If virtual server is used,
rather than traditional physical server, then this problem will not occur.
Virtual servers can be
provision with maximum harddisk space required by the content server – by
overprovisioned. Then configure the disk space required for thin provisioning –
so that unused space will be freed for other usage. However, if more disk
spaces are required, virtualised layer will provide more diskspace as needed
based on maximum diskspace required.
As virtualisation are
required to run on SAN storage, disk arrays will be available and disk spaces
will be abundant.
Replication through
virtualisation will be a breeze because a redundant VM for another video server
can be spawned easily and disk storage can be provisioned easily.
Also instead of replication,
another strategy for maintenance can be used – using vMotion through DRS
(Distributed Resource Scheduler) to move VM from one physical server to another
VM of another physical server. Then maintenance on the primary physical server
is a breeze as the vMotion ensure “work on the fly” with almost no downtime or
disruption while migrating to the secondary server – ensure a very high SLA to
the public.
This is another classic case
of SuperConvergence – where media video content are IP-based and stored in IP
video server running on TCP/IP. Multimedia comprising voice, image, video and data are now running on
digitised IP channel instead of the analogue broadcast channel ---- in this era
of SuperConvergence.
So H.323, H.320 voice and video protocol – now run on SIP (Session Initiation Protocol) which is IP-based.
Of course must be willing to pay for good experience IT professional.
But on retrospect, how come the contractor never check the diskspace required in Primary server and the diskspace available in the Backup server before replication?
For RAID 5 – if one disk fail, data still can reconstruct.
For RAID 6 – if 2 disks fail, data still can reconstruct.
For RAID 10 – if one set of disk fail, another set of disk can reconstruct.
But SAN storage using
fiber-channel or FCOE will be a more efficient and faster storage solution over
NAS that used IP.
Moreover, disk array allowed disk expansion to increase disk space. Also
redundancy can be configured for RAID such as RAID 5, 6, 10 - single parity
disk, dual-parity disk or mirroring and striping.For RAID 5 – if one disk fail, data still can reconstruct.
For RAID 6 – if 2 disks fail, data still can reconstruct.
For RAID 10 – if one set of disk fail, another set of disk can reconstruct.
No comments:
Post a Comment