Archive for the ‘Web security’ Category:
While our “Future of SaaS” Survey has only been open for a few days, the trend of the responses to the question: Are There Any Concerns or Reluctance to Using Cloud-based Services? are clear. Even though the remaining options are getting votes, the top four answers are as we expected.
- Security - Does anyone have unauthorized access to my data in the “cloud”? (59% of respondents)
- Uptime – Can I get to my data when I need to? (64%)
- Data Ownership - Who owns our data? (55%)
- Retrieve-ability – Can we retrieve our data if we want to take it back on-premise? (50%)
Respondents are permitted to select more than one concern.
We may see some shift in the ordering of these concerns but, we expect these to remain the top four. We’ll cover each of the four in the next several blog entries. The blog on Security was posted earlier.
Click the link to download the results in our Whitepaper- The Future of SaaS – Survey Results.
Uptime – Can I get to my data when I need to?
What good is secure data if you can’t gain access to it? Microsoft’s recent and recurrent downtime problems have brought this issue front and center again.
Having a service hosted is an important part of your business. To a SaaS provider, hosting services is their entire business. SaaS providers stake their entire reputation on uptime and security. A loss of reputation in either one will see a rush of businesses migrating to a more capable provider.
Concerns about uptime involve four main questions:
- What percent of time is the service guaranteed to be available?
- What counts as downtime?
- How does a SaaS provider prevent downtime?
- Are there penalties involved for excessive downtime expressed in the Service Level Agreement?
How many “9″s is enough?
Uptime is measured as a percentage of service availability against unavailability. An uptime of 100% may be desirable but if it were possible to achieve, it would be very expensive. So the question is, “How much uptime can we afford?” or “How much downtime is tolerable?”
Some large SaaS providers promise 99.9% uptime. That sounds like a great until you realize that 99.9% uptime means downtime of over 43 minutes per month or over 8¾ hours per year. This may tolerable if the downtime occurred only on Sundays at 2 AM. Unfortunately, that is rarely the case. The best SaaS providers will guarantee 99.999% uptime; only 25 seconds per month or 5 minutes of downtime per year.
As you consider migrating from an on-premise solution to a SaaS solution, compare the downtime promised against the downtime in your own data center. Usually, SaaS providers’ uptime far exceeds that of the data centers in the small- to medium-sized businesses.
Scheduled vs. unscheduled downtime
So, when a server needs a patch or a switch needs to be changed out. Does that count as downtime? Well, sometimes.
Some providers define scheduled maintenance as “planned outages” rather than downtime. Most on-premise data center tolerate a level of planned outages for upgrades, reboots & patches. But, what about critical services for which any downtime costs an organization lost revenue? It costs an airline reservation system approximately $89,000 per hour of downtime. In that case, downtime is downtime period.
When evaluating a potential SaaS provider find out their SLA has to say about which type of downtime counts against their uptime calculations.
How does a SaaS provider prevent downtime?
When an on-premise server is down, only that organization is affected. When a SaaS provider’s multi-tenant server goes down, all the affected organizations are down. A excellent SaaS provider has clusters of servers with built-in redundancy and replication. These clusters are often housed in geographically dispersed data centers. That ensures the services will be available when you need them. Our recommended email archive SaaS provider (Sonian) has a cross-cluster uptime of 99.999,999,999% (that’s 11 “9″s) of uptime.
With this level of SaaS providers, a client’s data is hosted on a primary server. That server is coupled with a redundant server in a different cluster as well as a backup server. We have a vendor-partner with such an arrangement which has a promised 99.999% uptime and an historic track record of 100% uptime over 10 years.
Penalties for excess downtime
An explanation as to why we are such strong advocates of cloud-based services may be in order. Several years ago the power distribution room in our office building blew a fuse. It took building management four days to restore power. What did they offer their tenants as recompense? A Mexican buffet complete with a Mariachi band. Now, I enjoy occasional Mexican food and music but, four days without power and all they come with is tacos and re-fried beans? Really?
Leaving the lessons learned about disaster recovery and redundancy plans aside, What financial compensation for excessive downtime does the SaaS provider offer? The best ones offer a sliding scale based based on monthly calculations. The more downtime, the greater the penalty.
So, how do you become comfortable with a potential SaaS vendor? Ask questions. Contrast their solution among vendors and your on-premise solution. For a list of good questions to ask, read our, Questions to Ask a Potential SaaS Provider.