Scalability and resiliency IaaS-solutions Dynamic Data Center from TelecityGroup
In the wake of interest in cloud services, many data center operators have deployed cloud platforms on their sites and have begun to actively promote "clouds" to the masses. TelecityGroup Finland is no exception. I bring to your attention a review of our IaaS solution called Dynamic Data Center .
Solution Chart Dynamic Data Center
TelecityGroup Finland knows about the “clouds” firsthand. The first cloud platform was deployed back in 2008, and during this time various hardware resources and options for building a cloud infrastructure were tested. Earlier this year, the platform was upgraded, and the IaaS solution offered to customers was called the Dynamic Data Center, or DDC for short. Why Dynamic Data Center, you ask? First of all, I wanted to reflect the scalability and the possibility of flexible capacity expansion. In addition, the services offered by various cloud providers are replete with terms such as “virtual server”, ”IaaS, PaaS, SaaS”, ”cloud”, which sometimes is nothing more than a marketing ploy.
')
What distinguishes the DDC service from others?
The main “chips” are the complete fault tolerance of the solution, unsurpassed storage system performance, an additional services package, a variety of resource accounting and a convenient self-service portal. But first things first.
Fault tolerance of the solution
It is provided at all levels, from the physical site and communication channels to the innovative storage system and server blade farms.
The physical platform on which the virtual infrastructure is deployed is no less important than the “stuffing” of the platform. If the data center is “up”, no penalties will save your critical IT systems from interruptions. The DDC solution is deployed on the basis of two own data centers TelecityGroup Hiomo and Uspenski . Both data centers have been certified to ISO 27001 , which serves as an additional confirmation of the reliability of information protection, as well as the availability of a set of measures to identify and eliminate potential risks. At the request of customers, you can replicate virtual machines in an additional data center, where they will run when the main data center is unavailable. Backups are stored in the third, physically remote, data center Suvilahti .
The hardware filling of the platform is represented by IBM Flex System blade servers and storage connected via SAN (Storage Area Network) using FC technology (Fiber Channel). The solution is deployed on the basis of VMware, the world leader in virtualization. Thanks to VMware High Availability, when a single physical server crashes, the VM automatically restarts on the other (for more details on the VMware website). In addition, a backup copy of the image of each virtual is created daily, from which you can restore a “fallen” VM at the request of customers. In addition, we provide an agent-based backup service, details below.
As for the storage systems, they differ depending on the category of disk resources. For the vStorage Silver category, this is, for example, the IBM DCS3700 and Hitachi. For the Gold category - IBM XIV Storage System , which I want to talk about in more detail.
With its innovative grid architecture, the IBM XIV Storage System provides unmatched performance and scalability. Grid-architecture consists of a set of similar components, united by a common interconnect. The disks in IBM XIV behave like “communicating vessels”, due to which their uniform load is achieved. All IBM XIV components are reserved under the Active-Active N + 1 scheme, ensuring a high level of reliability and availability. In case of failure of any component, the fastest recovery is ensured.
Comparison of traditional storage systems and IBM XIV Storage System
A few words about the channels of communication. Communication with the “outside world” is provided by our own 10 Gbps fiber-optic network connecting all data centers with redundant connections. In case of failure of one of the communication channels, an automatic switch to the other takes place. By default, a guaranteed bandwidth of 100 Mbps is provided, which can optionally be extended to 10 Gbps. For large customers, connection of dedicated communication channels can be arranged. Let me remind you that the data centers TelecityGroup - with neutral accessibility. If you want a specific communication provider on the site - always welcome. The Dynamic Data Center solution deployed on the basis of Finnish data centers → delays on communication channels to Russia (in comparison with other European countries) are insignificant.
Unsurpassed Disk Subsystem Performance
We were interested to compare the performance of our storage with other cloud providers. The test results were pleasantly surprised. For ethical reasons, I will not publish the names of a number of companies, but in general, I think the picture is clear. I will note (so that we are not reproached with bias) that now a similar independent test is being conducted by one independent company, of which we will definitely report the results.
Test results for random reading and writing blocks of 4 kilobytes
For testing, the utility FIO (Flexible I / O tester) was used . VM 2 vCPU, 2GB RAM, CentOS 6.4 64bit was installed.
More detailed statistics on our storage systems for levels Gold and Silver
Package of additional services
A range of additional services is provided in conjunction with the DDC service:
• data backup, two options: Advanced and Simple
For the Advanced option on the basis of the Commvalt Simpana software, a self-service portal is available, where you can manage the backup schedule yourself, back up and restore data without a provider. The Simple option is deployed based on Veeam Backup & Replication and provides the ability to backup a VM image with a storage horizon of 7 or 30 days at a remote site.
• Managed Firewall service
3 levels of service provision, from basic to advanced with support for IPsec tunneling and performance up to 200 Mbps.
• load balancing based on Citrix Netscaller
• hybrid solution (physical + virtual infrastructure)
• connection of dedicated communication channels
• server monitoring
Variety of resource accounting
TelecityGroup always for an individual approach to the client. We are pleased to offer various options for accounting for virtual resources, and always pre-conduct an audit of the current IT infrastructure of the customer to select the most optimal option.
Option 1: the upper and lower limits of the allocation of computing resources are individually negotiated. The lower limit is guaranteed and the resources between the upper and lower limits are allocated on demand and are paid upon consumption. This ensures payment flexibility and scalability. Disk resources are paid on the fact of consumption.
Option 2: a pool of resources with a fixed monthly fee for customers who can predict the amount of virtual capacity they need.
Option 3: Full-scale Pay-as-you-Go model with hourly billing.
Billing is automated, and in a short time it will be possible to view the statistics of spending funds directly on the self-service portal. Now you can get information from our support service, which serves your interests around the clock, seven days a week. Incidentally, regardless of the option prepayment is not required.
Convenient self-service portal
The self-service portal allows the customer to fully control IT resources deployed in the cloud.
Portal Start Page
It is possible to manage all levels of the virtual infrastructure, from one virtual machine (VM) to a VM group (vApp) and a whole virtual data center (vDC). For each vDC (for example, for Internal IT vDC and Production vDC), an optional billing and its own SLA are provided.
vDC consisting of four vApps
Through the portal you can:
• launch new virtuals, including using templates: both own and TelecityGroup templates;
• suspend and shut down the VM;
• select a physical site for each vDC;
• change server configuration and redistribute resources across multiple VMs;
• set the sequencing of VM launch within the framework of vApp;
• install the OS;
• manage network configuration;
• manage user rights.
Statistics of consumption of computing and disk resources for a certain period is available on a separate portal . Data are available both on the average consumption of resources per month and on the consumption of resources in hours.
Statistics is updated every hour.
From theory to practice (Or test drive)
Forewarned is forearmed. Agree that it is better to check the performance of the disk subsystem, the functionality of the cloud platform and communication channels before the actual migration to the “cloud”. Gentlemen, who want to check the “cloud” in action, we always meet you.
To access the DDC demo version, just send an e-mail to habrahabr@telecity.com or fill out the HERE form.
For those who wish, a resource pool of 10 vCPU, 10 vRAM, 50GB of vStorage Gold and 50 GB of vStorage Silver will be provided.
The validity of the demo version - up to two weeks.
Solution Chart Dynamic Data Center
TelecityGroup Finland knows about the “clouds” firsthand. The first cloud platform was deployed back in 2008, and during this time various hardware resources and options for building a cloud infrastructure were tested. Earlier this year, the platform was upgraded, and the IaaS solution offered to customers was called the Dynamic Data Center, or DDC for short. Why Dynamic Data Center, you ask? First of all, I wanted to reflect the scalability and the possibility of flexible capacity expansion. In addition, the services offered by various cloud providers are replete with terms such as “virtual server”, ”IaaS, PaaS, SaaS”, ”cloud”, which sometimes is nothing more than a marketing ploy.
')
What distinguishes the DDC service from others?
The main “chips” are the complete fault tolerance of the solution, unsurpassed storage system performance, an additional services package, a variety of resource accounting and a convenient self-service portal. But first things first.
Fault tolerance of the solution
It is provided at all levels, from the physical site and communication channels to the innovative storage system and server blade farms.
The physical platform on which the virtual infrastructure is deployed is no less important than the “stuffing” of the platform. If the data center is “up”, no penalties will save your critical IT systems from interruptions. The DDC solution is deployed on the basis of two own data centers TelecityGroup Hiomo and Uspenski . Both data centers have been certified to ISO 27001 , which serves as an additional confirmation of the reliability of information protection, as well as the availability of a set of measures to identify and eliminate potential risks. At the request of customers, you can replicate virtual machines in an additional data center, where they will run when the main data center is unavailable. Backups are stored in the third, physically remote, data center Suvilahti .
The hardware filling of the platform is represented by IBM Flex System blade servers and storage connected via SAN (Storage Area Network) using FC technology (Fiber Channel). The solution is deployed on the basis of VMware, the world leader in virtualization. Thanks to VMware High Availability, when a single physical server crashes, the VM automatically restarts on the other (for more details on the VMware website). In addition, a backup copy of the image of each virtual is created daily, from which you can restore a “fallen” VM at the request of customers. In addition, we provide an agent-based backup service, details below.
As for the storage systems, they differ depending on the category of disk resources. For the vStorage Silver category, this is, for example, the IBM DCS3700 and Hitachi. For the Gold category - IBM XIV Storage System , which I want to talk about in more detail.
With its innovative grid architecture, the IBM XIV Storage System provides unmatched performance and scalability. Grid-architecture consists of a set of similar components, united by a common interconnect. The disks in IBM XIV behave like “communicating vessels”, due to which their uniform load is achieved. All IBM XIV components are reserved under the Active-Active N + 1 scheme, ensuring a high level of reliability and availability. In case of failure of any component, the fastest recovery is ensured.
Comparison of traditional storage systems and IBM XIV Storage System
A few words about the channels of communication. Communication with the “outside world” is provided by our own 10 Gbps fiber-optic network connecting all data centers with redundant connections. In case of failure of one of the communication channels, an automatic switch to the other takes place. By default, a guaranteed bandwidth of 100 Mbps is provided, which can optionally be extended to 10 Gbps. For large customers, connection of dedicated communication channels can be arranged. Let me remind you that the data centers TelecityGroup - with neutral accessibility. If you want a specific communication provider on the site - always welcome. The Dynamic Data Center solution deployed on the basis of Finnish data centers → delays on communication channels to Russia (in comparison with other European countries) are insignificant.
Unsurpassed Disk Subsystem Performance
We were interested to compare the performance of our storage with other cloud providers. The test results were pleasantly surprised. For ethical reasons, I will not publish the names of a number of companies, but in general, I think the picture is clear. I will note (so that we are not reproached with bias) that now a similar independent test is being conducted by one independent company, of which we will definitely report the results.
Test results for random reading and writing blocks of 4 kilobytes
For testing, the utility FIO (Flexible I / O tester) was used . VM 2 vCPU, 2GB RAM, CentOS 6.4 64bit was installed.
More detailed statistics on our storage systems for levels Gold and Silver
Package of additional services
A range of additional services is provided in conjunction with the DDC service:
• data backup, two options: Advanced and Simple
For the Advanced option on the basis of the Commvalt Simpana software, a self-service portal is available, where you can manage the backup schedule yourself, back up and restore data without a provider. The Simple option is deployed based on Veeam Backup & Replication and provides the ability to backup a VM image with a storage horizon of 7 or 30 days at a remote site.
• Managed Firewall service
3 levels of service provision, from basic to advanced with support for IPsec tunneling and performance up to 200 Mbps.
• load balancing based on Citrix Netscaller
• hybrid solution (physical + virtual infrastructure)
• connection of dedicated communication channels
• server monitoring
Variety of resource accounting
TelecityGroup always for an individual approach to the client. We are pleased to offer various options for accounting for virtual resources, and always pre-conduct an audit of the current IT infrastructure of the customer to select the most optimal option.
Option 1: the upper and lower limits of the allocation of computing resources are individually negotiated. The lower limit is guaranteed and the resources between the upper and lower limits are allocated on demand and are paid upon consumption. This ensures payment flexibility and scalability. Disk resources are paid on the fact of consumption.
Option 2: a pool of resources with a fixed monthly fee for customers who can predict the amount of virtual capacity they need.
Option 3: Full-scale Pay-as-you-Go model with hourly billing.
Billing is automated, and in a short time it will be possible to view the statistics of spending funds directly on the self-service portal. Now you can get information from our support service, which serves your interests around the clock, seven days a week. Incidentally, regardless of the option prepayment is not required.
Convenient self-service portal
The self-service portal allows the customer to fully control IT resources deployed in the cloud.
Portal Start Page
It is possible to manage all levels of the virtual infrastructure, from one virtual machine (VM) to a VM group (vApp) and a whole virtual data center (vDC). For each vDC (for example, for Internal IT vDC and Production vDC), an optional billing and its own SLA are provided.
vDC consisting of four vApps
Through the portal you can:
• launch new virtuals, including using templates: both own and TelecityGroup templates;
• suspend and shut down the VM;
• select a physical site for each vDC;
• change server configuration and redistribute resources across multiple VMs;
• set the sequencing of VM launch within the framework of vApp;
• install the OS;
• manage network configuration;
• manage user rights.
Statistics of consumption of computing and disk resources for a certain period is available on a separate portal . Data are available both on the average consumption of resources per month and on the consumption of resources in hours.
Statistics is updated every hour.
From theory to practice (Or test drive)
Forewarned is forearmed. Agree that it is better to check the performance of the disk subsystem, the functionality of the cloud platform and communication channels before the actual migration to the “cloud”. Gentlemen, who want to check the “cloud” in action, we always meet you.
To access the DDC demo version, just send an e-mail to habrahabr@telecity.com or fill out the HERE form.
For those who wish, a resource pool of 10 vCPU, 10 vRAM, 50GB of vStorage Gold and 50 GB of vStorage Silver will be provided.
The validity of the demo version - up to two weeks.
Source: https://habr.com/ru/post/221873/
All Articles