email us at amarts@artifact3.com

Name: Amin Marts. No worries if you mispronounce the first name, I’ll correct you.

What I’m interested in: Simply, finding solutions for problems, excuse me, challenges.

What I bring to the table: Big picture thinking. Ruthless execution. The ability to distill complex topics into bite-sized and easy to digest contextual stories.

Turn On’s: Smart design. Simple solutions to complex problems. Billboards that make me stop, think, and take a picture to share with you. Jargon. Most of all, flowy singletrack.

Turn Off’s: Jargon. Overly complex solutions that don’t scale. Ambiguous messages. Unintelligible handwriting. Skiers who wear Starter jackets.

My Favorite Tools: iPad, Keynote, Mindmeister, DropBox, old skool whiteboards, OmniFocus, Glenn Beck.

Things I (admit) I read: HBR, Economist, anything Seth Godin writes, WSJ, on occasion the International Herald, when feeling snarky The New Yorker.

January’s Wish: To lose the twang I’ve developed from watching too much CMT during Christmas.

Best way to contact me: Through my twitter or email. I’m easy to find.

Posts Tagged ‘virtualization’

04.07.2011

Solve The “Mobile Backup” Challenge with Desktop Virtualization

In the March 2011 issue of Storage Magazine there is an article by industry guru Curtis Preston on “mobile backup”. The scenarios outline data protection scenarios for laptops and ROBOs (remote offices branch offices). Preston nails the caveats associated with ROBO backup. He’s spot on and there’s little I’d add to his arguments, management suggestions and thought process. On the topic of protecting the mobile user specifically, he missed.

Preston correctly asserts, as users have become dispersed they’ve taken data with them on their laptops, away from the core. This (migration away from the corporate LAN) makes it difficult to backup and protect local data. A common and easy method for ensuring locally created data is saved back to a centrally backed-up file server is by mapping local directories to a central file server. The file server in this case is part of the enterprise backup scheme. This works brilliantly when the user is on the LAN but falls short when they’re disconnected or the file sever isn’t accessible. Moreover, if the user saves data to a local directory that isn’t mapped to a file server he might as well be on the moon.

To mitigate the risk of IP dying on mobile endpoints, there are products such as Carbonite, Mozy and SmartSync which backup local data to either the cloud or the data center. These products work well as tactical point-and-shoot solutions but this is 2011 and you can do so much better. Think strategically.

Enter the distributed desktop. This is not a conversation on desktop virtualization per se, however a core capability of the concept is enduser environment protection. Note, I’m saying “environment protection” opposed to “data protection”. “Data Protection” is what you’re used to. It’s synonymous with flat files, such as Word documents, Excel spreadsheets, images, pdf’s, etc. “Environment Protection” encompasses the aforementioned in addition to applications, end user customizations to the operating system, and the operating system itself. This is accomplished via block level snapshots which are then transmitted back to a centralized server at specified intervals. To preserve bandwidth on the WAN or LAN only the deltas are transferred.

Native user data backup and restore capabilities coupled with smarter, granular and faster provisioning, and enhanced security makes the distributed desktop the go to concept for “mobile backup”.

 

Posted by on 04.07.2011 at 4:08am   Tags: , , , , , ,
0 comments   Email this post

01.03.2011

A Visual Look at Desktop Virtualization – Part 1

Desktop virtualization has been a point of interest for many SMBs of late. Like most newer concepts, it started as a cool idea but it’s moving steadily to an enterprise ‘must have’. An enterprise ‘must-have’ if you’re concerned with increasing efficiencies, enhancing security and protecting data, all from a common platform.

Over the next 2 posts I’ll walk through:

  • What desktop virtualization is
  • Challenges it solves
  • Enterprise benefits realized
  • Who the market leaders are and how they differ

Take a peek below of market landscape. We’ll talk it out in just a moment.

Posted by on 01.03.2011 at 8:30am   Tags: , , , , , , ,
0 comments   Email this post

11.01.2010

Terabytes of Storage Dirt Cheap…Go to Staples

In a recent conversation the question was proposed; why is enterprise storage so expensive? There’s a lot that went into the answer however I thought the baseline of the conversation was interesting.

Disks are getting faster however their mechanical architecture has remained largely unchanged. Beginning with the mechanics, envision a record player. There’s a platter where data resides and an arm with a head that reads and writes information. As the actuator arm slides across the media a two-dimensional map is created. Three-dimensional data storage exists however it’s not mainstream, the use cases are limited, it’s widely unsupported by traditional file systems and it’s crazy expensive so we’re not going to talk about it here. Back to the real world, the platters spin at a constant rate. Constant spin rate means the edge passes under the heads faster than the data at the center (of the platter) which creates a diacritical curve in read/write performance.

Performance degradation is mitigated by modern files systems by grouping data sequentially, writing individual files contiguously and by leveraging radial placement. Radial placement is the process of placing active data on the (fast) edge of the disk and slower less active data closer to the center. The prevailing theme is reduced seek time which can also limit head thrashing which can reduce drive life.

Due in part to the inherent dismal performance of disk in comparison to processor and memory performance and the capabilities derived from virtualization, how would you increase the aggregate performance of disk? Simple, add more spindles. Enter RAID. RAID is fundamental when talking fault tolerance in addition to being an elemental component of overcoming the limitations of disk.

In a nutshell, adding disk spindles moreover aggregating spindles for particular workloads (e.g. email, database, file sharing, application performance) is more readily applicable than adding faster spindles. RAID in conjunction with the spreading of I/O across multiple hard disks in concert with striping and mirroring is at the heart of that expensive array humming in your data center. The aforementioned is not at the heart of your multi-terabyte USB drive that sits on your desk. Moreover, when it comes to housing line-of-business applications and data, a good chunk of what you’re paying for is performance.

Posted by on 11.01.2010 at 6:45am   Tags: , , ,
0 comments   Email this post

09.09.2010

Data Protection Conversation Starters

Shared storage and virtualization form the computing foundation for many mid-sized and most larger enterprises. Migrating the storage footprint from direct attached storage (DAS) to a shared storage model often includes virtualization at some level. This migration requires modernizing your data protection methodology.

Although organizations differ in their line-of-business (LOB) applications, data storage size and backup method of choice, some commonalities exist. I’ve compiled a short list of the thought process you should work through on your way to modernizing your data protection capabilities

1. Standardize:

When possible, invest in the same data protection technology for both your physical and virtual environments. Doing so will ease management as well as reduce the costs associated with licensing and training.

2. Invite conversation:

Organizations aren’t static. The impact on computing resources certain business units have on storage continually change. Has your data protection policy evolved to account for SharePoint, ERP, CRM or other new LOB applications? Has your organization acquired another organization who is now a business unit with niche compliance pressures?

3. RPO’s & RTO’s (Oh My)

What are you recovery-point objectives (RPO) and recovery-time-objectives (RTO)? As your business changes, so do the expectations associated with what is an acceptable amount of data loss following a ‘disaster’. In lockstep with defining the RPO is the RTO. Unlike RPO, which is a technical capability, RTO resides in the realm of business process. RTO speaks to the service level agreements that define when data will be available post-disaster or disruption to the affected entity be it the business, a department or knowledge worker.

4. Crash-Consistent vs. Application-Consistent Recovery

Virtualization has a profound affect on not only the required back-end infrastructure, but also the incumbent data recovery application and methodology. Simply taking a point in time snapshot of the virtual machine does not leave you with a viable backup of the server and application. In some cases a snapshot of the virtual machine will only provide you with a crash-consistent copy of the server. The unaware application (running on the server) will be out of sync and thus requires special handling. This special handling is unique to the application, however it must be taken into account with modernizing your data protection policy.

Specifically, the question that requires an answer is:

“What special handling is required of my application to synchronize it with the snapshot of the virtual machine?”

This synchronization between the virtual machine and application is referred to as being application-consistent. Special handling is not always essential, however you’ll need to be aware of when crash-consistent recoveries and application-consistent recoveries are appropriate.

Posted by on 09.09.2010 at 9:52am   Tags: , , , , , , , , ,
0 comments   Email this post