Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure
Migrating and managing your data storage in the cloud can offer significant value to the business. Start by making good strategic decisions about moving data to the cloud, and which cloud storage management toolsets to invest in.
Your cloud storage vendor will provide some security, availability, and reporting. But the more important your data is, the more you want to invest in specialized tools that will help you to manage and optimize it.
First, know if you are moving data into an application computing environment or moving backup/archival data for long-term storage in the cloud. Many companies start off with storing long-term backup data in the cloud, others with Office 365. Still others work with application providers who extend the application environment to the vendor-owned cloud, like Oracle or SAP. In all cases you need to understand storage costs and information security such as encryption. You will also need to decide how to migrate the data to the cloud.
Second, once you are storing data in the cloud you need to manage it for cost and security. Never assume that your cloud providers are managing your applications for you. They will protect your active data with backup and built-in redundancies and are obligated to provide certain security measures. But if you need strong security like 2-factor authentication, or need to prove compliance with cloud storage, then you will probably need to invest in third-party cloud-based software.
The management of cloud storage migration poses plenty of complex choices.
Migrating data to the cloud takes a lot of work and some level of risk, and you need to carefully weigh benefits and risks before taking on the cloud storage migration project. You don’t have to go it alone; migration tools are available to assess applications for large-scale migration. Understand benefits and risks, and what is involved in migrating a specific application to the cloud.
· Scalability. The scalability of the cloud is one of its major assets. If your backup and archival data is consuming large amounts of storage, then the cloud can provide cost-effective scalability with dynamic provisioning. The same benefit holds for large volumes of active application data, such as video and games.
· Lower costs - maybe. Cloud computing can reduce operational costs by shifting to a subscription model. The company also spends less capital buying new storage equipment, and IT spends less time managing on-premise storage. However, watch your cloud storage costs: storing data above your basic capacity threshold gets very expensive, and frequent data retrieval comes with significant costs.
· Off-load some management tasks. IT spends significant time provisioning, load balancing, and capacity planning for stored data. IT retains responsibility for their cloud-based data, but can off-load tasks like storage provisioning, troubleshooting, and patching/upgrading storage intelligence tools.
· Compliance and data protection issues. Even if you have migrated your data to the cloud, you are ultimately responsible to prove compliance. You are also responsible for protecting your data long-term, ideally moving it into a searchable archive for eDiscovery and compliance.
· Proprietary technology. If you use proprietary technology in your on-premise such as WORM compliance, you might need to look at 3rd party vendors to supply that feature in the cloud.
· Latency. Although cloud-based performance can be very fast, there may be bandwidth and latency issues between clouds and on-premise. When you want the cloud as a backup target, clarify that cloud-based performance will be as good or better than on your network.
· Multi-tenant issues. Unless you pay for a dedicated environment, your stored data may be at risk in multitenant settings.
· Vendor lock-in. Once you have migrated data to the cloud, it might be difficult to move between platforms. Understand what cloud to cloud migration options are available to you.
Two primary cloud storage models are 1) long-term cold storage, and 2) storing active data on cloud computing platforms, like SaaS, PaaS and IaaS.
Storing long-term backup data to the cloud seems straightforward but you have some decisions to make. First, look for the ability to store data in cost-effective cold storage. The Big 3 hyperscaled clouds – Amazon, Azure, and Google – all offer cold storage tiers. But their storage and retrieval costs are different. Understand your pricing for basic storage capacity, above-threshold capacity, and the costs and consequences of data retrieval.
Also nail down if you can use third-party software to add value to your long-term stored data, such as WORM storage and searchability for eDiscovery and compliance.
Cloud computing platforms: SaaS, PaaS, and IaaS
If you are storing active data in a cloud computing environment, then you have some additional concerns around your data. There are three primary computing services in the cloud: Software as a Service (SaaS), Platform as a Service (PaaS), or Infrastructure as a Service (IaaS).
· SaaS. SaaS platforms will back up your active data – say within 30 days or so – but beyond that you are on your own. Remember that cloud providers are primarily responsible for availability, not long-term data durability. Understand your provider’s security and data protection measures for stored data. For example, SaaS giant Microsoft does not back up any of its Office services past 30 days, and most are considerably less. Look to third-party cloud backup and archival products that integrate with SaaS applications to protect your SaaS data long-term.
· PaaS. Your cloud provider may or may not provide long-term data protection to your development platform data. Understand what they are responsible for and fill in with third-party data protection products that integrate with your cloud vendor’s PaaS offerings.
· IaaS. One common use case for IaaS is migrating a corporate application and data to the cloud for simple rehosting instead of recoding. Since the organization is not making major changes to the application code base, it’s technically simple to migrate the code and its data with some configuration changes. The issue is it’s going to take time – a lot of time -- to migrate data to the new cloud-based infrastructure. Build in reasonable timeframes and data priority when you schedule the migration, which could stretch into months.
Before backing up or migrating large datasets to the cloud, do a proof of concept (POC). Run a series of tests including:
· Test performance milestones. Backup must occur with backup windows without impacting local computing performance. Test performance/latency at different times of the day, not just at night. Regular backup performance might be fine at night, but if you deploy continuous backup for high priority applications, you are going to need good performance any time of the day and night. If latency issues pop up, it’s better to know now then later.
· Test the complexity of data migration and management. How time-consuming was it to migrate data to the cloud? Backup and archival migration may be relatively simple, but once in the cloud how easy is it to manage?
· How good was the support from your cloud provider? Vendors are trying to sell you at the POC stage, so support will probably be the best you will ever get. Before signing on the dotted line, ask your peers about how good customer support really is. While excellent support bodes well for the future, indifferent support at this stage will only get worse.
When you migrate data to the cloud, you still need to manage it. This comes as a surprise to some businesses, who expect the cloud provider to take on most of management responsibilities.
The truth is that the business is ultimately responsible for their data. You can write management responsibilities into your service level agreement (SLA), but you must be very clear what you were asking for and getting.
In any case, the business will need to take an active role in managing their data for capacity, cost, data protection, compliance, and performance. Even if you write those responsibilities into your SLA, you are still ultimately responsible for protecting and securing your data in the cloud. This is where cloud management tools can be an important investment.
Some toolsets work across multiple clouds and others are dedicated to a specific cloud. Balance breadth and depth: if you have a multi-cloud strategy and store data in different clouds, then you want a toolset that gives you general storage information across all your cloud deployments. Or if you are exclusively or primarily in a single cloud, opt for a tool that is optimized for that cloud. Common data management toolset features include:
· Compliance. These tools manage cloud security and reporting requirements for SOX, PCI, HIPAA, and others. Most cloud providers offer some level of compliant security and reporting, but compliance toolsets enable customers to customize policies and get in-depth reports on sensitive data.
· Security. Basic security management in the cloud includes data access control and encryption. Specialized toolsets protect stored data with premium security management features such as key management, two-factor authentication, managing security across instance, data, and user access.
· Cost tracking. Cost-tracking metrics protect the customer from an unpleasant billing surprise. Cost tracking monitors data usage and cost, which allows you to adjust your services and control your expenses. This comes in especially handy when your cloud provider charges high prices for storing data over a basic capacity agreement.
· Consolidation. These toolsets pool storage resources from multiple clouds into a single platform. Others integrate compute, storage, networking, and more into a single console for centralized management and reporting across your cloud-based applications and storage.