Actifio GO for Google Cloud
Enterprise class SaaS data protection in Google Cloud and on-premises.
Actifio GO for GCP is a SaaS offering which enables powerful enterprise class backup and recovery for GCP resident and on-premises workloads. Actifio now supports backup, disaster recovery and rapid database cloning of Oracle on Bare Metal Solution on Google Cloud Platform besides other enterprise workloads including SAP HANA, SQL Server, and others.
Enterprise-Class SaaS Data Protection in Google Cloud and On-Premises.
Enterprises face many challenges with Data Protection in Google Cloud.
Enterprises have to deal with expensive cloud costs to protect workloads inside of Google Cloud and protect on-premises VM workloads in Google Cloud.
How Actifio Go for GCP Helps: Highly efficient incremental forever data capture minimizes required compute, bandwidth and storage. Advanced object storage support further reduces costs.
Enterprises need cost effective data retention while minimizing costs and maintaining data access
How Actifio Go for GCP Helps: Actifio Go for GCP leverages Google Nearline/Coldline for cost effective data retention and intelligent caching technology minimizes egress charges and API cost
Organizations must minimize RTO on databases to ensure rapid recovery from unexpected data outages.
How Actifio Go for GCP Helps: Deep integration with SQL Server and SAP HANA databases ensures application consistent data capture and recovery inside of Google Cloud.
Lifecycle management (patches, upgrades, etc.) for backup and dedupe virtual appliances running Google Cloud is cumbersome.
How Actifio Go for GCP Helps: Actifio Go for GCP takes care of all software updates thus simplifying management and reducing OpEx both in the cloud and on-premises.
Complex architecture of virtual dedupe appliances, backup servers and media servers.
How Actifio Go for GCP Helps:Simple to use SaaS model incorporates all needed functionality including application integration, SLA management and data replication.
1 / 5
Features & Benefits
Actifio GO For Google Cloud Features.
Self-service sign up, simple deployment and central management
Simple and easy to use SaaS platform to deliver backup and recovery of VMs, SQL Server and SAP HANA databases inside of Google Cloud and on-premises VMware VMs.
Application consistent incremental forever capture of databases and VMs
Deliver small and efficient backup windows while ensuring database recoverability inside of Google Cloud
SLA driven data management
Powerful SLA engine ensures consistent protection policies across on-premises and one or more Google Cloud regions.
Instant recovery on-premises or in the cloud
Minimize downtime by recovering VMs and databases in minutes regardless of size either on-premises or inside of Google Cloud.
Multi-region disaster recovery
Enable disaster recovery in a single Google Cloud region or across regions with the ability to instantly access 1000’s of datasets in parallel.
Data encryption in transit and at rest
Ensure that data is safely secured in block and object storage to meet business and regulatory requirements.
Frequently Asked Questions
Actifio GO for GCP Architecture
Actifio GO for GCP solution consists of the following components:
Actifio Global Manager (AGM): The management control plane that is automatically deployed in GCP. Users use AGM to setup backup SLAs, recover files/folders/VMs, DR in GCP. AGM is also used to administer, monitor, and manage one or more Actifio Sky data movers.
Actifio Sky Data mover: A VM appliance that is deployed in the customer’s data center to protect on-premises workloads, or customer’s GCP cloud project to protect GCP VMs. Sky is the core data mover engine that performs incremental forever backups, instant recoveries, instant database cloning.
Actifio Sky Data Mover leverages VMWare’s VADP API to protect VMWare VMs with application consistency & incremental forever backups. No agents need to be deployed in the Guest OS. Actifio Sky Data Mover also provides instant recovery for VMWare VMs, allowing VMs to boot directly off Actifio. Once booted these VMs can be moved to their production storage using VMWare’s Storage vMotion.
Actifio provides a light-weight connector that is required to be deployed on GCP VMs. Actifio Sky Data Mover establishes a secure connection with the connector to trigger application consistent incremental forever backups and instant recovery.
- Users can use simple google like search interface in Actifio Global Manager UI and perform granular file/folder restores.
- Users can perform instant mount of the backup image and access the entire database/volume/files/folders. The mounts are very fast as the data is not copied. To the recovery machine. Instead Actifio presents a virtual disk with all the data in it via iSCSI or NFS. With this approach, even a 10 TB database or volume can be accessed within minutes in GCP, thus reducing the RTO to minutes.
- Users can also perform full machine recoveries of on-premises VMs, physical servers and GCP VMs in any GCP region a s anative GCP VM. Actifio ensures that the system state is restored properly and mounts the rest of the non-system volumes instantly, thus reducing the RTO.
- And last but not the least, Actifio also allows users to perform the traditional restores where the data is restored from its backups to a recovery machine.
Actifio GO for GCP Differentiation
|GCP VM backup and recovery||Yes||Yes|
|Integrated Billing in GCP||Yes||Yes|
|Backup and Recover on-premises workloads to GCP||Yes||No||Others do not support in their SaaS offering|
|Application consistent & incremental forever backup of mission-critical databases||Yes||No||Others perform only crash consistent snapshot backups of databases|
|Instant Mount and Recovery from backups stored in GCP PD/NL/CL||Yes||No||Others can’t perform instant recoveries from persistent disk or nearline/coldline storage. They can perform instant mounts of native cloud snapshots but it takes too much time to hydrate data from snapshots to persistent disk thus increasing RTO.|
|Data Immutability option in software.||Yes||No||This capability in Actifio ensures even an admin, or a rogue user/admin can’t delete backups before their scheduled retention expiry.|
|Access data instantly even after Unsubscribing||Yes||No||Delivers peace of mind that data is always accessible even after unsubscribing from the SaaS. Eliminates lockdown to a vendor because of data retention needs.|
|Low Total Cost of Ownership||Yes||No||Unlike others, the option to store the majority of backups by Actifio in an incremental forever manner in Google NL/CL inexpensive object storage lowers the TCO.|
GCP Infrastructure Costs During DR
Each tenant uses their own cloud account. Thus each tenant is charged by GCP based on consumption.
Assuming 1000 VMs protected to GCP nearline storage, 400 restore jobs per month, 500 MB restored per restore job, the monthly cost would be approximately $26.00. Details below:
- 400 restore jobs x 500 MB per restore job = 200 GB of total data restored from the cloud per month.
- GCP nearline charges $0.01 per GB for data access. Thus data access charges = $0.01 per GB x 200 GB = $2.00 per month.
- Google also charges for the data that leaves GCP cloud to the internet at a rate of $0.12 per GB in USA. Thus network costs = $0.12 per GB x 200 GB = $24.00 per month.
- Thus total costs in the month = $2.00 access charges +$24.00 egress charges = $26.00
Note that it’s uncommon to restore dozens of entire VMs back to on-premises. This might happen in a situation where a farm of VMs got deleted or corrupted and there is no local backup copy.
Assuming GCP nearline was used to store backups, with each VM being 200 GB in size, when 50 VMs are recovered from GCP back to on-premises, the cost would be approximately $1,210. Details below:
- GCP nearline storage charges $0.01 per GB for data access, and network egress costs are $0.11 per GB in USA (for 1 to 10 TB)
- Total data required to be restored = 50 VMs x 200 GB per VM = 10 TB
- Access and Egress charges = ($0.01 access charges + $0.11 data egress charges) per GB x 10 TB = $1,200
- Assuming 1 MB object size was used for backup, # of objects read for recovering 50 VMs with 10 TB = 10 TB/1MB = 10,000,000 objects. GCP nearline storage charges $0.01 per 10,000 Read APIs. Thus, Read API costs = (10,000,000/10,000)*$0.01 = $10
- Total costs = $1,200+$10 = $1,210 for the specific incident when 50 VMs with each 200 GB size are recovered from GCP back to on-premises.
To recover 500 VMs on-demand in GCP with an average size of 200 GB per VM in Google Cloud Platform (GCP), with each DR test running for 3 days, using Actifio’s Instant mount of backups from GCP nearline storage, the total costs will be approximately $9,250.
Note that this approach delivers low RTO with post DR IO performance that is as good as what Google Nearline IO can deliver. Cost details below:
- For simplicity sake, assume all of them are RHEL and an average VM would need 4 cores and 15 GB of memory using n1-standard-4.
- The compute costs for 3 days would be approximately $9,000 per DR test.
- With Actifio GO for GCP, you can mount data instantly from GCP nearline without having to copy the data to block storage. Let’s assume that as part of your DR testing, you access 25% of the total data. Total data size = 500 VMs x 200 GB per VM = 100 TB. 25% of 100 TB = 25 TB. The data access charges for 25 TB = $0.01 per GB x 25 TB = $250.
- Total costs for each DR test = $9,000 for compute + $250 for data access = $9,250
To recover 500 VMs on-demand in GCP with an average size of 200 GB per VM in Google Cloud Platform (GCP), with each DR test running for 3 days without using Actifio instant mount, the total costs will be $11,780.
Note that approach increases RTO because of copy from NL to PD storage and delivers post DR performance that is as good as SSD PD. Cost details Below:
- As explained above for 500 VM DR in GCP, the on-demand compute costs will be $9,000 per DR test.
- In this use case, data needs to be ‘restored’ from nearline to persistent disk. Assume SSD class persistent disk in GCP is used. 100 TB of SSD persistent disk storage costs $560 per day. Thus block storage costs for a DR test = $560 per day x 3 days = $1,680 per DR test.
- Data access charges from object storage (during restore) = 100 TB x $0.01 per GB = $1,000 per DR test.
- Read API costs to read 100 TB from Google Nearline Object Storage = $100
- Total charges per DR test = $9,000 for compute + $1,680 for block storage + $1000 for data access from object storage + $100 for Read APIs = $11,780
Note that cloud infrastructure costs will vary in different regions and might change over time and will vary based on many factors. The examples are given below to provide rough approximations.