Join Newsletter
Trusted Business Advisors, Expert Technology Analysts

Taneja Blog

Taneja Blog / Cloud / Primary Storage

Cloud-Based Storage and Service Levels

Cloud-based applications (also called SaaS) and storage in the cloud both require making good service levels agreements with the cloud provider. The former is not particularly difficult to do. Customers must agree on uptime percentages, bandwidth requirements and performance speed of course but these are relatively straightforward.

Storing corporate data in the cloud is a larger can of worms. Requirements shift according to the type of data stored and its usage. Let’s look at the top four types of cloud-stored data before we talk more about storing to the cloud.

#1. Production data: Digital delivery to online customers. This is the realm of start-ups and some established production companies where building primary data centers would be cost-prohibitive. When Google or Amazon goes down, these are the customers who scream the loudest because their business is SOL.

#2. Production data: Hosted email applications. These vendors host email applications and primary email storage online. The advantages are security, searchability and a cost-effective method of handling email installations. Email is a critical application and uptime must be superb. Email vendors are also paying more attention to enabling email searchability and preservation for litigation requests.

#3. Secondary data: Hosting data for applications. This is hosted non-production data that is subject to active applications. Common examples include heavy number-crunching analysis on inactive data.

#4. Secondary data: Backup and archive. Backup to the cloud requires a backup application or third-party application that enables cloud as a storage target. This is becoming increasingly popular in corporate settings and represents excellent economies if bandwidth and performance are acceptable. Archiving is the go-to program for long-term data retention and searchability. Storing archives in the cloud makes a good deal of sense because they are by nature long-term and should be quite searchable.

What all of these cloud usages have in common is required service level agreements, and therein lies the rub. Case in point: Eli Lilly, as in really big pharma. Eli Lilly used Amazon Web Services successfully for running intensive analysis on non-production data. The company decided that they wanted to add high value archives storage to Amazon. They required certain service level agreements, the same they would negotiate with their internal IT service. But Amazon would not comply, refusing to customize their standard service level agreement for this flagship customer. As a result they lost this aspect of Eli Lilly’s business, and they suffered in the marketplace due to critical buzz.

The upshot is that Amazon felt that Eli Lilly’s requirements were too much risk exposure for them and they were willing to accept a major customer loss and bad publicity. Of course Amazon must negotiate agreements that work for them but they are trying to keep the wave from crashing on shore. Corporations will insist on service level agreements for their secondary data that will shift according to data usage. All of them should include security, activity support and availability.

  • Security applies to both digital and physical security. Digital security is protecting against data loss both from outside sources (such as hackers) and from internal errors (such as clueless storage admins). Physical security depends on strong data center protections against human intervention and natural disasters.
  • Activity support is the ability to run operations against the stored data. Eli Lilly’s data analysis is a good example. Another popular usage is running eDiscovery and compliance searches, preservation and analysis against stored data. The services might come from the data owner, service providers, law firms, or even the hosted vendor.  
  • Availability includes both uptime and longevity. Requirements differ according to backed up or archived data. Backup availability should center on fast disaster recovery. When a company also uses backup for long-term data retention, they commonly tier their backup infrastructure. They will retain current backups on site or replicate to a company remote site, and will upload copies of backups to the hosted site. Instantaneous availability for archives is not critically important but the cloud vendor should provide good-to-excellent uptime, especially if email archives remain available for transparent end-user access. Vendor longevity is vital so customers can be certain their archives will remain safely stored and available long-term.

Corporations cannot be shy about demanding service levels from their cloud vendors. Granted that companies need to understand what is and is not reasonable when dealing with cloud vendors, who have to maintain profit and lower their own risk. But the corporation is absolutely driving storage and storage application development in the cloud, and cloud-based vendors need to provide service levels accordingly.

  • Premiered: 08/15/11
  • Author: Taneja Group
Topic(s): eDiscovery compliance Cloud service level agreement Amazon Google Eli Lilly Hosting Backup Archive


There are no comments to display. Scroll down to leave your own!


Leave a Comment

You must be logged in to comment. Click here to log in or register if you don't have an account.