{"id":2972,"date":"2014-12-17T00:00:00","date_gmt":"2014-12-17T00:00:00","guid":{"rendered":"http:\/\/otava.test\/backup-is-the-sweet-spot-of-the-data-protection-spectrum\/"},"modified":"2014-12-17T00:00:00","modified_gmt":"2014-12-17T00:00:00","slug":"backup-is-the-sweet-spot-of-the-data-protection-spectrum","status":"publish","type":"post","link":"https:\/\/www.otava.com\/blog\/backup-is-the-sweet-spot-of-the-data-protection-spectrum\/","title":{"rendered":"Backup is the ‘sweet spot’ on the data protection spectrum"},"content":{"rendered":"

Data is growing exponentially. Not just in size, but also in importance to organizations. So how do you protect that data while simultaneously assuring it can be quickly recovered in case of an emergency?<\/p>\n

A recent Online Tech \u201cTuesdays at 2\u201d webinar discussed how organizations can turn offsite backup into a recovery option.<\/p>\n

In \u201c<\/em>Transforming Your Offsite Backup into a Real Recovery Option<\/em><\/a>,\u201d<\/em> guest host Steven Aiello led an in-depth dive into what he calls the \u201cdata protection spectrum\u201d and how companies should \u2013 and should not<\/strong> <\/em>\u2013 use products on that spectrum to create a better offsite backup and recovery solution.<\/p>\n

Aiello was formerly a Senior Product Engineer at Online Tech and currently works as a Technical Architect for Chicago-based IT consulting firm AHEAD<\/a>, where he helps build infrastructure and security solutions for large organizations in the financial, health care and other assorted verticals.<\/p>\n

Aiello discussed the following data protection spectrum: availability, replication, backup and storage. While discussing the pros and cons of each category, Aiello stressed that all of them can play a key role depending on an organization\u2019s needs.<\/p>\n

\u201cIt\u2019s not that any one is better or worse than the other,\u201d he said. \u201cIt\u2019s using the right tool for the right job, and how all of these pieces plays into your data protection strategy.\u201d<\/p>\n

He added later that a single organization could find that it needs to be using every entry on the spectrum. \u201cMost likely, for a large organization, that\u2019s exactly where they need to be.\u201d<\/p>\n

Below is a brief recap of Aiello\u2019s points. For a deeper dive, a replay of the webinar is available<\/a>.<\/p>\n

AVAILABILITY<\/strong><\/p>\n

By definition, availability means one thing: Applications never go down. It also means another thing: A huge capital investment. Aiello says that in exchange for six 9s \u2013 or 99.9999% uptime \u2013 companies need to be willing to hand over a six-figure check.<\/p>\n

But it could very well be worth the price. Aiello uses the example of a trading house that is trading billions of dollars of securities each day. That same company can lose tens of millions of dollars if an application is down for minutes.<\/p>\n

Availability has a very specific business use case, so we won\u2019t spend a lot of time on it here. But Aiello does note one distinct limitation: The laws of physics. \u201cThe fastest that we can shoot data across the globe is at the speed of light. If your application can only tolerate a couple of milliseconds of latency, the distance from point A to point B starts to come into play.\u201d<\/p>\n

REPLICATION<\/strong><\/p>\n

Replication serves another specific use case; reducing recovery point objective (RPO) and recover time objective (RTO) scenarios at a moderately expensive \u2013 yet considerably less than availability \u2013 price tag.<\/p>\n

Aiello said the benefit of replication is two-fold:<\/p>\n

First, data centers can be located further apart because the replication from data center A to data center B is not synchronous.<\/p>\n

Also, unlike availability, replication technology does allow for a moderate amount of data retention. Still, Aiello stresses that neither availability nor replication are the equivalent of data backup. Asking a company in the health care industry, as an example, to keep medical records for 21 years in a replicated model is not feasible. \u201cIt becomes too expensive,\u201d Aiello said.<\/p>\n

ARCHIVE<\/strong><\/p>\n

Skipping over \u201cBackup\u201d on the data protection spectrum for a moment \u2013 since backup was the primary focus of the webinar \u2013 Aiello described archiving as a business deciding \u201cthe likelihood of this data being relevant, past 30 days, is almost nonexistent.\u201d<\/p>\n

An archiving system is the cheapest on the spectrum because a business may never need the data. If it does, it will pay for the retrieval. That\u2019s a fair tradeoff considering the likely of needing it is minimal.<\/p>\n

\"\"<\/a>BACKUP<\/strong><\/p>\n

Aiello calls backup \u201cthe sweet spot\u201d in data protection because it\u2019s relatively inexpensive when compared to the replication and availability models but allows organizations to start building out a recovery plan.<\/p>\n

What is an appropriate backup solution? Along with being granular, able to restore specific files and having a catalog, Aiello pointed to the following \u201cmust-haves\u201d:<\/p>\n

Encryption.<\/strong> \u201cThat was something we took a great, great deal of pride here at Online Tech, in working together. It was always about protecting the customer data. That goes from the primary storage platforms that the production systems run on, all the way to those backups as they leave the production environment. They are encrypted in transit, and they\u2019re encrypted at rest.\u201d<\/p>\n

Data durability.<\/strong> \u201cBackups should be durable. When you pick a backup solution \u2026 you are essentially betting the company on how reliable that backup technology is. It\u2019s a big, big bet that you\u2019re making.<\/p>\n

\u201cWhen I look at a backup solution, the first question I ask, \u2018What\u2019s your data durability rating? Has your product literally ever lost data, because of a malfunction?\u2019\u201d<\/p>\n

While defining backup, Aiello found it important to discuss what backup is not. \u201cI\u2019m seeing some people in the industry try to use a couple of different things to be backup, which it\u2019s not designed for.\u201d<\/p>\n

The first is enterprise file sync & share (EFSS) products like DropBox or Box.com. There is no revision history in these products. Having multiple copies of any file that is only a point in time does not constitute a backup. It constitutes multiple copies of a file, but it doesn\u2019t allow you to step back in time to recover corrupt data.<\/p>\n

The second is snapshots, a tool used as a feature of a replication strategy that cannot be confused for being a backup. First, if an administrator is running servers with a bunch of snapshots on them, there will be a decrease in performance. They\u2019re not designed to be kept around for a long time. Also, they\u2019re not granular and there\u2019s no catalog. \u201cIf you have to find one file out of three-quarters of a million, and it\u2019s been given some sort of random label, your RTOs and RPOs just explode,\u201d he said. Finally, snapshots offer no encryption.<\/p>\n

\n

DISASTER RECOVERY WORKSHOPS<\/strong><\/p>\n

Want to hear more from Aiello and other disaster recovery experts? Join us at a one of our upcoming half-day workshops scheduled to be held at our Indianapolis Data Center on Thursday, Dec. 22 <\/a>and our Metro Detroit Data Center on Thursday, Feb. 19<\/a>. Attendance at both workshops is limited to 50, so register now!<\/p>\n

Aiello will be joined on our expert panel by Baseline Data Services founder and president Lance Thompson and Online Tech Director of Infrastructure Nick Lumsden. They’ll present the following agenda at both workshops:<\/p>\n