside, rather than self-designed solutions, and, as a maximum, disappoint customers with a service that builds on the needs of producers rather than users. Perhaps the vendor will be ready to cover all these “side” effects with the offered bonuses — anything can happen. But such an approach goes against our business model, which is based on the principle of customer orientation.

5. Is your own infrastructure also hosted on cloud platforms or do you use other solutions?

We also use all the technical solutions offered to our clients for our internal corporate needs - we carry out almost all the work on developing and maintaining the service in the same IaaS 1cloud. All company employees work with virtual machines - this greatly facilitates the process of remote interaction and removes the dependence of the team's performance on specific workstations and devices.

6. When did the last idle time of your service occur and how long did it last?

As with any company, sometimes there are "drops" of some parts of the infrastructure. The last downtime happened about 2 months ago, lasted less than 15 minutes, and affected about 10 percent of virtual machines that were unavailable over the network.

The reason for the downtime was getting into our internal network of excess traffic as a result of DDoS-attacks. This led to a partial loss of packets between the nodes of the virtual router. As a result, the high-availability system on the router failed: the slave nodes received false information about the unavailability of the master node and took over its function.

The problem was quickly resolved, to prevent its manifestation in the future, node status monitoring was promptly configured, and then a system of in-depth analysis of traffic and protection against DOS and DDoS attacks “ Skat ” was introduced. During and after the incident, we maintained constant communication with customers. Thanks to this, it was possible to gain valuable experience regarding actions in such situations and work to prevent them.

If for some of the clients we fail to ensure the availability of the service within the SLA , then we pay this client a monetary compensation, the amount of which is set by the same SLA.

7. Does the architecture offered to you be diversified enough to offset the risks of downtime? Due to what?

All nodes of our infrastructure are duplicated, which greatly reduces the risk of downtime. The equipment used has high availability, security and performance (you can get more detailed information about the components we use here : Cisco UCS, Dell, NetApp). In the organization of communication channels, redundancy is used with separation across different providers and physical routes.

1cloud virtual servers are located on NetApp FAS6240 / FAS8040 disk arrays equipped with Intel chipsets that use high-performance memory and NVRAM and guarantee high availability (> = 99.999%).

Each controller has a service processor, isolated from the main controller. It will continue to work even if the main part of the board fails. The equipment is located in data centers SDN (St. Petersburg) and Dataspace (Moscow), which have passed Tier lll level certificates.



NetApp Storage Expansion Shelves

8. What are the advantages of your SLA against competitors?

We do not undertake to evaluate competitors. We have a service level agreement, it is published on the website and implies real financial guarantees from our side. For example, the established level of equipment availability (uptime) by agreement is 99.9%. If our services do not correspond to this indicator, we pay the client a monetary compensation (5-100% of the cost, depending on the actual availability indicators).

9. Do you have a single management tool for all the services offered?

The client can fully control his IaaS through a single control panel with wide functionality. as well as through our API . Using these tools, you can create and delete servers, create snapshots and backups, deploy private networks and manage resource records of domain zones.



You can get a key to access the 1cloud API from the control panel at any time.

Thanks to the REST API, a full automation of the tasks of managing the user infrastructure is available without the need to perform manual actions via the 1cloud web panel.

For habrovchan who decide to use our API, we are ready to provide an extended test period (3 days for individuals. 5 days for legal entities) and provide full support (to activate a test period, just contact support at support@1cloud.ru ).

10. Do you offer one common bill for all services?

Yes, payment for services is made on the total bill. The client has one account from which funds are written off to pay for all services used as they are consumed. At any time, you can explore the history of debits: when, for what and how much the user paid, for how many days to use the service, there are enough funds in the account. In the control panel, you can in a couple of clicks confirming the payment detailed acts.

Here we talk about how we optimized the procedure for receiving and controlling payments. And since we work both with individuals and with client companies, we have also taken care of automating the formation of invoices and acts for legal entities. On the one hand, this direction cannot be called profile or critical for the provider, on the other hand, the possibility of obtaining full information about the account balance and copies of accounting documents at any time makes life easier for our corporate clients.



The process of generating acts on the accounts of our corporate clients

11. Do you have a unified system for monitoring the state of client infrastructure and notification of failures?

Automated monitoring systems register an alert in case of failure of an infrastructure component. Service desk specialists carry out work to eliminate such failures around the clock. If necessary, second-line support specialists are involved, who receive the already detailed request directly from the first-line staff.

This helps to make the malfunctioning more operational. As for the monitoring tools intended for the client himself, we wrote about it in our blog (you can read it here ).

12. What are your stated fault tolerance and disaster recovery rates?

These parameters are described in detail in our SLA . Based on the declared availability level - 99.9%, it can take no more than 2 hours to recover from failures, and this is a monthly rather than a one-time limit.

13. Is the Service Level Agreement (SLA) supported by your real financial responsibility?

Yes, we undertake financial obligations that guarantee the refund to the user (5-100%) if the service does not match our SLA . The amount of compensation is fixed and published in the above agreement.

14. Are virtual machines replicated and, if so, for how many copies?

The computational capacities of the servers (CPU, RAM) are provided from the fault-tolerant cluster and are separated from the disk arrays, so replication of virtual servers is simply not necessary. In the event of a hardware host failure, all virtual servers are restarted on another guests.

As for the replication of disk arrays, the client data is duplicated on three physically deleted disks. The workload balancing on the available virtual resources is implemented by means of VMware DRS, which provides the ability to deploy new capacities or migrate virtual machines without the need to suspend the service (for more details, click here ).

15. Which hypervisor do you use?

We use vSphere from VMware. This platform suits us both in terms of reliability and in terms of its flexibility.

16. What professional certificates do you have?

Our employees have various professional certificates, in particular: Cisco Certifications, NetApp Certificate of Completion, Microsoft Certificate of Excellence, numerous FastLane certificates.

17. Do you offer cross-platform solutions?

Yes. We offer servers running the Windows, Linux, BSD families. To work with our control panel you only need a web browser. The panel has a web console to connect to the server, which allows you to access it even in the absence of the necessary utilities on the client device.

18. How is your customer support team structured and motivated?

We work on the principle of “empty inbox folder”, i.e. if the client’s appeal for a long time remains unanswered, this is considered a problem. We also have internal metrics for the quality of technical support service, such as the established maximum response time and time to resolve the issue.

There are two lines of technical support: the first line consists of specialists with basic knowledge of Windows / Linux OS administration. These employees can monitor the infrastructure of 1cloud and understand the principles of its construction.

The second line is made up of employees who have in-depth knowledge of these issues, as well as all the skills to work with VMware software (vCloud Director, vSphere, etc.) and are professionally versed in network infrastructure and data storage systems.

If necessary, we can form an emergency conference at the level of the third line - responsible specialists of the company who work with the most complex and interesting questions, perfectly understand the specifics of 1cloud architecture, have broad competences in system administration and development on .NET.

19. Do you have quality control systems for customer support? How do they work?

We constantly collect customer feedback on our service, including the quality of technical support. The company's management weekly selectively checks a number of client requests: their current status and processing history.

20. What is the procedure for responding to emergency customer requests?

In the case of 1cloud.ru, an emergency client request can only be associated with the unavailability of the service. , Major-: - ( , , , , IT-), «» .

, – DDoS-. 15 : , , 5 , , 3- .

( , ), .

21. ? How?

, (, ). In this way, for example, we arrived at a three-tier structure of technical support.

, , , . , .