August 13, 2007
Hosted Software Delivery Assurance
Seldom do we incur a sales prospect who doesn’t inquire on our ability to assure continuous and uninterrupted hosted software delivery. In the wake of the recent San Francisco area power outage and subsequent downtime incurred by several popular SaaS solutions (fortunately not us), I’ve received several inbound emails from existing customers generally querying about our data center resiliency and whether we’re vulnerable to a similar outage. I’ll use this blog post to respond to these types of requests.
At a high and non-technical level, Aplicor hosted software delivery can be initially summed to disciplined governance, veteran staffing and multiple layers of redundancy. Our governance model has evolved over many years and incorporates both the ISO 27001 and NIST (National Institute of Science and Technology) frameworks. These frameworks are specific to hosted software delivery, data center operations, information security, disaster recovery and business continuity. Aplicor is audited and certified annually by both ISO and NIST.
For purposes of this post, I’m going to refrain from getting into the multiple technical aspects of platform and hardware redundancy and speak to some of our core tenants, which include the following:
- High Availability (HA) global software delivery. We deliver our hosted software from multiple global data centers. The data centers are mirrored in real-time so that if either were to partially or entirely fail, the remaining mirrored location would continue with no service interruption. It surprises me that Aplicor is the only CRM or ERP SaaS organization to implement this procedure.
- N+2 equipment redundancy. Most data center operations strive for N+1 redundancy – meaning that for any point of failure, there is one backup mechanism in place. In 2003, Aplicor began upgrading to N+2 redundancy for an extra level of assurance. For reference, that extra assurance paid off for us last year when we incurred a router issue which subsequently corrupted the immediately backup. It also surprises me that Aplicor is the only ERP or CRM SaaS organization to implement this procedure.
- Isolated tenancy. We thoroughly evaluated multi-tenant application hosting many years ago and while we acknowledge it lowers cost to the publisher, it is also clear that the advantages do not extend to clients using the shared delivery model. We therefore designed an isolated tenancy architecture whereby every client is provided a unique and private database for their hosted application. In addition to the benefits of system performance, more granular security, assured privacy and application versioning for our clients, this architecture also provides a diversification for some data center delivery failure situations – meaning that if some systems we’re to fail, the situation would not necessarily affect all customers. To my knowledge, the only other ERP or CRM software company to implement isolated tenancy is SAP.
As the saying goes, the proof is in the pudding. Aplicor has been fortunate to achieve uninterrupted and continuous data center delivery since implementing these measures in April 2003. In order to achieve transparency and further provide visibility to both uptime and performance factors which affect users system performance and the user experience, we implemented the “Performance Visibility” web page at www.aplicor.com/sla.htm. There’s much more to our data center operations than I can include in a blog post. Please let me know if I can expand upon this topic in any way.
Technorati: Aplicor, SaaS  Add to Technorati Favorites
Del.icio.us: SaaS, Aplicor  Save this page to del.icio.us
View CC license
Posted by Chuck Schaeffer on August 13, 2007 in Hosted Delivery
Permalink | Comments (4) | TrackBack (0)
Why the Netherlands?
Nice talking to you at the Gartner CRM summit in Hollywood last week. We're about to embark on a project to find a data center in Europe. I know you guys have your data center for that part of the world in Amsterdam. Why did you choose the Netherlands over London or other locations?
Posted by Shawn Hannon on September 24, 2007
Hosting in Amsterdam
When it came time for us to implement globally redundant data centers we performed a fairly comprehensive data center destination search before selecting the Netherlands. Our primary goals were reliability and performance so selection criteria were focused on fibre infrastructures, zero-mile hops, pan-European latency optimization and several technical variables which influence thin-client application delivery. After about a nine month technical analysis, it became very clear the two best alternatives were London and Amsterdam. We ultimately selected the Netherlands. The Netherlands is counted as one of the most wired countries in the world, is home to the largest internet hub in Europe and our application delivery analysis showed particularly impressive IP performance metrics from Amsterdam to the European Union, Eastern Europe, the Middle East and Africa. Let me know if you would like to review our anlaysis, happy to share it.
Posted by Chuck Schaeffer on September 24, 2007
SaaS on a Seismic Fault Line
Hi Chuck. I read a Information Week article and an online article (at www.online-crm.com) indicating that SaaS vendor NetSuite filed an S1 to go public last month, and in that document NetSuite admitted to having no redundant data centers for their application or their client data. The Information Week article quoted numerous alarming statements, including that the only NetSuite data center resides near or on a seismic fault line. This statement that really brought it all home for me was:
"In the SEC filing dated July 2, NetSuite acknowledged using a single data center to deliver its services. "We do not currently operate or maintain a backup data center for any of our services or for any of our customers' data, which increases our vulnerability to interruptions or delays in our service," the company (NetSuite) said.
How can this be? Is this the norm or the exception among software as a service vendors?
Posted by Joe Wassem on August 20, 2007
SaaS on a Seismic Fault Line Response
I also read the many industry articles discussing NetSuite's S1 filing and presumed admissions and I can only assume that NetSuite's forwardness in publicly recognizing these deficiencies has much more to do with mitigating liability (and the potential exposure for class action lawsuits from soon to be new shareholders) and was unfortunately for them picked up by several customers and media organizations. There's no doubt that this behavior is unacceptable by a custodian entrusted with their clients' mission critical business data, however, I'm confident in saying that this is the exception and not the rule among the many credible SaaS software organizations.
Posted by Chuck Schaeffer on August 22, 2007
Trackback 
Trackback URL for this entry: http://www.aplicor.com/blog/070813.htm
Post a comment 
Comments are moderated and will not appear on this weblog until approved.
|
