Now that we’ve explored the challenge of too many products in Part 1 of the series, it is time to turn our attention to a familiar topic. I’m sure you are all “shocked” by the title and wondering if this blog has turned into the “Captain Obvious” report. Indeed, we all know backup takes too long, but sometimes it’s good to compare yourself to others to see how your experiences stack up.
As someone who has been around the backup world for quite some time (I started off at Cheyenne Software, which my fellow old-timers will remember well), the one constant in this business has been that backups take too long. Technology has changed quite a bit over the years, but the complaint remains the same.
Lately, the problem has increased exponentially as data growth has exploded. So as part of our IT survey we asked users about it, and 26 percent answered, “Yes, data growth has had a very significant impact on our data protection and is causing problems.” That number may seem a bit low, but it is a very serious statement about a problem that exists right now for those respondents. The scope of the problem is fleshed out by the next response, where 51 percent answered, “Yes, data growth has had some impact on our data protection but we are able to handle it.”
That’s not a statement that data growth isn’t a problem, but rather that data growth isn’t a problem yet. It also doesn’t reveal what the users may have done to deal with it. There may be quite a bit of radical backup re-engineering behind the comment that “we are able to handle it.” And for others, it’s very likely just a matter of time.
A final 23 percent said that “data growth has not had an impact on our data protection.” That is partly a function of company size. When we broke out the data between companies above and below the 500 employee limit, it turns out that only 16 percent of the larger companies said “no impact.”
Well, what about actual backup times? We broke out those questions in various forms. Since most people are still stuck in the mold of nightly incremental backups and weekly full backups, we started with “How long does it take your nightly incremental backups to complete?” and got these results:
We see that 32 percent of respondents are in what I would call the “danger zone” of taking more than 7 hours to get a nightly backup done. This is becoming an excessive length of time. What about full backups?
This is where things go from bad to worse. More than half of respondents are taking more than 11 hours for a full backup cycle, and a significant number are in the 24 hours and beyond range. This is cause for serious alarm.
One last cut at the data was to ask about the length of the single longest backup job? This is that one server that stands out amongst all the others as the toughest backup nut to crack.
We have 39 percent of respondents having one server taking more than seven hours to protect. And what happens if that particular backup job fails right towards the end? Trouble is what happens!
Backup times are unreasonable for a significant number of the respondents in our survey. Data growth has overwhelmed legacy technologies and the problems this creates are obvious.
The answer to this isn’t putting in more tape drives or adding more media servers (which you should get rid of anyway). The answer is fundamentally changing your approach to backup by moving less data and not backing up the same data over and over again.
Reading and moving data only once is a core technology strategy of the NetApp Syncsort Integrated Backup (NSB) solution, and it’s the only realistic way to deal with massive data growth. That’s how our users such as Campbell Alliance can reduce backup times from 13 hours to “an hour or two”. You can do the same and we’d be happy to show you how.
